Friday, 8 December 2017

T-72: Part 2

Due to the length constraints imposed by Blogger, the original T-72 article was split into two parts. This part covers the second half of the article. This is not a new article. You can view the first half here.


  1. Protection
  2. Common Characteristics
  3. T-72 Ural
  4. Gill Armour
  5. T-72A
  6. Kontakt-1
  7. T-72B
  8. How Does NERA Work?
  9. Kontakt-5

  10. Fuel Tanks as Armour
  11. Smoke Screen
  12. Escape Hatch
  13. NBC Protection
  14. Firefighting
  15. Entrenchment Dozer Blade
  16. Storage

  17. Mobility
  18. Engines
  19. Cooling System
  20. Transmission
  21. Suspension
  22. Water Obstacles
  23. Road Endurance
  24. Driver's Station


A good indication of a tank's true survivability is its resistance to catastrophic destruction, which can refer to the tendency for a fire to start and the likelihood of that fire spreading and consuming the entire vehicle or the possibility of the ammunition exploding. In this sense, the T-72 stands on equal footing with many of its contemporaries and surpasses some of its rivals. However, many modern tanks now include separated ammunition storage - something the T-72 lacks. Nevertheless, the protection level of the T-72 was remarkably high for its time as a result of its combination of thick armour and low silhouette, while the low placement of ammunition in the hull reduces its chances of suffering a direct hit and detonating. However, the location of the ammunition means that any fuel or hydraulic fluid leak (both substances are plentiful in the fighting compartment) will inevitably pool on the floor of the hull and a fire will reach the ammunition eventually. Nevertheless, this danger is mostly nullified by the automatic fire extinguishing system, so the survivability of the tank could be considered to be among the highest by Cold War standards.

As for armour alone, the protection qualities depend greatly on the variant being considered. As the years go on, the protection value markedly increases, reaching its zenith with the T-72B2 variant with the Relikt armour package. We shall examine the protection qualities of all the main variants in detail armour-wise.

The myth of the T-72's inferiority in terms of protection is just that - a myth. Various T-72s have proven their worth in various conflicts when placed under competent command, but the lack of media coverage on the successes tend to skew the view in favour of the image of burning wrecks. To list one incident in Grozny, in the year 2000, a T-72B with the tail number 611 took 3 hits from Fagot anti-tank missiles and 6 hits from RPGs during 3 days of intense fighting and remained in battle with only minor damage. Most of the hits landed on the sides of the tank, with one rocket impacting the lower rear of the hull These are the same types of weapons that an Abrams or a Challenger 2 faced during campaigns in the Middle East, and there are plenty of other cases to be found in the second Chechen war. One only needs to be motivated to search.

A lot can be said about the inherent design issues of the T-72, but one cannot accuse it of being made from inferior steel. As a testament to its quality, an ex-GDR T-72M1 tested in Meppen (details here) withstood 24 hits from a mix of 105mm and 120mm APFSDS and HEAT shells on the turret front without a single fracture or crack. Whether or not the shots defeated the armour is a different matter.


The hull side, hull roof, hull bottom and rear armour of all T-72s are identical, regardless of the variant. The hull side and the turret side are both 80mm thick, but the hull side thickness over the engine is slightly thinner at 70mm. The side armour is more than enough to withstand 20mm armour-piercing ammunition fired from various aircraft, such as the AH-1 Cobra firing the 20x102mm round, or A-1 Skyraider, firing the 20x110mm round. Ad hoc use of M61 Vulcan gattling guns on non-ground attack aircraft such as on the F-4 Phantom would not have yielded any better result.

Drive sprocket area. Note the thickness

This picture shows quite clearly how the upper hull side is thicker than the lower sloped side.

The side armour is thickest at the top half, visibly appearing bulkier (as shown in the picture above) both outside and inside, thinning down to 20mm with a modest slope at the roadwheel region. The upper and lower sides are not the same plate. The upper sides are a single, very long piece of steel, while the lower side is actually the same plate as the belly armour. The belly armour was bent into a tub shape and welded to the upper sides. This probably helped increase the resistance of the hull to the explosions of anti-tank mines under the tracks, as the shape encouraged the deflection of blast waves. This is speculation only, but it is supported by evidence that the Soviets were well aware of the improved mine resistance of the M48 Patton due to its arched hull belly design.

The interior surface of the hull sides is coated in a 20mm layer of anti-radiation lining, which can help absorb spall and other secondary penetrator fragments or even stop residual penetration from lower energy projectiles. This is discussed later in the "Anti-radiation" section below.

The thickness of the side armour can be clearly seen here

It is without a doubt that the sides of the tank were only sufficient for a very limited period of the service life of the T-72. Being only 80mm thick, the side armour plate could offer only a fraction of the protective value of the front armour, and this was not a trifling issue. The number of hits sustained by a tank's sides were statistically significant, as shown by the analyses conducted by Dr. Manfred Held in "Warhead Hit Distribution on Main Battle Tanks in The Gulf". The charts below are from the study.

The sides would have been mostly resistant against 105mm APDS like the L28 and its American derivative at combat ranges within a somewhat reasonable 40 to 50 degree arc, but this narrow arc limits the tank's freedom to maneuver in open spaces. The appearance of 105mm APFSDS rendered the side armour completely inadequate as protection against contemporary anti-tank firepower. 

The hull roof is 20mm thick, the rear armour plate is 40mm thick, and the hull floor is 20mm thick. The hull bottom is only sufficient against explosive charges with a mass of less than 10 kg detonated over the tracks and not directly under the hull. These parts of the hull are most likely constructed from the same steels used in the same locations in the T-54 and T-62; 49 S grade steel for rear armour plate and the hull roof, 43 PSM grade steel for the floor. These grades of steel were first used in the T-54 obr. 1953. The hull bottom is constructed from a single plate of rolled steel, which is then stamped into a complex shape with protruding ribs for the installation of torsion bars and a depressed section in the floor to accommodate the driver. Reinforcing nubs were pressed into the plate between every torsion bar rib to improve the stiffness of the floor. The side edges of the plate were bent upward at a 30 degree angle to join with the side hull plate, thus forming a tub shape. This was possible due to the ductility of 43 PSM steel, which is a soft annealed steel and cannot be considered equivalent to RHA. 43 PSM has a yield strength of 400 MPa and a tensile strength of 600 MPa, and a hardness of 180-250 BHN. These qualities make the steel plate easy to stamp, but also make it unsuitable as mine protection since steels like 43 PSM perform poorly under high blast loading compared to high hardness and high toughness steels. This is possibly compensated by the extensive ribbing of the belly plate, although the effectiveness of such a measure is unclear and probably insufficient. It does not help that there are a multitude of structural weaknesses in the front part of the belly plate where a tilt rod mine will detonate, namely the escape hatch and drain plugs.

The armour plates used in the side hull and front hull armour of the T-72 is medium hardness steel, most likely 42 SM steel with a hardness of around 340 BHN. The cast turret probably employs MBL-1 armour grade cast steel with a hardness of 270-290 BHN. This grade of steel was first used in the turret of the T-62. The HHS (High-Hardness Steel) employed in the tank is BTK-1Sh with a hardness of ~450 BHN. The appliqué armour plate used in the 1983 modification of the T-72A and earlier variants is most likely BTK-1Sh as well.

The lower glacis is a 80mm plate, sloped at 64 degrees. The properties of the plate are identical to the other welded plates used for the hull, like the side armour plate and the front plate of the upper glacis. Being a traditionally weak area of the tank, the relatively poor armour of the lower glacis is partially counteracted by its small size and low exposure to enemy fire. Furthermore, the presence of the upper glacis armour array partially reduces the height of the weakened zone. The photos below (credit to Stephen Sutton for left photo) shows the thickness of the front plate of the upper glacis array by the weld seam joining it to the lower glacis plate so that you can visualize the approximate reduction in the size of the weakened zone. The tank on the left is a T-72M1 (formerly Iraqi, disabled by air attack and abandoned almost fully intact) and the tank on the right is a T-72M, so both have a 215mm array with a 60mm front plate. The dozer blade overlaps with the array to minimize gaps in the armour.

At a 64 degree obliquity, the lower glacis is immune from anything short of 105mm APDS, but even so, the chances of L28A1 or M392 APDS penetrating this plate in real conditions are diminished by the strategically placed integrated dozer blade. The dozer blade is probably made from some high hardness armour grade steel, but it is also possible that it is made from high strength structural steel with a hardness of around 200 BHN which was used for common general-purpose commercial bulldozers at the time and is still used in modern ones today, but a high hardness armour steel grade dozer blade is far more likely for a military vehicle like the T-72 because a high hardness blade can be used to shift abrasive rock and frozen soil as well as provide additional ballistic protection. Case in point: high hardness impact resistant steels like Hardox 500 are standard for bulldozer blades and loader buckets used in mines and construction sites where large quantities of rock must be shifted, and high hardness steel dozer blades are also standard for military combat engineering vehicles.

Having the lower glacis backed by the upper glacis array at the top third of its profile and supplemented by the overlapping dozer blade for the other two thirds means that the vulnerability of the area to earlier 105mm APDS is significantly reduced such that it may possibly be immune to L28A1 from 500-1000 meters. However, later APDS rounds with more elongated tungsten cores and with tungsten alloy tilting caps would not find this part of the tank to be much of a challenge at any range. This is an unavoidable artifact of the glacis design that followed the design of the tank well into the 21st Century as it evolved into the T-90 series.


The entire turret is made of cast steel. The side armour is curved at a considerable rearward angle to form a point at the very back of the turret. This was especially exaggerated in the T-72B variant due to the bulging turret cheeks. The rear of the turret of all variants have a distinct step joining the roof to the cast base.

The stub ejector port is also visible here

As mentioned before, the side of the turret is 80mm thick, thinning to around 40mm at the rear. The vertical slope of the turret provides a nominal increase in relative thickness to around 88mm when viewed perpendicularly, but the amount of vertical sloping is very minor as the turret is built with a heavy emphasis on horizontal shaping. With a thick layer of anti-radiation lining backing it and with the storage bins (plus cargo) acting as rudimentary additional armour, the sides are more than enough to withstand any 20mm and 23mm shell at point-blank and any 25mm autocannon shell at typical combat ranges (in the vicinity of 1500m) when hit at a perpendicular angle. This is including the 25mm M919 APFSDS shell. However, the armour is not thick enough to reliably resist 30mm, 35mm and 40mm shells. Still, with some extra angling, the side turret would have very good prospects. The rear of the turret, however, is completely hopeless.

The shape of the turret of the T-72A and T-72B is such that the sides will be completely unreachable by enemy fire from within the frontal 70 degree arc. This means that if the turret was shot at a relative angle of 35 degrees, the thin sides of the turret will be hidden behind the turret cheeks. If the relative angle is increased to 45 degrees, the sides will be visible, but then the angle of incidence will be so steep (80 degrees) that shaped charge warheadwillil fail to fuse and almost all APDS or APFSDS projectiles will ricochet. Even if an attacking projectile manages to dig into the armour, the LOS thickness of the thin side armour at such an angle is very formidable at 460mm and it is further enhanced by the reduced penetration efficiency at such an angle, so that it is close to the thickness of the turret cheeks. In other words, the turret of the T-72 is a very, very tough nut to crack from a wide range of angles. The unique teardrop shape of the turret makes it possible to present a uniform thickness of armour across the frontal arc and also do so without adding excessive weight to the tank, but the shape could not have been implemented without a two-man turret, which in turn could not have been accomplished without an autoloader.

On another note, it is interesting to observe that although the turret of the T-72 lacks handrails for tank riders like preceding Soviet tanks, the practice of hitching a ride was still occasionally taught and exercised.


The T-72 Ural, or Object. 172M, was the original T-72 and is the least technologically gifted among its "brothers". The hull glacis armour benefited from a composite construction taken directly from the T-64A, and the turret was a single homogeneous steel casting. The all-steel turret was closely based on the turret of the T-64A but differed somewhat in the shape of the frontal profile and most of all in the shape of the back half, as it had a distinctive step between the almost-flat rear armour and the heavily sloped roof to house the AZ autoloader ammunition lifter and rammer mechanisms.


The timeline of the evolution of the hull array is as follows (front to back):

1973: 80mm RHA + 105mm STEF + 20mm RHA

1976: 60mm RHA + 105mm STEF + 50mm RHA

1983: 16mm Appliqué + 60mm RHA + 105mm STEF + 50mm RHA

The original upper glacis armour for the T-72 Ural from 1973 is a composite sandwich consisting of a 105mm "steklotekstolit" (glass textolite) layer sandwiched between an 80mm RHA front plate and a 20mm RHA backing plate. The total thickness is 205mm to the normal, but the glacis is angled at 68 degrees to produce a total LOS thickness of 547mm. The source for the 105mm thickness figure of the glass textolite layer is "Kampfpanzer: Heute und Morgen" by noted German armour expert Rolf Hilmes and various other Russian documents. Two glass textolite plates were pressed together to form the 105mm layer. Some publications mention that the 105mm glass textolite interlayer was split evenly into two 52.5mm layers, but according to Rolf Hilmes, the 105mm glass textolite interlayer was split into a 60mm plate and a 45mm plate.

The armour array is identical to the array used in the T-64A, so we will be using documentation for the T-64A as part of our short analysis.

According to the page below, the original requirements for the new prospective main battle tank of the USSR dictated that it had to be immune to 100mm armour piercing shells fired at 1000 m/s (normal muzzle velocity is 895 m/s) and 105mm subcaliber shells (does not differentiate between APDS and APFSDS) fired from the American M68 cannon at a distance of 1000 meters. The armour was also required to be immune to 85mm HEAT as well as 105mm HEAT fired from an M68 cannon. However, these figures were corrected by Alexander Morozov on the very same page (chief designer at the KMDB design bureau responsible for inventing the T-64) so that the armour was required to be immune to 105mm subcaliber shells at 500 meters and 115mm HEAT shells with a copper liner (ordinary ones used a less effective steel liner). The required RHA equivalence was around 330mm against KE threats and around 450mm against HEAT threats. It is noted in the margins that the data on the performance of the M68 was estimated and was subject to change, which may explain the revision of the protection requirement from immunity from 1000 m to immunity from 500 m.

It seems settled, then, that the hull armour of the T-64A is worth 330mm RHA against KE threats and 450mm against HEAT threats so the T-72 Ural would have the same amount of armour. After all, there is no source more authorative than the original specifications table, and it was edited by Morozov himself, no less! However, the table below states that the resistance of this same armour (on the T-64A) is equivalent to 305mm of RHA steel against KE threats and 450mm RHA against shaped charges. Either the table is a fabrication, or clearly, the figures for KE threats are more nuanced than they appear and cannot be taken at face value.

The high obliquity of the glacis armour presents a mixture of advantages and disadvantages, but the composite nature of the array makes the true value of the armour much more nuanced than it appears at first glance. The most obvious advantage of steep angling is that the penetration power of earlier APDS rounds will be drastically reduced and some HEAT warheads may even fail to fuse on impact, but the disadvantage is that long rod penetrators will actually penetrate more armour at higher angles up until the critical ricochet angle, which is usually around 80 degrees and above and depends on the length to diameter (L:D) ratio of the penetrator. It is known that the higher penetrative power of long rod penetrators on high obliquity plates is caused by the asymmetry of forces acting on the back of the plate as the penetrator passes through, but the impact and breakout effects for a finite thickness plate are often ignored. In reality, the optimum armour angle to maximize the penetration of long rods is 60 degrees, but between 60 to 80 degrees, the penetration of long rods will be affected by increasingly severe losses that may outweigh its increased performance.

Against HEAT warheads, the principle benefit of the high slope of the armour is that some warheads may not detonate properly. During the famous Yugo tests, the 90mm M431 HEAT shell with the M509A1 PIBD fuze was demonstrated to have a very high probability of failing to detonate against the 60-degree upper glacis of the target tank (a T-54) when the tank was angled 20 degrees sideways. Using the compound angle table on page 47 of WWII Ballistics: Armor and Gunnery, we find that a 60 degree vertical slope and 20 degree horizontal slope creates a compound angle of 62 degrees. The 68 degree slope of the T-72 upper glacis exceeds this in excess. Although 90mm guns were obviously obsolete in the face of the T-72, the newer 105mm M456A2 HEAT shell also uses the M509A1 PIBD fuze, so the results of the Yugo tests imply that M456A2 will also struggle to properly fuze on the upper glacis of the T-72. If a shaped charge succeeds at detonating on the upper glacis, it will be handled by the composite armour array which is reportedly rated for 450mm RHA against shaped charges. Against kinetic energy threats, the array is reportedly rated for 305mm or 330mm RHA, but the type of threat was not specified, i.e steel long rod, tungsten carbide APDS, tungsten alloy long rod, etc.

Almost all of the amateur attempts to distill the relative RHA efficiency factor of glass textolite use the thicknesses of each individual component of the armor as given, and almost all of these attempts are fundamentally incorrect. For example, the most common method is to subtract the physical LOS thickness of the steel plates of the array (80mm + 20mm) from a relative RHA thickness figure, which we can take to be 330mm in this case. 80mm + 20mm is 100mm, which when divided by the cosine of 68 gives a physical LOS thickness of 267mm, and 330mm minus 267mm is 63mm. By this logic, the 63mm figure therefore denotes the resistance of glass textolite in terms of RHA. The 105mm glass textolite layer has a physical LOS thickness of 280mm when angled at 68 degrees, so 280mm of glass textolite is ostensibly equal to 63mm of RHA steel, giving it a thickness efficiency of 0.225 against KE threats and a mass efficiency of 0.98. This is fundamentally incorrect.

As most of the armour community knows, long rods penetrate more armour at higher obliquity than at lower obliquity. For a plate angled at 60 degrees, a generic tungsten alloy penetrator pierces 1.17 times more steel in physical thickness compared to a plate at 0 degrees, whereas for a plate angled at 68 degrees, the same penetrator pierces 1.24 times more steel compared to a plate at 0 degrees. Therefore, the 100mm steel layer would not be directly equivalent to 267mm RHA in effective thickness, but actually equivalent to 267mm multiplied by the reciprocal of 1.24, or 215mm. Subtracting this from 330mm gives us 115mm, so 280mm of glass textolite would therefore be equivalent to 115mm of RHA steel and it should have a relative thickness efficiency of 1.79. Unfortunately, this is also incorrect. The increased performance of long rod penetrators at high obliquity affects glass textolite panels as much as it would affect steel plates, so 280mm of glass textolite would be worth much less than 115mm RHA. When all is said and done, adding up the steel and glass textolite as individual components gives us much less than 330mm, so it is plain to see that something important is missing and that these crude methods are far too simple to apply to composite armour. The first and most obvious mistake would be to combine the 80mm heavy front plate with the 20mm back plate and treat it as a single 100mm steel plate, and the second mistake would be to assume that the angle of the plates invariably weakens it against KE threats. For example, the Lanz-Odermatt equation reveals that a generic tungsten alloy long rod penetrator will defeat around 1.24 times more armour at 68 degrees obliquity compared to 0 degrees, increasing to as much as 1.48 times more at 80 degrees, but in reality, the rod may ricochet from the surface of a thick plate sloped at 80 degrees and leave only a shallow crater on the surface of the plate. Soviet engineers paid close attention to the problem tank protection and gave each element of the armour array its own special purpose so that all of the individual layers added together would be more than the sum of its parts. To better understand the capabilities of the complete array, it is necessary to know the function of its individual parts, beginning with the glass textolite interlayer.

Glass textolite is a material consisting of layered sheets of glass textile bonded by resin and pressed together. Glass textolite is not the same as fiberglass, because glass textolites are manufactured using laminated sheets of glass matting bonded together by resin, whereas fiberglass is manufactured using continuous glass fibers or chopped strands suspended in resin. Both contain glass fibers, but the use of fiber sheets in glass textolite makes it stronger than regular fiberglass.

A U.S Army technical translation of the "Plastmassy v bronetankovoy tekhnike" (Plastics in Armor Materiél) technical document originally published by the USSR Ministry of Defence in 1965 gives us some information on the glass textolite and fiberglass types used in the Soviet Union that would have been used in the armour of the T-72. The Eurokompozit website also gives a description of the glass textolite used in the T-72 which we can cross reference with the Soviet document. It mentions woven glass roving (rovings are woven bundles of glass fibers) and special phenolic resin as the matrix material, and the phenolphenolic resin-based glass textolite (steklotekstolite) listed in page 24 of "Plastmassy v bronetankovoy tekhnike" matches the description exactly. From this, we can be absolutely certain that the density of the glass textolite used in the T-72 is around 1.8 g/cc. Referring to the table of material properties, the specific type of glass textolite used in the armour has a tensile strength of 274.6 MPa, compressive strength of 294.2 MPa, flexural strength of 382.5 MPa and a specific impact strength (toughness) of 4.7-5.4 MPa.

Rolled AG-4S phenol resin-based fiberglass from the AG-4 series of fiberglasses matches the description to some degree, but this conflicts with Russian sources that explicitly state that "steklotekstolit" was used. Furthermore, AG-4S uses continuous parallel glass threads, not woven glass rovings as described in the Eurokompozit website. The Eurokompozit website states that the glass textolite used in the armour uses a specially modified phenolic resin for the matrix, so it is not likely that a commercial glass textolite was used. The purpose of the modifications made to the phenolic resin in the glass textolite for the T-72 is not known, but it is well known that glass-reinforced plastics like glass textolite lose a significant amount of strength at very low temperatures where they may become susceptible to brittle failure, but phenol-based GRPs are less sensitive to lower temperatures and are generally more ductile at the cost of reduced mechanical properties compared to GRPs based on epoxide resins. Based on this information, the choice of a phenol-based glass textolite for the armour and the use of a modified phenol resin is probably related to the inflexible requirement for Soviet tanks to be operable in conditions of -50°C to +50°C.

According to the old NII Stali website from way back in 2003, the efficiency of multi-layered armour against APFSDS ammunition increases as the filler density increases at obliquities of 0 to 40 degrees, but conversely, the efficiency increases as the filler density decreases when the armour is angled at an obliquity of 60 degrees and more. The final remark is that the absence of a filler (air gap) leads to a "negative result", which can be interpreted in two ways: it could mean that a long rod penetrator is not significantly affected by air gaps so the penetration into simple dual-layer spaced armour is more than the penetration into a triple-layer composite sandwich with any of the aforementioned fillers at any angle of attack, or it could be interpreted to mean that the penetration of a long rod penetrator is increased when an air gap is present. The latter option is not supported by any scientific literature whatsoever and is directly contradicted by prior claims regarding spaced armour on the same web page, so the former option is most probably correct.

The site mentions that high strength steels, titanium, aluminium, ceramics and glass textolite were among the materials studied for composite fillers and that a 15-30% increase in mass efficiency could be gained from the use of composite armour of this type. Apparently, the largest improvement in mass efficiency was achieved with glass textolite. This is presumably related to the follow-on claim that the efficiency of composite armour increases when higher density fillers are used at an obliquity of 0-40 degrees whereas lower density fillers are used at an angle at 60 degrees or more, so the ~1.8 g/cc density of glass textolite would make it the least dense and therefore the most optimal of all of the tested materials for the 68-degree glacis. An aluminium filler was used in the turret of the T-64 (115) instead of glass textolite, which makes sense because aluminium is denser (2.7 g/cc) and would be a more efficient filler at the low 25-30 degree slope of the turret.

These claims also appear to hold true for shaped charge threats as well, as shown in "Jet Penetration into Low Density Targets". The simulations and experiments detailed in the paper used a 100mm plate of variable density placed in front of a filler of variable density to find the most optimal combination. It was found that the velocity of the shaped charge jet tip emerging from the 100mm plate tended to be lower as the filler density decreased, but the jet increased in velocity when the density of the 100mm plate was decreased. As you can see from the graph below, the most serious reduction in jet tip velocity occurs when low density material is placed behind a 100mm plate with high areal density (m=500 kg/m^2). Since the thickness of the plate is fixed at 100mm, achieving the 500 kg/m^2 areal density figure requires the plate to be made from a material with a density of 5.0 g/cc. The relatively high 7.85 g/cc density of steel makes it even more suitable for this purpose.

The paper goes on to detail that low density materials are more effective against particulated jets than continuous jets. The graph above was plotted with the assumption that the jet emerging from the 100mm front plate is continuous, but the mass efficiency of a filler increases as the density of the filler decreases if the jet is particulated as it enters the filler. As you can see in the graph below, the most serious reduction in jet tip velocity occurs when the jet passes through a high density plate (500 kg/m^2) and enters a low density filler, with the biggest reduction in velocity occurring when the filler density falls below 0.3 g/cc.

The most optimal configuration is to have a front plate of high areal density in front of a filler of low areal density. This ensures that the jet is particulated as it emerges from the front plate, so that the low density filler performs at an optimum level. This conclusion is reinforced by other studies on the topic of shaped charge jet penetration into multi-layered targets such as "Theory Of Penetration By Jets Of Non-Linear Velocity And In Layered Targets" by P. Chou and J. Foster, from which the drawing below was taken. For a double layer of two plates each with an equal thickness of 50mm, the residual jet emerging from an RHA-AL double plate is shorter and slower than the residual jet emerging from an AL-RHA double plate, indicating that the efficiency of the RHA-AL combination is more efficient and that the behaviour of multi-layered composite armour is directional in nature.

Having a 80mm front plate sloped at 68 degrees for a LOS thickness of 213.6mm, the armour of the T-72 Ural is more than enough to particulate any shaped charge jet from the era and beyond, yielding very high efficiency from the glass textolite filler. A filler with an even lower density may be preferable as the mass efficiency would improve, but the performance against KE threats may suffer and an excessive thickness of filler may be needed to achieve the same level of protection.

From all of this information, it can be deduced that the most optimal configuration uses a steel front plate angled at a very high obliquity with a very thick glass textolite back plate, but the final array design of the T-64 and T-72 Ural featured an additional 20mm steel back plate behind the glass textolite layer. The presence of the steel back plate was most likely meant as a final barrier against KE threats rather than shaped charges, although a steel back plate would certainly be effective at stopping residual jet particles. The path taken by Russian engineers to reach this solution is detailed by Andrei Tarasenko in his article on the armour of the T-64, where he also describes the armour of an early prototype: the Object. 432. The armour of the Object. 432 had the 80mm steel front plate, but had a 140mm low density filler of glass textolite behind the plate, which would be highly optimal against both KE threats as well as shaped charges. According to Tarasenko, this configuration was estimated to provide protection equal to 450mm of RHA against shaped charges. However, this configuration was changed to the familiar 80-105-20 combination in the Obj. 432SB-2 variant.

It is interesting to note that the original glacis design used a single piece 140mm glass textolite panel while the Obj. 432SB-2 used two 52.5mm panels instead. The true reason why the glass textolite layer was divided into two panels in the new glacis design is unclear, but the fact that a 140mm panel existed and was experimentally tested dispels the myth that two panels were used in the T-64A instead of one was because of some deficiency in the Soviet plastics industry. Still, it could be surmised that two smaller panels are easier to manufacture than one large panel, so this decision may have had the effect of reducing the cost of the tank or making it possible for smaller plants to produce the panels during wartime.

From one perspective, it could be said that the newer 80-105-20 configuration added another layer to the original two-layer design to make it a sandwich, but it would be more accurate to say that the new configuration substituted 35mm of low density filler for a 20mm plate. If we look at this design solution from the perspective of mass efficiency against shaped charges, the efficiency of the armour clearly decreased, because 20mm of steel is obviously much heavier than 35mm of glass textolite and the level of protection offered by the new configuration against shaped charges did not change; it was still equal to 450mm of RHA steel, as shown by the table below (row: T-64A, column: "KC").

According to the table, the 80-105-20 array at 68 degrees has a mass equivalence of 335mm of steel, and its resistance to KE threats is equivalent to 305mm of steel while its resistance of HEAT threats is equivalent to 450mm of steel. The mass efficiency of the array against KE threats is apparently equal to 0.9 and the mass efficiency against HEAT threats is 1.35. However, the mass efficiency of the array is unusually low so it has to be mentioned at this point that the 305mm RHA figure against KE threats definitely does not reflect the performance of the array at a 0 degree angle, nor does it reflect the performance of the array against APDS rounds or the composite APFSDS shells fielded by the U.S Army during the 70's. The first hint is that the armour value of the upper glacis of the T-62 is attributed with the same amount of steel as its LOS thickness: 200mm. In reality, a hypothetical modern tungsten alloy long rod APFSDS shell with 180mm of penetration at 0 degrees at distance "x" would go through the plate, because a long rod penetrator would penetrate more steel at 60 degrees than at 0 degrees by a factor of 1.168. To obtain its relative armour value at 0 degrees, the 200mm LOS thickness of the T-62 is multiplied by the reciprocal of 1.168 and 171mm is obtained. Only now, the fact that the same hypothetical long rod can penetrate the upper glacis of the T-62 at distance "x" is reflected in the numbers. However, APDS shells and earlier composite APFSDS shells do not perform better at higher angles and the obliquity of the 60 degree plate would make it more effective than 200mm. The 105mm L28A1 APDS round, for instance, penetrates 300mm at 0 degrees at 1 km but 120mm at 60 degrees at 1000 yards. In the same way as before, the 200mm plate would therefore be worth more than the given relative thickness of 200mm at 0 degrees. If the same treatment is applied to the armour array of the T-64A, the effectiveness should increase to somewhere above 305mm.

The second hint is that the table does not differentiate between cast and rolled steel. The T-62 turret is listed as having a thickness of 200mm, and the resistance to KE threats is also listed as 200mm even though it should be slightly lower. The third hint is that the 305mm figure would mean that the armour should not exist, because the original requirement was for the T-64A tank to be immune from 105mm guns from a minimum distance of 500 meters. L28A1 APDS penetrates 300mm of steel at 0 degrees at 1 km, so the armour would be handily defeated at 500 meters if the 305mm figure were true, whereas in reality, the shell would be stopped by the 80mm front plate alone due to the slope. Therefore, the mass efficiency against APDS must be higher than 0.9.

Note the given areal density figures: 785 kg/sq.m for the T-62, 980 kg/sq.m for the T-64A. Strangely enough, these figures are only for the physical thickness of the armour and not the actual LOS thickness. This can be demonstrated fairly easily: The steel upper glacis plate of the T-62 is 100mm thick, or a tenth of a meter, so naturally, the areal density of the plate is a tenth of the volumetric density of 7,850 kg/m^3, which is 785 kg/sq.m. When the plate is inclined at 60 degrees it has twice its areal density at a perpendicular angle so the figure goes up to 1,570 kg/sq.m. The upper glacis plate of the T-64A/T-72 Ural should be treated the same way as the given areal density of the array is for its physical thickness of 205mm. When the plate is inclined at 68 degrees, the true areal density is 2,616 kg/sq.m. This is practically the same as the 2,609 kg/sq.m figure given in Figure 9 from this famous declassified CIA report on the T-72.

A solid 1-meter cube of steel would have an areal density of 7,850 kg/sq.m, and we can use this to find the mass of the glacis array relative to steel in terms of thickness by finding the quotient of the two terms, with the areal density of the array being the dividend and the areal density of a pure steel cube as the divisor. In this case, the answer is 0.33325 meters, or 333.25mm, which is almost exactly the same as the claimed mass equivalence of 335mm. This is the relative mass of the array in terms of steel, not its effective thickness in terms of steel. Areal densities can be used with some degree of accuracy to predict the amount of protection offered by composite armour, but the accuracy of such predictions hinges entirely on the mass efficiency of the armour. To obtain its effective thickness in terms of steel, 333.25mm is multiplied by the mass efficiency factor of the armour, and in this case, the claimed mass efficiency against KE threats needs to be investigated.

On a side note, the areal density of the array can be used to confirm the 1.8 g/cc density figure of glass textolite deduced earlier. 100mm of steel gives a LOS thickness of 267mm, and 267mm of steel has an areal density of 2,096 kg/sq.m, so the areal density of the glass textolite interlayer must be 2,096 kg/sq.m subtracted from 2,616 kg/sq.m = 520 kg/sq.m. This is corroborated by the 520.3 kg/sq.m given in Figure 9 from the CIA report on the T-72. To find the density of the glass textolite, we only have to divide 520 kg/sq.m by the 280mm LOS thickness of the interlayer, and from that we obtain 1,857 kg/cu.m, or 1.857 g/cc, which is very close to the figure obtained from "Plastmassy v bronetankovoy tekhnike". On the whole, it can be seen that the data presented in the table is formatted quite oddly and none of it should be taken at face value.

It is common knowledge that low density materials like glass textolite offer less resistance to shaped charges than high density materials like steel, so on the surface, it does not make sense that substituting 35mm of glass textolite with 20mm of steel did not increase the protection level of the armour, so the explanation must lie in the intricacies of the interaction between particulated jets and low density fillers. When we consider the fact that the effectiveness of the low density glass textolite is increased due to the particulation of the shaped charge jet, 35mm glass textolite in this specific configuration is nominally equivalent to 20mm of steel. Not bad, considering that the density of glass textolite is only ~1.8 g/cc whereas the density of RHA is 7.85 g/cc.

Conversely, the heavier 20mm steel plate is more efficient in terms of thickness, but it is much less efficient in terms of mass, so the substitution of 35mm of glass textolite for 20mm of steel can only be to improve protection against APDS and APFSDS threats while maintaining the same level of protection from shaped charge threats. This gives us some context for how the array would behave. Part of the function of the heavy front plate is obviously to erode the penetrator, and it would have been enough for less powerful threats on its own. After all, the Obj. 432 relied almost entirely on a single 80mm plate for protection against KE threats. There would have been no serious issues with that simple configuration since the L28 and L36 series of APDS shells would not be capable of penetrating the 80mm plate at a 68 degree angle, and the 140mm-thick glass textolite layer would behave like a giant spall liner to absorb spall and residual fragments. However, if the penetrator is not completely stopped by the single heavy front plate or at least shattered upon exiting it, the glass textolite layer would not be up to the task of stopping it on its own. This was a legitimate concern if the armour was hit by an APFSDS shell with a tungsten alloy rod.

In order to appreciate the function of the heavy front plate under circumstances where it is fully breached, the L15A3 can be seen as a good example of a more sophisticated APDS shell, as it has a tungsten alloy core with a relatively high elongation. Thanks to this, the shell has very good performance on highly sloped armour plate compared to earlier APDS and the high yield strength of tungsten alloys (a W-Ni-Cu alloy in this case) limits the severity of disintegration after the core emerges from a target plate. The emphasis is on the word 'limit', as demonstrated in the picture below (full page originally shared on tankandafvnews). The L15A4 is believed to be the primary APDS round of the Chieftain, but it is the same as the L15A3 for all intents and purposes because nothing related to its terminal ballistic performance was changed.

The graph on the left shows that the penetration of the APDS shell on a steel target at a 68 degree obliquity is around 110mm at 1000 yards (914 m) and around 100mm at 2000 yards (1,828 m), so it is obvious that the 80mm front plate of the upper glacis array is not enough to stop the L15A3 on its own out to 2 km and more, but it is definitely thick enough that the tungsten alloy core is shattered as it exits. This is exemplified in the photo on the right below.

It is not the best example, of course, because the thickness of the defeated plate shown in the picture was at the limit of the capabilities of the shell at that range and at that angle, so it is not quite as intact as it would be if it was facing the same plate at a shorter range, but the same concept applies: the asymmetry of forces acting on the plate due to the different relative thickness of metal above and below the penetrator cause the part of the plate below the penetrator to buckle, resulting in the early structural failure of the plate compared to a vertical plate. In parallel to this, the penetrator also experiences asymmetrical stresses as it penetrates the oblique plate, causing it to fracture into pieces inside the plate and to break apart as it exits due to the sudden release of the built-up stresses. This phenomenon becomes more pronounced at higher obliquity because the asymmetry of forces increases with the angle of the plate. The defeated plate in the photo on the right of the picture above was angled at 60° and you can already see how large the exit channel is. The photo below also illustrates how fragmented the tungsten core becomes after passing through a thick steel plate. As you can see, the core is shattered into dozens of smaller fragments, although many fragments of the steel plate itself undoubtedly contributed to the total amount of damage.

It's worth noting that the distribution of plate thicknesses in the two steel layers of the composite armour array is directly opposite from common simple spaced armour arrangements, which typically consist of a thin hard steel plate in front of a thicker but softer base armour plate. Such arrangements worked well on solid steel shells and APCR shells during tests conducted during and after WWII, but the effectiveness of such measures declined if attacked with capped shells including capped APCR and APDS shells, although they still performed better than homogeneous armour of the same mass.

On the other hand, the tungsten alloy armour piercing cap or tilting cap on advanced APDS rounds like the L15A3 would not contribute to improved performance on the composite armour of the T-72 itself, even though it contributes to better penetration on oblique plate targets in general. This is thanks to the heavy front plate of the 80-105-20 array which deflects and erodes the cap, forcing the tungsten core to penetrate the plate itself so that it loses much of its velocity and becomes fractured in the process. The low velocity fragments exiting the plate must then penetrate the glass textolite interlayer, and the non-optimal shape of the individual fragments of the core makes it much easier for the glass textolite to stop them as compared to an intact core with a pointed tip. If the armour array had a hypothetical 20-105-80 arrangement instead, only the cap will be destroyed during the interaction with the hypothetical thin front plate while the intact core will penetrate the thick interlayer. In doing so, it would not experience strong internal stresses so it does not fracture, nor does it lose any significant amount of its velocity. The naked core will have poorer performance on the high slope of the hypothetical thick back plate due to the lack of an armour piercing cap, but it is still a formidable threat as it retains almost all of its velocity and is fully intact. This can be combated by having a high thickness spaced plate in front of the base armour, as hinted by a conference paper titled "Spaced Armor" on page 4.

So a thin spaced plate in front of a thick base armour plate would be highly effective against APDS shells with unprotected tungsten carbide cores lacking an armour piercing cap. APDS shells of this description include the MK.3 shot for the British 20 pdr. gun, but more modern 105mm and 120mm APDS shells were made with tungsten alloy armour piercing caps to partly rectify this weakness, so generally speaking, a hypothetical 20-105-80 armour configuration would perform more poorly than the existing 80-105-20 array. This is confirmed by the following paragraph from the same document:

Indeed, this all but confirms that having a thick front plate and a thin back plate for the composite armour of the T-72 is a more reasonable design than its opposite as it has the effect of rendering the armour piercing or tilting cap of APDS rounds redundant. As the photos below show, the cap of an L28A1 APDS shell is fully destroyed during the initial penetration of a thick steel target, leaving a crater of large diameter while the tungsten core continues to penetrate efficiently.

The disintegration of tungsten core penetrators would be more pronounced at higher obliquity, so the 68° angle of the upper glacis is highly effective at defeating APDS shells. If the target plate is monolithic, the many fragments created by the breakup of the tungsten alloy penetrator is hugely beneficial as it greatly increases the post-penetration lethality of the shell. On the other hand, the same phenomenon would be hugely disadvantageous against the oblique composite armour of the T-72, as the penetrator would successfully perforate the heavy front plate but the broken pieces (and also the fragments of the armour plate) will have to penetrate the glass textolite layer and may fail to defeat the back plate due to the bad shape of the individual core fragments, especially since the residual velocity of the fragments exiting the heavy front plate will be reduced. Of course, it should not be forgotten that an anti-radiation lining is present behind the 20mm steel back plate of the upper glacis array, so the likelihood of these penetrator fragments going through the last layers of the array is very low.

The 305mm figure for the armour shown earlier point towards this conclusion as well. L15A3 has a penetration of 382mm at the muzzle at 0 degrees and 355mm at 1000 yards at 0 degrees, indicating that it would perforate the armour with ease at 914 meters if the given protection values were taken literally. However, we have already established that the 305mm figure needs to be processed further due to the format of the presentation. When we multiply 305mm by the cosine of 68 degrees, we find that 305mm RHA vs KE is equal to 114.3mm RHA at 68 degrees, and there does not seem to be anything obviously wrong with this at first glance - the cumulative thickness of steel in the array is 100mm, and 105mm of glass textolite is ostensibly equivalent to 14.3mm of RHA which is not very high. However, the penetration graph shown earlier states that L15A3 can defeat only 110mm of steel at a 68 degree obliquity at a distance of 1000 yards (914 m), so L15A3 should not be able to defeat the armour of the T-72 Ural at 914 meters. Suddenly, the same armour becomes immune to L15A3, and the importance of the slope of the armour becomes immediately clear. When we consider the lower performance of L15A3 and other APDS rounds at an impact angle of 68 degrees compared to 0 degrees, the 305mm figure is obviously misleading. It is much more likely that this figure is referring to APFSDS rounds and not APDS. However, other information indicates that the armour of the T-72 is functionally equivalent to a 130mm RHA plate sloped at 68 degrees for APFSDS rounds, which is 347mm RHA when normalized to 0 degrees. This is much closer to the 330mm figure stated by the initial requirements for the T-64A and is much more likely to be true since the initial requirements represent a prerequisite level of protection, so the armour cannot be less than 330mm RHA but can be more.

Indeed, it appears that these low figures may not necessarily apply to certain types of APFSDS rounds. Referring to page 16 of this CIA intelligence report on Soviet tanks, it appears that the M735 APFSDS round was considered to be incapable of defeating the armour of the T-72 except on the thin monolithic lower glacis armour based on the stated probability of kill per hit of 0.22 for both the lower and upper bound estimates of the armour. These results were based on test firings conducted by the Ballistics Research Laboratory.

The lack of any change in the probability of kill for both estimates indicates that the M735 round cannot defeat the turret and upper hull armour under any circumstances, but can defeat weaker zones such as the lower glacis which remains at a constant thickness regardless of any changes in the upper glacis and turret armour of the tank. The M735A1 round with a DU core is claimed by Janes to penetrate 370mm of RHA steel at 0 degrees at 1 km, so the M735 round with a tungsten core and identical ballistic properties should penetrate around 10% less under the same conditions, or around 330mm RHA. Therefore, the armour of the T-72 must be worth more than this amount even for a lower end estimate. This does not contradict the 330mm RHA figure provided by our Russian sources because the M735 round uses an unusual teardrop-shaped tungsten alloy penetrator with a round tip and a large diameter. This type of penetrator does not perform well at highly oblique armour and does not perform well against composite armour in general, so the armour may be worth more than 330mm against M735. Unsurprisingly, the M774 round with a long rod depleted uranium penetrator is rated much more favourably than the M735, having a probability of kill per hit of 0.71 for a lower bound estimate and 0.50 for a higher bound estimate. The 21 percentage point difference is difficult to interpret because the table is not stating the probability of penetration, but the probability of kill on the first hit. This includes a multitude of factors including the probability of hit and the post-penetration effects in the event of armour perforation. As such, the estimations for M774 will not be taken into consideration. Overall, these estimations appear to support the notion that the T-72 upper glacis armour cannot be less than 330mm RHA but can be more.

To find the virtual equivalent relative thickness of the armour array against a threat like the L15A3, we must divide the penetration of the shell at 0 degrees (355mm) with the penetration at 68 degrees (294mm LOS) to obtain the slope efficiency coefficient, which is around 0.83. When 305mm is multiplied by the reciprocal of 0.83, we find that the effective thickness of the armour against APDS is 367mm RHA, normalized for 0 degrees. The mass efficiency of the array increases from the miserly 0.9 claimed in the Russian table to a much more believable 1.1, rounded up from 1.095. The 355mm penetration of L15A3 at 1000 yards is insufficient and would fail against this target. When we use the 330mm figure instead, the protection jumps up to an excellent value of 397mm RHA at 0 degrees in relative thickness against APDS and the mass efficiency increases to 1.18. These numbers may seem far too high, but the original 305mm or 330mm figures cannot possibly represent the protection of the armour against APDS, because the 100mm of steel in the array already has a LOS effective thickness of 321.6mm after taking the slope coefficient of 0.83 into account. The effective thickness of the entire package must be higher than that by at least some amount.

This is an excellent lesson in caution, because practically all protection figures for tanks fitted with composite armour are formatted as a fixed number normalized for a 0 degree impact unless specifically stated to be otherwise. The reason is obvious upon further inspection. For instance, the NERA panels in the armour of the M1 Abrams are arranged at a different obliquity than the front and back plates encasing it, so an attacking projectile will have to first perforate the thin front plate placed at low obliquity, then the NERA panels placed at high obliquity, and then finally the thick back plate placed at low obliquity. How should the equivalent protection figure be expressed? If it is expressed as line of sight thickness from a particular angle, then which angle is most appropriate? The angle of the steel front and back plates or the angle of the internal NERA panels? Since the numbers attributed to the T-64A upper glacis armour are too low to be for APDS projectiles, they must have been normalized to be equivalent to a steel plate at 0 degrees against long rod projectiles for the sake of simplicity. The only question is what type of long rod projectile they had in mind.

It is interesting to see that page 3 of the brochure states that the L15A3 shell is "the first high velocity shot of its type which effectively defeats multiple targets", which may be referring to the NATO Heavy Triple target. It also probably hinting that it was different from previous APDS designs and that previous designs would have performed worse against spaced armour or perhaps even composite armour. This is most likely referring to the use of a tungsten alloy core, as older APDS shells like the L28 and L36 have all the features of the L15 including the shape of the core and armour piercing cap, only differing in having a tungsten carbide core. However, it appears that the 105mm L52 was the first APDS round created for this purpose, not the L15. Indeed the L15 projectile is observed to be almost identical to L52 except for the scale.

The diagram below shows the L15A5, which is structurally similar to the L15A3 and differs only in the alloy of the core. Note the sharp-tipped conical steel tilting cap, labelled "nose pad", on top of the hemispherical nose of the tungsten carbide core.

From what we now understand, all the evidence indicates that the heavily sloped upper glacis array of the T-72 Ural should be quite adept at defeating all  contemporary APDS projectiles up to the 120mm caliber, which were still credible threats even while new APFSDS was being developed on the other side of the Iron Curtain. The primary threat for some time was the Chieftain, which relied on the powerful L15 series of APDS shells that had some chance of penetrating the upper glacis at close range, but the American M60A1 rapidly overtook the Chieftain due to the emergence of M735 composite APFSDS ammunition. Other 105mm-armed tanks would not be considered nearly as dangerous because APDS was the only option besides HEAT, and neither would be enough.

Furthermore, blast attenuation is an aspect often overlooked when referring to tank armour. The T-72 has an advantage over tanks with monolithic steel armour due to the use of a composite construction. By placing multiple layers of multiple materials of drastically different properties in the path of the blast waves, the effectiveness of the array in attenuating explosions is significantly improved as compared to homogeneous materials of the same weight. This was quite important seeing as HESH (High-Explosive Squash Head) shells were and still are a British favourite. Indeed, the primary impetus for the development of the L11 120mm rifled gun for the Chieftain was because a larger HESH shell could be fired from it. Even if HESH were somehow effective against the armour of the T-72, the spall could be absorbed by the thick anti-neutron lining, which also provides blast attenuation along with the layered hull armour array.

Against APFSDS ammunition, the functions of the individual components of the armour array are similar as for APDS rounds but with additional nuances. The most important component is the heavy 80mm front plate and its ability to deflect long rod penetrators. This topic is not often discussed yet carries crucial importance pertaining to long rod penetrator performance on spaced and composite armour designs. For one, the complexity of the issue regarding the performance of certain nose shapes versus certain armour arrangements makes it almost impossible to create a standardized model for long rod performance against complex armour. Multiple models need to be combined to simulate specific penetrator and armour designs and the combined model needs to be verified with experimental results, making it a laborious process that is only truly useful on a case by case basis. This is one of the reasons why protection equivalency figures in RHA are often misleading, as there often are many differences between APFSDS penetrator designs that may improve or degrade its performance against specific types of composite armour, and this is also the reason why the RHA figures published in this article should not be taken at face value.

However, one fact is known for sure, and that is that certain nose shapes perform better on semi-infinite plate at high obliquity and certain nose shapes perform worse. For instance, it is known that DU long rod penetrators have significantly better performance compared to tungsten alloy penetrators across the board (up to 10% better for alloys of the same density), but the same guidelines for penetrator design apply for both types. Conical noses, for example, are rarely found on service ammunition partly because conical-nose rods yield the best results on perpendicular plate but perform very poorly at high obliquity. However, the clincher is that it does not offer nearly enough improvement on perpendicular plates to justify the huge losses in performance on oblique plates, not to mention that even flat targets will often be attacked at some angle during real tank battles as very few hits are sustained from a 0 degree angle of attack. All taken together, it is easy to understand why conical noses are never found on any long rod penetrator. Blunt noses are the most popular as they offer the best performance on highly oblique plates with only slightly reduced performance on perpendicular plates, which is largely irrelevant for modern tank armour design anyway. Frustrum-nosed rods offer a compromise between conical and blunt nosed rods, and for this reason it has found some usage, although it is not nearly as widespread as blunt noses.

This was examined in great detail in the study "Effect of Nose Shape on Depleted Uranium (DU) Long-Rod Penetrators" by W. Leonard. This study is particularly relevant for U.S-made depleted uranium long rod penetrators from the 70's and 80's as the material used for the DU test rod is the same U-3/4 Ti alloy as used in the M774 and M833. The DU test rod is not a replica of the two rods since the aspect ratio is less - 10 vs 14.3 and 18 respectively - but this does not affect the nature of the effect of nose shape.

As the tables below show, the effects of nose shape are consistent for both depleted uranium and tungsten alloy rods. For a 1" thick RHA plate sloped at 70.5 degrees, the velocity limit for a DU rod with a blunt tip is 1,088 m/s and the velocity limit for a DU rod with a conical tip is 1,355 m/s. This is a sizable difference of 267 m/s or 24.5%, meaning that a conical rod of the same mass and aspect ratio would need to have 24.5% higher impact velocity to defeat the same target. For a 3" thick RHA plate placed perpendicularly, the velocity limit for a DU rod with a blunt tip is 1,373 m/s and the velocity limit for a DU rod with a conical tip is 1,239 m/s. The difference is only 134 m/s or 9.76%, so in other words, the conical rod has an extremely minor advantage compared to the blunt rod.

The same pattern is observed for tungsten alloy rods. For a 1" thick RHA plate sloped at 70.5 degrees, the velocity limit for a tungsten alloy rod with a blunt tip is 1,186 m/s and the velocity limit for a tungsten alloy rod with a conical tip is 1,470 m/s. The difference is 284 m/s or 24%, which is very significant. For a 3" thick RHA plate placed perpendicularly, thvelocity limit for a tungsten alloy rod with a blunt tip is 1,440 m/s and the velocity limit for a tungsten alloy rod with a conical tip is 1,333 m/s. The difference is only 107 m/s or just 7.4%. As you may have noticed, the velocity limit of the tungsten alloy penetrators of all three nose types are higher than the depleted uranium penetrators of the same nose types for all angles, translating to lower penetration values at the same impact velocities.

Although nose shapes other than these three were not tested, it was mentioned that the studies by John Zooks - "The Penetration Performance of Tungsten Alloy L/D=10 Long Rods With Different Nose Shapes Fired At Rolled Homogenous Armor" and "Comparative Penetration Performance of Tungsten Alloy L/D=10 Long Rods With Different Nose Shapes Fired at Rolled Homogeneous Armor" - using short frustum, hemispherical, and blunt nose shapes showed that tungsten alloy rods performed worse than a conical nose rod on a perpendicular impact but better on an oblique impact. 
In the conclusion of the paper, it is asserted that cone-nosed penetrators displayed an increased tendency of ricocheting off the face of oblique plates instead of digging into the target. The independence of nose shaped design from penetrator material was also noted, and the insignificant value of higher performance on perpendicular plate in real world conditions is also made clear.

This means that the findings by Zooks et al. with tungsten alloy penetrators can be directly applied to DU rods as well. Additionally, the study 
"On The Role of Nose Profile In Long-Rod Penetration" by Z. Rosenburg and E. Dekel shows that these results are also applicable to long rod penetrators of a wide variety of different metals with wildly differing mechanical properties including titanium, aluminium and copper. Titanium is known to be a good representative of depleted uranium as it fails by adiabatic shearing under high strain rates, thus simulating its penetration properties. Steel targets were used for the tungsten rods and aluminium targets were used for the titanium and copper rods, simulating a roughly similar ratio of penetrator to target strengths. The final penetration depth for the five nose types for all metals is presented in the table on the right, normalized in terms of penetrator rod length.

From the simulation results gathered from the study, it can be seen once again that conical nosed rods perform the best on perpendicular impacts compared to blunt nosed rods, and the difference between the conical nose rod and the T-shaped nose rod was by a factor of 3. The performance of the hemispherical nose rod lay between the conical nose rod and the blunt nose rod as predicted earlier using the information from Zooks et al., so the performance of hemispherical nose penetrators such as the American M735 and M774 shall be less efficient compared to a blunt-nosed penetrator on a highly oblique plate such as the 68 degree heavy front plate of the T-72 array but more efficient than a conical nose penetration. The design shifted to an ogive nose shape in the M833 depleted uranium long rod penetrator, but the official reason is not yet known. According to various studies on the subject, an ogive can be considered the most optimal nose shape because it is conducive towards greater armour penetration at a perpendicular angle compared to a conical nose, but it does not possess the same weakness as a conical nose on high obliquity impacts. However, the performance of ogived nose rods on oblique targets is still significantly poorer than blunt nose rods and it has also been shown that the poor performance of this nose shape is apparent at a medium obliquity. From the results of the study "Penetration of 6061-T6511 aluminum targets by ogive-nosed VAR 4340 steel projectiles at oblique angles: experiments and simulations" by T. Warren et al., it can be seen that a steel long rod penetrator with an ogive nose experiences a large rotational moment as it penetrates the target, and the rotational forces are large enough in the case of a 45 degree impact that the rod is bent in the direction parallel to the surface of the target and its penetration path forms a curve towards the free surface of the plate - towards the region of lower material resistance.

For practical tank armour applications, these results indicate that the large bending forces experienced by a long rod penetrator with an ogive nose on a highly oblique plate will likely lead to the ricochet of the tip and a subsequent decrease in penetrative efficiency. All of this hints that M735, M774 and M833 may have reduced effectiveness on the upper glacis of the T-72 as a result of less-than-optimal tip design for high obliquity plates.

There have been a variety of unique solutions devised by munitions engineers to combine the benefits of conical and blunt nosed penetrators. An early solution was the use of rounded noses on the American M735 tungsten alloy and M774 depleted uranium long rod penetrators, but this was a compromise solution and did not give the penetrators the best possible performance on both perpendicular and oblique plates. The L27 "Charm-3" APFSDS round features a more recent solution which was to install a conical tip segment on top of a blunt nosed monobloc rod, connected via a weak joint. In the case of an impact with a target at a perpendicular angle, the conical tip yields better penetrative power and the mass of the tip contributes to the penetration depth. If the projectile impacts an oblique plate, the conical tip segment readily ricochets off the plate and is detached from the rest of the rod at the weakened joint, leaving the base rod with its blunt tip to dig into the target. This enables the penetrator to have optimal performance on both perpendicular and oblique targets with minimal sacrifices made to projectile design.

However, Charm-3 may be a little too recent to be considered a contemporary threat to the T-72 Ural, considering that the latter was only produced from 1973 to 1975. The armour array was changed in 1976 as part of a modernization programme.


In 1976, a new glacis array was introduced for the T-72 Ural-1 modernization. The new array retained the 105mm glass textolite filler, but it now had a 60mm RHA front plate and a 50mm backing plate instead. The total thickness becomes 574mm when angled and the mass of the array in terms of steel was increased to 362mm. The areal density increased from 2,616 kg/sq.m to 2,826 kg/sq.m. The vast majority of the Red Army's T-72 tanks incorporated this newer armour scheme, and this is instantly obvious when we examine the production record of the T-72; the production volume at Nizhny Tagil in 1976 alone was 1017 units, whereas only 950 units were released during the entire production run of the original T-72 Ural model from 1973 to 1975. This new 60-105-50 array was also the standard configuration for most if not all T-72M tanks.

A multitude of different array layouts with different ratios of layer thicknesses were tested against tungsten alloy long rod penetrators of differing aspect ratios.

The array shown below is an array design 1.2:2.12:1.0 ratio of layer thicknesses with steel front and back plates with a glass textolite interlayer, equivalent to the 60-105-50 design of the Ural-1.

The redistribution of thicknesses of the steel plates in the Ural-1 variant was done to improve the mass efficiency of the array, but from a technical standpoint it might be for two possible reasons; a tendency of the thin 20mm back plate of the original array to buckle or bulge excessively when impacted by the remnants of an APDS or APFSDS shell, and possibly its inability to reliably absorb the fractured or otherwise degraded penetrator of new long rod APFSDS projectiles due to the low efficiency of thin plates.

According to page 246 of "Particular Questions of Terminal Ballistics" 2006 (Частные Вопросы Конечной Баллистики) published by Bauman Moscow State Technical University on behalf of NII Stali, the optimum thickness of the front plate of a spaced or composite armour array against a long rod APFSDS penetrator where maximum destructive effects are achieved is equal to 1 to 2 diameters of the penetrator rod for a plate angle of more than 60 degrees. The relevant paragraph is shown below.

"Полученные результаты позволяют утверждать, что практически независимо от удлинения сердечника наиболее эффективное максимально деструктивное и дестабилизирующее воздействие на сердечник при α = 60° и более оказывают экраны с горизонтальной толщиной в 1-2 калибра сердечника. Это воздействие в максимальной степени реализуется на удалении переднего торца сердечника от тыльной поверхности пробитого экрана не менее чем на 0.5 длины остатка сердечника. Если анализировать красную) кривую, то видно, что при отсутствии или малых величинах воздушных зазоров, т ее. В диапазоне, где эффекты, связанные с изгибом или разрушением сердечников, еще не успевают проявиться в достаточно полной мере, всегда имеет место проигрыш в стойкости до 10...12% по сравнению с РМБ."


"The obtained results make it possible to assert that practically independent of the core elongation, the most effective destructive and destabilizing effect on the core at α = 60 ° and more is provided by screens with a horizontal thickness of 1-2 diameters of the core. This effect is maximally realized by removing the front end of the core from the back surface of the punched screen by at least 0.5 lengths of the core remainder. If we analyze the red curve, we see that in the absence or small values ​​of air gaps, that is, in a range where the effects associated with bending or destroying the cores do not yet have time to manifest themselves to a sufficient extent, there is always a loss in durability up to 10 ... 12% in comparison with monolithic homogeneous armour."

From this, it appears that the 60-105-50 array was designed to be optimized from contemporary
APFSDS rounds unlike the earlier 80-105-20 array which was primarily intended to defeat APDS. It is not a coincidence that the thickness of the 60mm heavy front plate of the array is approximately 1-2 times the diameter of APFSDS shells fielded by NATO militaries from the late 70's to the late 80's. For instance, the American M774 and M833 rounds for the 105mm M68 have a penetrator diameter of 26mm and 24mm respectively. The British L23 and L23A1 rounds for the 120mm L11A5 have a penetrator diameter of 29mm, and the German DM23 round for the 105mm L7A3 has a penetrator diameter of 26mm. The increased elongation of later APFSDS rounds do not make a difference in changing the effectiveness of the heavy front plate, so the increasing elongation of APFSDS rounds during the 80's and 90's does not change the fact that 60mm remained the optimum thicknesses. This is rather important since the T-72B continued to make use of a 60mm heavy front plate in all of its three hull armour designs, and the T-90 continued to make use of the same front plate thickness as it inherited the T-72B obr. 1989 hull and hull armour.

It is possible that the front steel plate was made slightly harder to further improve its protective characteristics and because it is thinner and therefore easier to treat to a higher hardness than the earlier 80mm plate. However, there is no evidence for or against this notion, so it is pure speculation. The stimulus behind the change to a new array is unknown, but there is a possibility that it may have been influenced by some yet-unpublished analysis of enemy anti-tank weapons from the 1973 Yom Kippur war. It could also be due to some new projections and predictions regarding APFSDS developments in the West or new intelligence regarding the soon-to-arrive M735 APFSDS (1978). Whatever the reason was, the implementation of the new array was purely beneficial. For comparison, the T-64B (1976) continued to use the older 80-105-20 array well into the mid-80's, but the protection was not significantly inferior to the redesigned Ural-1 array because high hardness BTK-1 steel was used as a replacement for the medium hardness steel of the T-72. Similarly, the T-80 used the older design but incorporated BTK-1, and the T-80B featured a slightly thinner 60-100-45 array that had comparable or slightly better protection as it used BTK-1. According to Andrei Tarasenko, the upper glacis of the T-64B is equivalent to 360mm RHA against APFSDS.

Since the thicknesses of the front and back plates of the array have changed, there is no doubt that the nature of the interactions between the armour array and the threats against it have also changed in some way. A 60mm plate will not be as effective as an 80mm plate at particulating a shaped charge jet, so a more continuous jet will penetrate the glass textolite filler. As we have already established, the low density glass textolite filler is less optimal against a continuous jet, so it may be that not only was the thicker back plate intended to absorb the rest of the jet, but the back plate had to 30mm thicker than the original 20mm plate in order to make up for the reduced effectiveness of the interlayer. This can be viewed as a compromise to improve protection against KE threats while keeping the protection against SC threats at the same level, with the penalty of a reduction in mass efficiency. From this perspective, the additional 10mm of steel should not be regarded as additional armour against shaped charges, but as compensation for the reduced front plate thickness so the mass efficiency should remain the same at 1.35. Therefore, any attempts to add 27mm of armour equivalence (10mm / cos 68°) to the previously given 450mm RHA figure would be fundamentally invalid. A decidedly more reasonable approach is to divide the relative mass of the array (362mm) by the reciprocal of 1.35. This gives us 490mm RHA, which is supported in Kampfpanzer: Heute und Morgen by German author and military expert Rolf Hilmes who also attributes the T-72M with an armour protection level of 490mm RHA against shaped charges. According to page 159 of "Боевые Машины Уралвагонзавода: Танк Т-72" published by the Uralvagonzavod Production Association, the hull armour of the T-72A (the same as the Ural-1) is equal to 500mm RHA against shaped charges.

The effectiveness of the new array against KE threats is not known, but if we apply the same 1.1 to 1.18 mass efficiency coefficient against advanced APDS and 0.985 coefficient against APFSDS of the original array, then the effective thickness of the new 60-105-50 array would be worth around 399mm to 427mm RHA against APDS and around 357mm RHA against APFSDS. However, the point of introducing a new array would be to improve the mass efficiency against APFSDS, otherwise the same increase in effective thickness could be achieved by adding a much simpler and cheaper 10mm appliqué armour plate on top of the original array, but the problem is that it is difficult to estimate how much the mass efficiency factor changed. According to Rolf Hilmes, the upper glacis armour of an ex-East German T-72M provided a protection level of 400mm against APFSDS rounds (and 490mm against HEAT rounds as mentioned before), implying a mass efficiency of 1.1 against APFSDS. These numbers are in good agreement with our prior assumptions. If the 80-105-20 array has a mass efficiency of 0.985, then the 60-105-50 array should have a higher ME, and a 1.1 coefficient represents a perfectly reasonable increase of only 11.6%. According to "Боевые Машины Уралвагонзавода: Танк Т-72", however, the upper glacis armour of the T-72A is equivalent to 360mm RHA against APFSDS threats.

The exactness of our previous 399mm estimation with Hilmes' 400mm vs APFSDS claim is purely coincidental because 399mm was obtained using the slope efficiency factor of APDS and does not apply for long rod APFSDS rounds, and the newer 60-105-50 array may have a higher efficiency against APDS anyway.

We can corroborate the 400mm figure using the table below.

The key lies in the 400mm figure mentioned in the margins and not in the equivocation of the T-72 armour with a 130mm RHA plate. The requirement to penetrate 400mm RHA at 3.5 km is not a coincidence, nor is the fact that the official maximum effective range of the round is reported to be 3.5 km. The 400mm figure refers to the effective thickness of the 60-105-50 array against APFSDS which we have established earlier to be equivalent to 400mm RHA in line of sight thickness, the penetration figures listed for the L23A1 table have nothing to do with a 130mm plate of any kind, but are the LOS penetration of the round for a semi-infinite plate at 68 degrees. Indeed, inputting the known dimensions of the penetrator (29x410mm with 1x50mm frustrum) and the density of the W-Ni-Fe alloy (18.6 g/cc) into the Lanz-Odermatt equation with the target set at 68 degrees, we find that there is a perfect match between the calculations and the reported experimental results. Case in point, the calculations indicate that the penetration at the muzzle is 482.7mm (compared to 482mm actual penetration), 460.4mm at 1.0 km (459mm), 437.6mm at 2.0 km (435mm), 413.2mm at 3.0 km (412mm), and 400.8mm at 3.5 km (401mm). The average error is practically nil - only 0.0027%.

Of course, this is assuming that the L23A1 has a conical tip and not a hollow-point tip or a two-piece tip as this drawing appears to show. Even if the rod does not have a conical tip as assumed, this interpretation of the data appears to make the most sense considering that the table is definitely listing the penetration of the round into semi-infinite plate.

With this interpretation of the data, the 400mm RHA figure supplied by Hilmes is somewhat validated and finding out the range limit of L23A1 against other armour arrrays is almost trivial. The round will cut straight through the hull of the T-72 Ural (305-330mm) at 5-6 kilometers and defeat the hull of the T-72 Ural-1 and T-72A (400mm) at 3.5 km or less. This seems imply that the T-72 is hopeless whenever it encounters a Challenger 1/2 or a Chieftain Mk.5/2, but it is important to note that the original L23 base model entered service in 1985 and the L23A1 followed suit soon after during the late 80's, so the L23A1 came only a decade after the T-72 Ural-1 was fielded. By this time, the superior T-72B had already entered service and the earlier models had already been uparmoured with an additional 16mm armour plate.


In 1983, an additional 16mm high hardness plate was added to the upper glacis armour by welding, which came about as a result of live fire testing of captured Israeli M111 tungsten-cored shells from Lebanon (in the 1982 war in Lebanon). Contrary to popular belief, the Israelis did not "discover" that their M111 Hetz could perforate the T-72 from the front "at about 650 meters". The Israelis never got their hands on an intact T-72, nor did they ever face them with 105mm guns in combat. Strong evidence has indicated that at best, the Syrian T-72s were destroyed in an ambush by TOW missiles fired at their flanks from gunships.

However, it is true that the M111 "Hetz" was acquired by the Soviet Union and extensively examined and tested. A very popular theory is that these rounds came with the captured Israeli M48A3 that was until recently on display in Kubinka. The original American M48A3 doesn't have a 105mm gun, of course, but Israelis had a habit of upgrading their tanks. Having captured M111 Hetz rounds in some quantities, it was discovered by the Soviets that the upper glacis of the T-72 was vulnerable at short ranges (the exact range was never publicized), thus necessitating the installation of the appliqué armour plate. As the appliqué plate is only 16mm thick, the boost in armour protection is not particularly high, but it increased the mass of the array to the equivalent of 371.7mm of RHA. It is known that the T-64A, T-64B and T-80, T-80B also received weld-on appliqué armour at around the same time as the revelation, but all of these variants received a 30mm plate rather than a 16mm plate. The reason for this is not that the T-64A/B and T-80/B were more valuable assets than the T-72 and deserved better armour, as some may assume based on common perception. Rather, it was because the vast majority of T-72 tanks at the time were already using the 60-105-50 hull glacis armour scheme while the others were still built with the older 80-105-20 armour array. As it had a similar but slightly thinner new armour scheme (60-100-45), the T-80B received a 16mm appliqué armour plate in 1983 like the T-72, whereas the original T-80 received a 20mm plate in 1979 as part of an unrelated modernization effort to increase its protection to the same level as the T-80B. The T-64A and T-64B both received a 30mm appliqué armour plate in 1984-1985 during a modernization effort to improve the armour up to the level of the new T-64BV, which had a completely new 5-layer upper glacis array.

The value of the 16mm plate against KE threats is difficult to quantify, and it is certainly inaccurate to simply add the LOS thickness of the plate (42.7mm) to the known effective thickness of the array. The fact that a tungsten long rod penetrator would penetrate around 1.24 times more armour plate at a 68 degree obliquity is not accounted for, nor are the differences between HHS and RHA, or the positive effects of the obliquity of the plate, or the advantages of layering a high hardness plate over a lower hardness plate to create a DHA (Dual-Hardness Armour) arrangement. High hardness steel is best used as appliqué armour, like in this case. The higher hardness and strength of HHS yields the best results for defeating KE threats especially at a very high obliquity, but very high hardness steels tend to be difficult to weld, so it is more than likely that the appliqué plate has a hardness in the lower spectrum of hardnesses for vehicle armour-grade HHS as listed in MIL-DTL-46100E, so it may be around 450-500 BHN. Regardless, the lack of confirmation on the composition of the plate combined with a multitude of other factors make it much more difficult to find out the true contribution of the appliqué plate, so no attempt will be made to perform any calculations. Instead, we must rely on secondary sources.

The November issue of the famous Russian Tekhnika i Vooruzhenie 2006 (Журнал Техника и Вооружение) magazine mentions in page 14 that in 1993, a report published in the specialized magazine "German Airspace" by A. Mann states that the armour of the T-72M1 exhibited protection equivalent to 420-480mm of rolled homogeneous armour when tested against modern 105mm and 120mm ammunition from West Germany. The upper glacis armour of the T-72M1 is the same as the 1976 modification of the T-72 Ural plus the 16mm appliqué armour plate (16-60-105-50). For all intents and purposes, "420-480mm" can be interpreted to mean that the hull armour is equivalent to 420mm RHA against modern long rod tungsten alloy ammunition like the 105mm DM33 (standard 105mm APFSDS in the late 80's and the decade after), while the turret is equivalent to around 480mm RHA, probably at the cheeks. The use of modern APFSDS rounds instead of earlier types like the M735, M774, 105mm DM23, or the 120mm DM13 and DM23 is responsible for the reduced mass efficiency of the uparmoured array.

On the other hand, it is claimed on Andrei Tarasenko's website that the armour of the T-72A with the appliqué plate is equivalent to 405mm of steel against M111. The relevant paragraphs are cited verbatim here:

"Одним из выводов специальной комиссии по определению противоснарядной стойкости отечественных танков было то, что М111 имеет преимущества перед отечественными 125 мм снарядом БМ22 по дальности пробития под углом 68° комбинированной брони ВЛД серийных отечественных танков. Это дает основание полагать, что снаряд М111 отрабатывался преимущественно для поражения ВЛД танка Т72 с учетом особенностей ее конструкции, в то время как снаряд БМ22 отрабатывался по монолитной броне под углом 60 градусов.

В ответ на это по завершении ОКР «Отражение» на танки вышеуказанных типов в ходе капитального ремонта на ремзаводах МО СССР на танках с 1984 года осуществлялось дополнительное усиление верхней лобовой детали. В частности на Т-72А устанавливалась дополнительная плита толщиной 16 мм, что обеспечивало эквивалентную стойкость 405 мм от ОБПС М111 при скорости предела кондиционного поражения 1428 м/с."

This translates to:

"One of the conclusions of the special commission for determining the anti-ballistic stability of domestic tanks was that M111 has advantages over domestic 125 mm with a BM22 projectile at the penetration range at an angle of 68° Combined armor of VLD serial domestic tanks. This gives grounds to believe that the M111 projectile was developed primarily for the destruction of the VDL of the T72 tank, taking into account the features of its design, while the BM22 projectile was tested on a monolithic armor at an angle of 60 degrees.

In response to this, after the completion of the ODF "Reflection" for tanks of the above types during the overhaul at the Soviet factory of the Ministry of Defense of the USSR, on tanks since 1984 an additional reinforcement of the upper frontal part was carried out. In particular, the T-72A was equipped with an additional plate thickness 16 mm, which provided equivalent durability 405 mm from the OB11 M111 at the rate of the limit of the conditioning defeat of 1428 m/s."

Based on the known velocity drop of 105mm APFSDS rounds, the velocity of 1428 m/s roughly corresponds to a distance of 500 meters given that the muzzle velocity of M111 is 1455 m/s, so the armour is immune to M111 up to 500 meters. Not coincidentally, this is the same distance at which the original T-64 was designed to be immune against 105mm APDS.

Now, we are left with two distinct figures for the same 16-60-105-50 armour configuration: 420mm by Mann and 405mm by Tarasenko. However, this discrepancy can be explained by the unique stepped tip of the M111 penetrator (refer to the drawing earlier) having superior performance at high angles of impact whereas standard monobloc rounds from West Germany like the DM33 have blunt tips that presumably perform worse than the M111. This adequately explains the differences in the claimed armour values. If our previous interpretation of the L23A1 firing table was correct, then L23A1 should behave much like "modern ammunition from West Germany" so the 420mm RHA figure may apply. In this case, L23A1 would be able to defeat the up-armoured upper hull array at a distance of between 2.0 to 3.0 km. Not so good considering that combat distances are still much shorter than that, but at least it has some effect in increasing the survivability of the tank. The mass efficiency of the armour did not increase, as the upgraded array now has an areal density of 3,144 kg/sq.m and has the relative weight of 400mm of steel, but has a maximum effective thickness of 420mm RHA. The mass efficiency is therefore 1.05, or just 5% better than a homogeneous steel plate. Of course, a homogeneous steel plate would not be worth its own weight at an angle of 68 degrees since the penetrative performance of long rod penetrators is enhanced on highly sloped plates, but this is largely irrelevant in this context.

It is worth noting that M111 and DM23 were the most advanced 105mm APFSDS rounds of the early 80's and were the best anti-tank ammunition available for the Leopard 1A4 and 1A5, not to mention that they were practically standard for the remainder of the decade. The DM13 round was a licence-produced version of the L28A1. It must be understood that the up-armouring programme was merely a temporary stopgap measure to keep the Red Army's large fleet of T-72 Ural and T-72A tanks viable against common 105mm APFSDS threats for the next few years, whereas the emerging threat of 120mm guns required a much more serious upgrade in armour protection that took the form of the T-72B.

Against shaped charges, it is not possible to simply add the LOS thickness of the appliqué plate to the total effective thickness. The mass efficiency of the array may possibly increase due to the added thickness of the heavy front plate but it is too complex to quantify the changes, not least because the mechanical properties of the plate material must be known. The effectiveness of high hardness steel may not be higher than RHA because the penetration of shaped charges is not related to the yield strength, but to the ultimate tensile strength (UTS). A high hardness steel plate with a lower UTS would offer less resistance than an RHA plate with lower hardness but higher UTS. In the example below, Hardox 500 steel offers significantly less protection compared to RHA steel. Table taken from the study "Hypervelocity Penetration against Mechanical Properties of Target Materials" by K. Kamarudin et al.

If the appliqué armour plate is BTK-1Sh steel, estimations are complicated by the lack of information on the material properties. If the plate is BT-70Sh, the shaped charge resistance will be lower as the steel has a UTS of only 1700 MPa, lower than both RHA steel and Hardox 500. This is consistent with an article published on Andrei Tarasenko's website that increasing the hardness of the front plate of the three-layer sandwich leads to a decrease in the shaped charge resistance of the armour. The relevant claim, verbatim:

"Уровень трехпреградной (сталь + стеклотекстолит + сталь) броневой защиты танков Т-64А, Т-72А и Т-80Б обеспечивался подбором оптимальных толщин и твердости материалов лицевой и тыльной стальных преград. К примеру, повышение твердости стального лицевого слоя ведет к снижению противокумулятив­ной стойкости комбинированных преград, установленных под большими конструктивны­ми углами (68°). Это происходит вследствие снижения расхода кумулятивной струи на внедрение в лицевой слой и, следовательно, увеличения ее доли, участвующей в углублении кавер­ны."


"The level of the three-barrier (steel + glass textolite + steel) armored protection of the T-64A, T-72A and T-80B tanks was ensured by the selection of optimal thicknesses and hardness of the front and rear steel barriers. For example, increasing the hardness of the steel face layer leads to a decrease in the anti-cumulative resistance of the combined barriers installed at large structural angles (68 °). This is due to the reduction in the flow rate of the cumulative jet during penetration into the face layer and, consequently, to an increase in its length involved in the deepening of the cavity."

It should be noted that the paragraph is referring to the performance of the entire array as a whole system and not the front plate alone. Unlike the example given where the hardness of the entire heavy front plate of the steel-glass textolite-steel was changed, the 16mm appliqué plate is added on top of the existing 60-105-50 array so the optimal arrangement of thicknesses and hardness is not modified but augmented. This means that the mass efficiency of the array does not decrease as a result of adding the appliqué plate, but the additional effective thickness from the appliqué plate is less than its LOS thickness. However, the exact reduction in efficiency compared to normal RHA steel is unclear. If we proceed to simply add the LOS thickness of the appliqué plate to the array, then an effective thickness of 532.7mm is obtained. The enhancement to the protection level could be speculated to be an increase from 490mm to 530mm against HEAT, rounded down from 532.7mm.

Regardless of the exact increase in protection, the armour is clearly insufficient against modern anti-tank missiles fielded during the early 80's. The standard TOW and Dragon anti-tank missiles would not be effective against this armour, but they were also ineffective against the original 80-105-20 array so the up-armoured design did not provide any real qualitative improvement. The armour is still nominally insufficient against the rather old MILAN missile (580mm penetration) from 1972, and the newly appearing ITOW (1984) and MILAN 2 (1983) missiles were far too powerful for the simple three-layer array to handle. A huge leap in mass efficiency would be required to achieve a sufficient level of armour protection from these threats as well as against future threats without encumbering the tank too much even if some weight gain is normal for all tank designs. This huge leap took the form of Kontakt-1 reactive armour.


The turret is made from MBL-1 armour-grade cast steel, assembled from two pieces. The turret front, sides and rear are cast as a single piece, but the roof is cast separately and welded on. This slightly degrades the structural integrity of the roof, as the weld seams can be weak points. This is probably some byproduct of the close imitation of the T-64A turret design, since the Nizhny-Tagil plant had already mastered the production of one-piece turrets for the T-62 and demonstrated the ability to produce a one-piece turret with composite armour for the Object. 167M. However, the decision to use a monolithic steel turret even though a composite turret with an aluminium filler could be produced is unexplained and rather difficult to justify, given that significant weight savings could have been made without any real drawbacks. Indeed, the Object 172 prototypes developed from the six T-64A tanks sent to Nizhny Tagil had the later standardized T-64A turret with tool steel inserts. This turret should also have been relatively simple to master in production, but this was also not done for some inexplicable reason.

According to a well known CIA analysis of a diagram from a captured Soviet T-72 manual, the thickness of the turret at the mantlet area is 350mm. The mantlet is the area immediately next to the cannon. The area directly next to the machine gun port is already 475mm thick, and from there, the turret only gets thicker, so even the weakest part of the turret can survive a hit from 105mm M392A2 APDS from 500 meters or less and the rest is thick enough to be largely invulnerable to any 105mm APFSDS shell when hit from straight ahead. The diagram is shown below. The areal density of the cheek (475mm) is 3,729 kg/sq.m.

The area immediately next to the gun barrel is especially weak due to the gun trunnion. The diagram processed by the CIA is reproduced rather poorly, so an original diagram from a higher quality Soviet T-72A manual gives us a better idea of the armour profile. The trunnion is highlighted below:

Combining the cast steel from the turret with the gun trunnion, the total physical thickness amounts to only 320mm, but this does not reveal the actual protection value of the steel. While it is definitely quite robust, the material properties of the steel gun trunnion may not be on the same level as the armour-grade cast steel surrounding it. The difference is not likely to be very big, but it is probably big enough to have some meaning for threats with a very small margin of penetration. Adjusted for the lower effectiveness of cast steel, this part of the turret is worth between 280mm to 290mm of RHA steel - thin enough that L28 and M392 APDS rounds with 300mm of penetration at 1 km has a chance to defeat this part of the turret at under 1 km.

The diagram appears to show that only the turret cheek on the right has a thickness of 475mm, and the turret cheek on the left appears to be substantially thinner, but both cheeks are equally thick. Both sides of the turret are symmetrical, and the gunsight interface port constitutes a weak point on the left side of the mantlet, mirroring the machine gun port.

The turret roof over the crew positions are 45mm thick and sloped at 78 degrees, and the thickness of the roof above the gun breech is more than twice as thick, angled at between 78 to 80 degrees. Adjusted for the lower hardness and strength of cast steel, the roof armour is still more than capable of consistently causing all contemporary APDS and APFSDS rounds to ricochet harmlessly. Even depleted uranium long rod monobloc projectiles like the M774 from 1979 would not be up to the task as it had a low aspect ratio (L:D) of 13.32, an ogived tip, and a relatively low velocity of 1508 m/s at the muzzle. When newer and longer long rod penetrators began to appear in the mid-80's, the invulnerability of the roof was seriously challenged.

Due to the geometry of the turret, the maximum physical thickness of the cheeks of around 475mm is not replicated anywhere other than the area immediately beside the gun mantlet. The cheeks become progressively thinner as it nears the edge of the frontal profile of the turret, but the line-of-sight thickness from the front increases due to the rounded shape of the cheeks. From a side angle, however, the relative thickness of the cheeks is significantly lower than 475mm, although still extremely formidable. According to Baryatinsky, the thickness of the turret cheeks at a side angle of 30 or 35 degrees is 400 to 410mm with a vertical slope of 10 to 25 degrees. The thickness of the side armour of the turret (80mm thick) varies between 395mm to 440mm at a side angle of 20 to 25 degrees. Due to the curvature of the turret, the base sections of the turret cheek is less sloped than the upper sections so the claimed 10 degree vertical slope must be for the thicker 410mm section while the 25 degree vertical slope must be for the thinner 400mm section. According to page 159 of "Боевые Машины Уралвагонзавода: Танк Т-72" published by the Uralvagonzavod Production Association, the LOS thickness of the turret at a 30 degree side angle is 410mm.


Based on the available information, the effective thickness of turret ranges between 369mm RHA to 427mm RHA against KE threats in a 60 degree frontal arc. This is not much worse than the early T-64 composite turret with a cast aluminium filler mainly because of the very high physical thickness of the cast steel, but it is significantly worse than the T-64A turret with tool steel inserts. Due to the exclusive use of cast steel in the Ural turret, the mass efficiency of the turret is no higher than 0.9 against KE threats and 1.0 against HEAT whereas the composite armour of the T-64 and T-64A invariably had higher mass efficiency against both types of threats, especially HEAT.

According to this Russian drawing, the LOS thickness of the T-64 turret from the front at a 0 degree angle is 599mm - much thicker than the 475mm of the Ural turret at the same location, but out of that 599mm of turret armour, the first 199mm is cast steel, the 295mm center layer is cast aluminium and the 185mm back plate is cast steel. It should come as no surprise that the areal density of this armour is 3,781 kg/sq.m which is almost exactly the same as the areal density of the Ural turret, so the turret of the T-72 Ural has roughly the same armour mass as the T-64. At a 30 degree side angle, the physical thickness of the turret of the T-64 is 575mm at 20 degrees for a LOS thickness of 612mm. It is equivalent to 410mm RHA against KE threats and 450mm against shaped charges which is significantly higher than the Ural turret which is worth 369mm and 410mm respectively for the two threat types. It would not be possible to improve the protection of the Ural turret to the level of the T-64 turret while retaining the same weight without using a composite construction, making this comparison a perfect demonstration of the benefits of composite armour.

Comparing the T-64A turret with tool steel inserts to the Ural turret, the Ural still loses out, especially in KE protection. The physical thickness of the two turrets is very similar, but the use of tool steel inserts and an RHA back plate in the T-64A turret armour led to a significant improvement in both mass efficiency and thickness efficiency compared to homogeneous cast steel while also reducing the overall thickness of the turret from the T-64 turret with an aluminium filler. When contrasted to the later T-64A turret from 1975 with the corundum balls, the Ural turret is simply crude in comparison which is highly regrettable since the T-72 Ural entered service just two years prior.

Of course, the thickness of the T-64 and the T-64A turrets increase towards the sides due to the curvature of the turret so that the LOS thickness can be 625mm or more at the cheeks directly in front of the gunner and commander, but the same is true for the Ural turret as well. For all intents and purposes, the LOS thickness from the front at 0 degrees will only refer to the minimum thickness of the turret cheeks.3

Using the 400-410mm figure by Baryatinsky and 410mm figure from "Боевые Машины Уралвагонзавода: Танк Т-72", the line-of-sight thickness of the same part of the turret from the front at 0 degrees can be determined with some degree of accuracy by dividing 410mm by the cosine of 30 degrees. This gives us around 473mm, which is extremely close to the 475mm figure obtained by the CIA. Based on this, we can confidently conclude that the turret cheeks of the T-72 Ural offer a generally uniform LOS thickness of steel of 475mm from the front, which is worth 427mm RHA against KE threats when converted from cast steel. When shooting at the turret cheeks from a 30 degree side angle, the cheeks are equivalent to 369mm RHA against KE threats. For context, note that the 120mm L15A5 APDS round penetrates 355mm of steel at 0 degrees at 914 m, while the advanced 105mm M833 APFSDS round penetrates around 360mm of steel at 0 degrees at 1 km and M900 penetrates around 440mm under the same conditions. In other words, the armour is proofed against the most powerful APDS ammunition ever fielded by NATO and also proofed against 105mm APFSDS appearing ten to twenty years later on its 60 degree frontal arc from a distance of around 1 km. Penetration figures for the 105mm APFSDS were taken from a Nitrochemie presentation.

The lack of a composite filling in the turret is disadvantageous when the tank has to deal with HEAT and HESH ammunition, but this is compensated to some extent by the extreme thickness of the steel. HESH works well on homogeneous plate, but there is a limit to how thick the plate can be. As far as the Ural is concerned, HESH is no more deadly than any other high explosive round, which is to say that the turret is completely immune. A bigger challenge would be 105mm HEAT shells. The most common 105mm HEAT round of the day, the M456A2, could reportedly penetrate 425mm of steel armour at an impact angle of 0 degrees or 210mm of steel at 60 degrees. Bearing in mind that the cast-to-rolled armour conversion coefficient does not apply for shaped charges, 425mm of penetration is far too low to go through the 475mm turret cheek in a head-on attack, but it has just enough power to have a chance of success on a shot from the side at an angle of 30 degrees where the LOS thickness of the turret is only 410mm. The fuze on the M456A2 may not work on the roof of the turret due to the extreme slope. Despite the lack of composite armour, the chances of defeating the turret armour from the frontal arc with 105mm HEAT seems to be very slim indeed. However, the Ural turret is still inferior to the T-64 turret with an aluminium filler because it cannot guarantee protection in a 60 degree frontal arc, let alone a 70 degree frontal arc as achieved with the T-64.


In addition to solid armour protection elements, the T-72 Ural is also equipped with four flip-out panels on each side of the hull, known as "gill" armour. "Gill" armour was notoriously fragile. These panels took the place of traditional side skirts and were originally found on the T-64A. Why they never made the effort to combine both side skirts and gill armour on standard production model tanks is not known. From frontal angles, "gill" armour augmented the high resistance of the T-72 to HEAT weapons.

The purpose of these panels was to detonate shaped charge warheads at a great distance from the sides of the tank, thus providing a great deal of spaced armour. However, the coverage offered by these "gills" was limited, as gaps will begin to appear past 35 degrees obliquity. Though they could still work at greater angles, the chance of intercepting an incoming warhead becomes slimmer and slimmer. Even when folded, the panels still provide a modicum of spaced armour, as you can see in the second B&W photo below. It is interesting to note that the suspension is rather densely packed, so there is hardly any room for a shaped charge jet to slip through without colliding with some part of a track or a roadwheel or something. As long as the jet does not break a track link, all that can act as additional armour, especially the roadwheels.

The panels are made of hard vulcanized rubber flaps mounted on aluminium sheets. They offer absolutely no protection whatsoever from any type of KE projectile, although it is very clear that there was a lot of missed potential in improving the relatively thin side armour protection of the tank. The panels are probably great for drying clothes.

The primary disadvantage to gill armour is that the panels are very easy to knock off when maneuvering in wooded areas. The gills are spring loaded so that they bend quite easily if one of the panels happen to cross paths with a tree or some other obstacle, and the heavy duty hinges upon which the gills rotate are very robust. However, the heavy duty hinges are secured onto the fragile rubber mini-side skirt with only two small bolts, as you can see in the photo below.

Notice the thick L-shaped wire; it's the spring that flips these panels out.

Struck squarely in the center of any one of the panels from a 30 degree angle from the central axis of the hull, "gill" armour can provide around 2.2 meters of air space from the hull side armour or more if the panel is struck at the outer edge and less if struck at the inner edge. Under such conditions, a great deal of spaced protection can be achieved. Depending on the exact point of impact on the "gills", the air space between the panel and the side hull armour can range from 1.8 meters to a whopping 3.5 meters.

This would have given the T-72 Ural a great amount of protection from guided missiles and man-portable rockets fielded during the 1960's within a 70 degree frontal arc, but this may not be true for anti-tank missiles of the 1970's. To fully convey the peculiarities of the effects of spacing on the penetration of shaped charges, the drawing below can be of great help. The drawing shows the depth of penetration of a 100mm shaped charge increasing to 700mm (7 CD) when the standoff distance is increased to 0.6 meters, but the penetration drops down to less than 400mm at a standoff of 1.2 meters, less than 200mm at 2.4 meters, and less than 50mm at 4.8 meters.

The normal achievable penetration of the 100mm diameter warhead would probably correspond to the penetration at a 15cm standoff distance or less, since the typical built-in standoff for a rocket-delivered shaped charge warhead with a typical pointed aerodynamic fairing without a standoff probe or a spike tip is usually less than 2 CD. As you can see in the chart, an additional 0.45 meters of space in front of the built-in standoff yields the best penetration obtained from the warhead, and this really helps to communicate the peculiarities of shaped charges. Spaced armour can be effective, but only when applied properly and in sufficient portions. For example, if an APC with a ~400mm-wide track had a sideskirt installed to cover the suspension, it would actually become even more vulnerable to a shaped charge grenade. Even at 30 degrees, the side skirts of a typical tank would not provide sufficient spacing to defeat a tank-fired HEAT shell. Because of this, the primary incentive side skirts to be installed on tanks is to reduce the amount of dust kicked up into the air by the tracks, mainly to prevent the enemy from spotting the tank from faraway distances and also to improve the visibility for other tanks at the back of a convoy. Protection from shaped charges would not be one of the reasons.

If a "gill" armour panel was struck at 30 degree angle by the warhead described in the diagram, there is a good chance that it could have offered just enough space to protect the 160mm side armour (80mm at 60 degrees). Protection would be guaranteed at angles steeper than 30 degrees since the amount of air space provided would increase drastically, but at such angles, the possibility of striking the thinner (70mm) engine compartment side armour arises. All taken together, the side aspect of the tank is quite evenly protected. However, this is only a hypothetical scenario with a nondescript shaped charge. By comparing the specifications of actual anti-tank missiles with the spacing of "gill" armour, it is clear that the results could vary wildly.

Older missiles like the SS.11 (1956) using older shaped charge technology form less cohesive jets due to imperfections in the manufacturing of the shaped charge liner, so the shaped charge jet dissipates more quickly over spaces. A large but old large diameter missile like SS.11 will fail to perforate the side armour of the T-72 despite having a 164mm diameter warhead with a 125mm diameter shaped charge capable of 600mm of penetration, whereas the much newer TOW missile (1970) would go through with ease. Besides the TOW, another interesting example is the ITOW from 1982, as it has a 127mm warhead (152mm missile body) and a 124.2mm diameter shaped charge but only 630mm of penetration - sometimes stated to be as high as 700mm - compared to only 430mm from the original TOW despite having a reduced explosive filler (2.08 kg vs 2.45 kg). This was achieved by adding an extendable probe to increase the standoff distance to 370mm (14.6 inches) as opposed to only 107mm for the original TOW. The implications of these details will be immediately apparent after deciphering the graph below.

For more precise estimations, it is important to keep in mind that the actual shaped charge liner in all missiles is actually smaller than the diameter of the missile warhead, but this is often ignored for the sake of simplicity, but this cannot be done for a more precise estimation like the one in this simple example: The SS.11 missile has an external diameter of 164mm, but the shaped charge in the warhead is only 125mm in diameter. If this 125mm warhead impacted the side skirt of the T-72 at an angle of 30 degrees from the axis of the hull, the 2.2 meters of standoff from the warhead to the side armour would be equivalent to 17.6 CD, or around 19 CD when the built-in standoff of the missile nose fairing is accounted for. This cuts down the penetration of the warhead to less than one CD, or in other words, less than 125mm. The 80mm side hull armour of the T-72 will be more than enough to resist such an attack, having 160mm of effective thickness when angled at 30 degrees.

As another example, the total amount of standoff for the ITOW missile from the 2.2 meters of air space would be 17.7 CD, or 20.7 when considering the built-in standoff distance. If the warhead in the ITOW missile had a shaped charge liner made using older technologies, this would reduce the penetration to only 0.8 CD, or 100mm, but thanks to the superior performance of precision-made shaped charges, the actual penetration of the missile would be around 2.5 CD, or 298mm. There is always a chance that one of the roadwheels or a track link could be in the path of the shaped charge jet, but otherwise, the missile will punch straight through the side hull armour and cause some serious damage. Even more interestingly, the original TOW would be even more lethal, as the total amount of standoff would be only 18.57 CD due to the lack of a standoff probe. Referring to the graph above, this would generate 2.8 CD of penetration, or 347.8mm.

Seeing as the TOW achieved IOC (Initial Operating Capacity) in 1970 and the T-72 only entered service in 1973, it is immediately apparent that "gill" armour would not be very useful in the years to come, except perhaps against smaller shoulder-fired weapons. When "gill" armour was first implemented on the T-64, the most powerful anti-tank missiles at the time were slow, manually guided types with paltry penetration power for their size and weight. Of course, the best case scenario where the "gill" panels create 3.5 meters of air space would be able to neutralize most anti-tank missile threats, but it is not possible to do this consistently. All taken together, it is much easier to understand why this unusual solution was replaced with conventional side skirts in the T-72A upgrade.

Gill armour is useless from the side

These panels are no longer seen even on unmodernized T-72 Urals, having being rapidly replaced with conventional side skirts as seen on the T-72A. This could be due to two reasons already mentioned above; fragility and incomplete coverage. One concrete advantage of the conventional side skirts is that it keeps the amount of dust kicked up by the tracks under control, but why not combine the two?

Such a modification appears to only exist on Czech T-72M1 tanks, but even then, it is not a standard modification for Czech derivatives of the T-72 or even a large scale modification for the T-72M1 as part of some overhaul plan. It is rather likely that the panels were installed as part of a modernization programme, but they simply kept falling off and it became tedious to replace them after every exercise, so they were removed once and for all, leaving only the standard side skirts.


Protection-wise, the production model T-72A differs from the T-72 Ural and Ural-1 mainly by the implementation of composite armour in the turret. The gill armour had also been replaced with conventional side skirts. The front hull armour was the same as in the Ural-1. The T-72A can be directly compared to the Leopard 2A0, as both were introduced in 1979.

Glacis Array

The upper glacis armour on the T-72A was identical to the T-72 Ural-1, which was introduced just three years prior. In 1983, the T-72A received a 16mm appliqué armour plate alongside its predecessors. The total thickness of the glacis with the appliqué armour plate now becomes 231mm, or 616mm when angled at 68 degrees. As we have already examined the armour in full detail, there is nothing else to talk about.

Determining the presence of appliqué armour is simple business. The tow hook area is a good indicator. If the cut-out over the tow hook is present, then appliqué armour is present. This is a good way of distinguishing earlier T-72 models from the T-72B when the turret is not visible, as the T-72B has thicker armour but no appliqué aarmou on the glacis as sometimes claimed.


Notice the characteristic ledge on the middle of the turret cheek

It should be noted that the T-72A was not the first T-72 model to feature the composite armour turret, which is colloquially referred to as the "Kvartz" turret. According to Mikhail Baryatinsky in "Т-72: Уральская броня против НАТО" (T-72: Ural versus NATO), the T-72 began to receive the "Kvartz" turret from 1977 onward in some unknown quantity, which would mean that virtually all of these turrets went to the T-72 Ural-1 model as the production life of the Ural-1 was from 1976 to 1979. This relatively small batch of turrets had composite armour, but also had the extension for the second optic for the TPD-2-49 optical coincidence rangefinder. This unusual model is not seen very often, but the photo below shows at least three of these tanks participating in some exercise.

The T-72A had this turret upon its introduction in 1979, and continued to be manufactured with this turret for five more years until 1984. The composite turret features a cast armour cavity on each cheek filled with a material known as "Kvartz", sometimes referred to as "sandbar armour" or "sand rods". "Kvartz" translates to "Quartz", so quartz is probably an ingredient, but the exact composition of this compound is unknown, though the name implies that it includes granules or powdered substances. Based on an ARMOR article penned by James Warford, the armour of captured Iraqi T-72M1 tanks was thoroughly analyzed in the U.S but the composition of the filler has not yet been disclosed to the public. Warford emphasizes that typical sand is probably not used, and he speculates that the name "Kvartz" hints that quartz may be used and recalls the use of quartz gravel as an ingredient in HCR2 add-on armour kits during WWII. The full ARMOR article can be read here.

The three-layer arrangement of the armour may help it attain greater standards of protection than homogeneous armour of the same mass against shaped charges. As noted with the hull array, the composite nature of the T-72A's turret should give it an added damping effect against high explosives and high explosive squash heads, but also against the shockwave of nuclear explosions as well as the radiation.

The thickness of the T-72A turret is unclear, but we can use the same method employed by the CIA to determine the thickness of the turret of the T-72 Ural.

As mentioned before regarding the turret of the T-72 Ural, the CIA determined the thickness of the turret by scaling it against the known length of the barrel of the co-axial machine gun. Comparing the diagram used by the CIA and the diagram from the T-72A manual, we can clearly see that the "Kvartz" turret is thicker. Taking the machine gun barrel to be 680mm long, we find that the thickness of the cast steel around the machine gun barrel is 370mm - just slightly thicker than on the T-72 Ural. The beginning of the turret cheek to the immediate right of the co-axial machine gun measures approximately 514mm, which is 8.2% thicker than on the Ural turret.

The turret measures 514mm at the start of the cheek, increasing to 600mm at the area directly in front of the commander's cupola and increases to more than 700mm as the cheek progresses to the edge of the inhabited space of the turret. According to According to page 159 of "Боевые Машины Уралвагонзавода: Танк Т-72" published by the Uralvagonzavod Production Association, the LOS thickness of the T-72A turret from a side angle of 30 degrees is 530mm. From a frontal view of 0 degrees, the same spot of the turret should have a thickness of around 612mm which is reasonably close to the 600mm figure obtained from the manual drawing, so the two numbers are mutually supportive to some extent. Of course, the curvature of the turret cheeks makes it impossible to perform a simple conversion like this without a rather large margin of error.

The drawing shows that the 600mm-thick cheek is made from homogeneous cast steel, but the presence and form of the filler was probably omitted for security reasons because it is known without a doubt that the turret cheek at this location is filled with "Kvartz" as shown in the photo below. The subject of the photo is the turret of an ex-GDR T-72M1, purchased by Sweden in the early 90's and used for testing purposes. Many of the vehicles purchased by Sweden during that time are still used today as OPFOR assets for training purposes. Looking closely at the photo below, you will notice that the turret is rusted on the surfaces of the cut, but the filler retains its original colour and some amount of it has fallen out of the cavity.

According to the slide below taken from a presentation file from 2012 belonging to Rickard O. Lindström who worked for the Swedish FMV (Defence Materiel Administration), the T-72M1 turret has a center layer of "sand or kvartz (quartz)". There are two ways to interpret this information: There are different types of turrets, one with a sand filler and presumably meant for export, and one with a quartz filler presumably meant for domestic use. Alternatively, it could mean that the terms "sand" and "quartz" are fully interchangeable in this context. Since the Swedes purchased ex-East German T-72M1 tanks, the first interpretation is out of the question, so it appears that "sand" and "quartz" refer to the same substance. As it happens, quartz sand is the most common type of foundry or moulding sand used for casting steel because the minimum silica content of moulding sand must be at least 98% so that the other chemical substances in the sand do not affect the steel. Quartz is a mineral with a very high silica purity in its natural form, making it a suitable material for moulding sand especially for casting large and thick steel structures such as tank turrets. Other substances are needed for casting such as a chemical binder and clay, so the filler in the turret is definitely not pure quartz sand. The density of the quartz sand filler is hard to determine, especially as it is a sand and not a solid block so not all of the space is occupied by quartz sand grains but by air. According to this website, the density of solid quartz is 2,643 kg/m^3 and the bulk density of quartz sand is 1,201 kg/m^3. Of course, the size of the sand particles used for the T-72A turret is unknown and the filler also includes other substances besides quartz sand, so we can use the 1,201 kg/m^3 number only as an approximation.

Based on the thickness of the steel and "Kvartz" in the cutaway photo, it can be surmised that the cavity containing the "Kvartz" layer, whatever it is, is present in a 1:5 ratio to the steel aspect of the turret, as you can see in the photo below. Knowing that the thickness of the turret cheek from a 30 degree side angle is 530mm, the outer wall of the cavity should be 212mm thick, and the inner wall should be have the same thickness. The cavity containing the "Kvartz" filler should be 106mm thick, assuming that it is about half the thickness of the steel cavity walls. This is very different from the distribution of thicknesses in the turret of the T-64 and T-64A, which had almost the same thickness of filler as the steel walls of the composite armour cavity. The low thickness of the filler in the T-72A turret indicates that it has relatively low mass efficiency (ME) but relatively high thickness efficiency (TE) against both KE threats and shaped charges, as the bulk of the work of defeating both types of threats is still accomplished by the cast steel of the armour.


If the thickness of the cheek in front of the commander's cupola and gunner's sight housing is 600mm when viewed straight ahead, then there should be 480mm of steel and 120mm of "Kvartz" based on the previously estimated 1:5 distribution of thicknesses of the composite sandwich. The pure thickness of the steel armour applies directly to shaped charges, but when it is adjusted for the lower effectiveness of cast steel against kinetic energy threats, the steel component of the turret should be worth around 432mm of RHA steel. The "Kvartz" layer probably has little effect on KE threats although it is still quite thick and should have some contribution, so a reasonable estimate of the value of the composite armour against APFSDS rounds should be around 480mm RHA. This is an improvement of less than 10% over the T-72 Ural, but the improvement in protection against shaped charges is probably more substantial, likely to the tune of 550mm based on the fact that the thickness of the steel alone is already 480mm. Sergey Suvorov gives the slightly more generous figures of 500mm against armour-piercing subcaliber threats and 560mm against shaped charges in his article "T-72: Yesterday, Today, Tomorrow", published in the July 2004 issue of the "Техника и Вооружение" magazine. According to "Боевые Машины Уралвагонзавода: Танк Т-72", the resistance of the T-72A turret from a 30 degree side angle is equivalent to 410mm RHA against APFSDS rounds and 500mm against HEAT rounds.

These estimations also align quite well with information from the November 2006 issue of the famous Russian "Техника и Вооружение" magazine, which mentions in page 14 that in 1993, a report published in the specialized magazine "German Airspace" by A. Mann states that the armour protection of the T-72M1 exhibited protection equivalent to 420-480mm of rolled homogeneous armour when tested against modern 105mm and 120mm ammunition from West Germany. It is known that the turret of the T-72M1 is the same as the T-72A, and the 420mm figure probably refers to the upper glacis while the 480mm figure probably refers to the turret, so if this assumption is correct, then our 480mm estimation is completely accurate. It is quite possible that Mann's 480mm figure is lower than Suvorov's 500mm figure because Mann's figure specifically relates to modern German APFSDS from the late 80's and early 90's which were invariably designed with improved performance against composite armour whereas Suvorov's figure may have been sourced from Soviet testing using known Western APFSDS shells from the early 80's like M111 "Hetz". With a total thickness of 600mm but an effective thickness of 480-500mm RHA, the thickness efficiency of the composite armour is 0.8-0.83.

Compared to the T-64B turret armour, the "Kvartz" turret is slightly superior in both KE and shaped charge protection, as the T-64B obr. 1976 turret offers 450mm RHA of protection from APFSDS and 550mm RHA against HEAT. However, this was not achieved with better efficiency but by the higher thickness of cast steel and, therefore, higher mass of the armour. The total thickness of the T-64B turret ranged from only 400-475mm at 0 degrees obliquity and around 60% of the total thickness is composed of lightweight aluminium with ceramic balls, a far cry from the 600mm total thickness and 20% non-metal filling of the "Kvartz" turret.

It can be calculated that the areal density of the "Kvartz" filler would be 144 kg/sq.m assuming that the bulk density of quartz sand is applicable and the estimated 120mm thickness is correct. Also, the cast steel would have an areal density of around 3,768 kg/sq.m assuming that its estimated thickness of 480mm is correct. Overall, the areal density of the array is around 3,912 kg/sq.m which is equal to the weight of a homogeneous steel block measuring 498mm in thickness. An unopposed rule of thumb is that the basic nature of any type of composite armour inherently increases its mass efficiency against shaped charges compared to a homogeneous equivalent, the only variable factor being the magnitude of the increase in efficiency. However, the thickness efficiency of a simple sandwich-type composite armour design is almost always less than its LOS thickness unless the density of the filler is greater than steel. Based on these two simple rules, the HEAT protection offered by the armour must be more than its mass relative to steel in terms of thickness so it should be equal to more than 500mm RHA (rounded up from 498mm RHA), but it must also be less than the LOS thickness of the entire armour sandwich so it should be less than 600mm. The middle point of these two numbers is 550mm, so Suvorov's 560mm figure is perfectly reasonable and is very likely to be the truth. Besides that, the effective relative thickness of 480-500mm RHA against KE implies that the mass efficiency is 0.96-1.0 which is better than the 0.9 mass efficiency coefficient for the monolithic cast steel turret of the T-72 Ural. Against HEAT, Suvorov's 560mm figure implies a mass efficiency coefficient of 1.12 which is an small improvement over the 1.0 mass efficiency of monolithic cast steel armour, but quite reasonable considering the overall simplicity of the "Kvartz" turret design.

Whatever "Kvartz" really is, it appears to be effective against 3BM-15 APFSDS. This was demonstrated by the famous T-72M1 turret in the Parola Tank Museum, Parola, Finland. Tag number 5 in the photo below marks the impact of a 3BM-15 shell into the left turret cheek. Photo by Andrej Smirnov.

According to the placard (see below) at the museum hosting the turret, the shell was stopped completely after digging only 170mm through the multilayer armour. This is incredibly strange, as this would mean that the shell passed either failed to defeat the outer cast steel wall or passed through the first wall but then stopped in the "Kvartz" layer. More importantly, we do not know the range (simulated or otherwise) at which the shot occurred, and we have no idea how they determined the depth of penetration. The inner wall of the turret was obviously not cut up to examine the armour, so they must have poked a stick into the shell crater until they hit solid resistance. Since we that know "Kvartz" is a quartz sand, then it is very likely that the filler substance simply refilled the hole where the shell passed through and the measuring stick compacted the filler as it was pushed in. It would not show how deeply the tungsten carbide slug of the 3BM-15 shell entered the inner wall of the armour array. On the other hand, the statement on the placard can be interpreted to mean that the shell defeated the outer cast steel wall, passed through the "Kvartz" layer and penetrated 170mm into the inner cast steel wall, where it stopped. Either way, it is strange and may be a typo of some sort.

Nevertheless, this example is valid enough to understand why it is often misleading to express the protection value of armour in terms of RHA steel. 3BM-15 is known to be capable of penetrating 400mm of steel at 0 degrees at 2 kilometers, but none of that performance could be seen on the composite armour of the T-72A turret. One has to contend with the fact that there are are multitude of unique APDS and APFSDS penetrator designs. M735 APFSDS, for instance, has a tungsten alloy penetrator with a teardrop shape encased in a steel sheath which peels away upon impact.

And the 3BM-15 along with all Soviet APFSDS designs before "Vant" and "Mango" were comprised of a steel projectile encasing a small tungsten carbide slug. These penetrators will not behave in the same way as the composite penetrator of the M735 or the long rod penetrator of the M774 when striking the same composite armour. As such, it would be rather foolish to assign a fixed armour value to a composite array. That said, we can still express the cast armour component of the composite armour array in terms of RHA. Some penetrator designs may be badly affected by "Kvartz", and some may be less so.

The resilience of the cheeks against contemporary APDS against kinetic energy projectiles of all sorts should still be very high, definitely high enough to resist 105mm APFSDS from well into the 80's. It should not, however, be able to resist 120mm DM13 at combat distances of 1500 meters, unless the composite penetrator design of DM13 is badly affected by non monolithic armour in the same manner as 3BM-15. If DM13 is indeed much worse off from not being a monobloc penetrator, then it is perfectly possible that DM13 cannot penetrate the cheeks at combat ranges or at least the upper boundaries of normal combat ranges.

According to first hand accounts on the performance of ex-East German T-72M1s during Canadian testing, found here, new experimental 105mm shells, presumably designed in the late 80's, claimed to be "jazzed up" to match 120mm rounds in performance, failed to perforate the turret armour. Apparently, the impact only formed a "slight [dinner] plate sized bulge in the armour and cast some paint flakes around the turret wall". The hull armour fared worse, but still quite respectably. If this anecdotal account is true, these tests echo the initial relationship between M111 "Hetz" and the T-72A, as "Hetz" was able to defeat the glacis armour at close ranges while the turret was effectively invulnerable.


The T-72A introduced steel-reinforced plastic side skirts (interwoven textile skirt), which provided complete coverage for the sides, excluding the roadwheels. They were mounted 610mm away from the side of the hull, and could thus still drastically reduce the effectiveness of a small HEAT warhead when impacted at a steep angle, though certainly not to the degree that the gill armour configuration could achieve.

In general, "soft" side skirts like the type which the T-72A uses do not provide enough protection from serious shaped charge warheads at most angles of attack. At angles of 30 degrees or so, the amount of spacing provided (1220mm angled) would be enough to dissipate the shaped charge jets of most tube-launched HEAT grenades and ATGMs of the 50's and 60's enough that the 80mm of side armour (160mm when angled) might be able to handle them, but the chances of even such modest hopes are slim. Referring again to the graph below, we can see how much or how little protection the side skirts would offer.

An older missile like the SS.11 with and older shaped charge liner would only have its total standoff distance increased to around 11.2 CD due to the air space between the side skirt and the side hull plate of the tank, generating a penetration power of 1.7 CD according to the graph. That would be around 212mm of penetration - more than enough to go through the 160mm effective thickness of the side hull plate when it is angled at 30 degrees. Against the TOW or the ITOW, there is no need to even consider the possibility of a failure to penetrate. For the TOW, a total standoff distance of 10.7 CD is created by the air space, generating a penetration power of 5 CD, or 621mm of penetration. That is much, much more than the actual 430mm of penetration from the warhead on its own. It is only slightly better with the ITOW, as 12.8 CD of standoff is created, generating a penetration power of 4.3 CD or 534mm. On the other hand, the amount of space may be enough to decrease the penetration of the 66mm warhead of a LAW to something low enough that the side hull armour plate is able to handle, depending on the angle of impact. The good chances of resisting shoulder-fired anti-tank rockets fired from the flanks (where there was none with "gill" armour) was likely enough to outweigh the lack of protection from obsolete anti-tank missiles, seeing as both solutions would be incapable of protecting the tank from contemporary missiles like the TOW. When we also consider the lack of durability associated with "gill" panels, it is obvious that the decision to switch to a conventional side skirt was a completely pragmatic one.


Kontakt-1 is a type of ERA first introduced in 1982. Upon beginning mass production, the Soviet army immediately embarked on a large scale upgrading programme in 1983 to equip all tanks in active service with the new ERA. The extensive programme saw the majority of T-72 tanks outfitted with Kontakt-1 within the year, and almost all tanks were outfitted by the end of the 80's. 

There are two types of Kontakt-1 blocks - full sized and reduced size. The reduced size block is used to protect special areas of the tank, like behind the headlights.

Mounting the blocks are easy. Each one is bolted onto a tinny spacer mounted all over the surface of the hull and turret. The ease of installing and replacing the blocks meant that the entire modification could be done as part of regular scheduled maintenance. However, simplicity comes at a price in this case. The ERA boxes are rather fragile, and can be quite easily knocked off when the tank is travelling through densely wooded areas, or perhaps traversing obstacles in urban sprawl. This is perfectly illustrated by the example below:

Each Kontakt-1 block consists of two 4S20 explosive elements, which are plastic explosives sandwiched between two flat steel plates. The operating principle of the armour lies in the disruption of shaped charge jets through the violent separation of the steel plates sandwiching the explosive layer upon detonation. It is sometimes claimed that the large number of small gaps between the individual blocks leaves a statistically large portion of the tank surface vulnerable, but this is only partially true at very specific angles. This is examined in the diagram below, taken from "Защита Танков" (Protection of Tanks) by V.A Grigoryan. The column of numbers 'N' to the left indicates the number of reactive plates that a shaped charge jet must pass through depending on the point of impact. As you can see, even if a warhead impacted the edge of one of the Kontakt-1 blocks, the design of the blocks is such that the jet must pass through at least two 4S20 elements. If a warhead impacted the middle of a Kontakt-1 block, the shaped charge jet will intersect with the 4S20 element of the first block, and then continue into the next block, where it will intersect with both 4S20 elements for a total of three intersections.

At the 68-degree angle of installation on the upper glacis and on the turret cheeks, the 40mm gap between the Kontakt-1 blocks does not significantly weaken the overall protective qualities of the reactive armour. The overlap between the blocks when viewed frontally is also sufficient to counteract the decrease in ERA efficiency from edge effects (shaped charge jet impacting the edge of the ERA element).

From a frontal perspective, Kontakt-1 provides uncompromising coverage despite the presence of gaps between the individual blocks. The same can be said of the blocks installed on the side skirts. It is obvious that the height, angle and spacing of the reactive armour package was tailored specifically for an installation angle of 68 degrees, and problems arise when the blocks are installed as smaller angles than that. As long as the blocks are installed at the appropriate angle, there are only a few circumstances in which the gaps between the blocks become weak points, and even so, they are quite small.

The 4S20 elements are arranged in a V-shape with an angle of 9 degrees between them. The mass of the explosive material in each element is 260 grams and the explosive power is equivalent to 280 grams of TNT. The explosive elements are highly insensitive to ensure that they can survive rough handling, being hit by machine gun fire, autocannon fire, set aflame in napalm and any other other form of damage as long as it is less powerful than a shaped charge. Kontakt-1 is so safe from external damage that the one thing that destroyed tanks clad in Kontakt-1 are universally observed to retain their Kontakt-1 blocks even if the metal of the tank itself is completely burnt out from a cook-off of catastrophic detonation. Here are some examples:

Kontakt-1 on this T-72B shows damage from splinters and the right turret cheek has sustained a hit, 

And this Georgian T-72B:

And on this Georgian T-72AV:

The highly insensitive nature of the plastic explosives contained inside 4S20 explains why Kontakt-1 has absolutely no effect on KE rounds - they are so insensitive that they fail to detonate when hit. The low sensitivity also makes Kontakt-1 easier to defeat by tandem warheads using the non-initiation approach. Neither the 4S20 elements themselves or the thin steel walls of the Kontakt-1 block have any resistance to KE threats.

The weight of each block is 5.7 kg, while the reduced size block weighs slightly less. A full set covering the entire tank weighs approximately 1.2 tons. The 4S20 explosive elements can be removed from the block by simply unbolting it, essentially leaving empty metal boxes bolted to the tank. This was always done as a safety precaution before putting tanks into long term storage. 

Kontakt-1 is extremely easy and simple to install. All that are needed are some bolts and nuts.

The Kontakt-1 blocks are installed on bolts that are welded to the surface of the tank armour. On uparmoured T-72 models featuring a 16mm appliqué armour plate, the ERA blocks are simply installed on top of the appliqué plate as you can see in the photo below.


When a shaped charge jet passes through the explosive elements, the resulting explosions will propel the walls of the box at a very high velocity at oblique angles to the jet, thereby cutting off most of the body of the jet. Compared to the Israeli Blazer ERA, Kontakt-1 is much more powerful, has more flyer plates, is better angled, much less sensitive to changes in angle, and has a more optimized sandwich arrangement.

Each individual 4S20 explosive element is technically considered an explosive reactive armour panel by itself. In Russian nomenclature, each explosive element is classified as a so-called "Dynamic Element", as it can work adequately on its own, like "Blazer", for example. The explosive element consists of of two medium hardness steel sheets sandwiching a layer of plastic explosives. The steel box containing the two explosive elements has walls measuring 3mm thick. The steel box does not merely function as a container for the explosive elements; it also contributes to the overall disruptive effect against shaped charges when the explosive charge is detonated. 

The thicknesses of the three layers of 4S20 is not disclosed, but from the photo above, it appears that the ratio of the thickness of the steel plates sandwiching the explosive layer thickness is 1:2. By scaling the known thickness of a full 4S20 plate to the 3mm walls of the steel box, the thickness of the steel sheets should be around 2.3mm or less, while the PVV-5A plastic explosive layer is around 5.4mm or more. Other sources state that the sandwich composition is 2-7-2 but this is in contradiction to this information placard that states that the thickness of a 4S20 element is 10mm. Regardless of the exact thicknesses, 4S20 has a slightly better ratio of flyer plate thickness to explosive layer thickness compared to "Blazer" ERA, which had a simpler 3/3/3 steel-explosive-steel sandwich configuration according to Rickard O. Lindström This would be a 1:1 ratio.

Using the known characteristics of the PVV-5A plastic explosive used in 4S20, we can apply the Gurney equation for symmetric sandwiches to calculate the velocity of the flyer plates. As mentioned before, the mass of a complete 4S20 element is 1.35 kg, while the mass of the explosive charge is 0.26 kg. The mass of the flyer plates sandwiching the explosive layer is obtained simply by subtracting the mass of the explosives from the total mass of the sandwich. The detonation velocity of PVV-5A is 7400 m/s, so we obtain a Gurney constant of 2.46 km/s. From all this, the velocity of the flyer plates is predicted to be approximately 1.156 km/s. The Gurney method of predicting plate velocities is detailed in "Gurney Energy of Explosives: Estimation of the Velocity and Impulse Imparted to Driven Metal".

In "Stopping Power of ERA Sandwiches as a Function of Explosive Layer Thickness or Plate Velocities", Dr. Manfred Held observed that the performance of 1mm thick flyer plates increased exponentially as the explosive layer increased, concluding that the increases in the flyer plate velocity is responsible for the increased performance. 

This is further supported by the theoretical model proposed by Yadav in "Interaction of a Metallic Jet with a Moving Target". Yadav's model showed that the magnitude of the reduction of penetration of a shaped charge jet was primarily affected by the velocity of the flyer plate, and not by the density of the plate, and that by increasing the ratio of explosive charge thickness to the flyer plate, the penetration of a shaped charge jet could be reduced. A reduction in the density of the flyer plates resulted in an increase of performance due to the subsequent increase of the velocity of the plate. 

Held states that the experimental data obtained by M. Ismail in "Optimization of performance of Explosive Reactive Armors" using 1-3mm flyer plates and explosive layers with thicknesses ranging from 2-5mm fits well into his model, to his surprise. Since access "Optimization of performance of Explosive Reactive Armors" is not currently available, the reproduction of Ismail's data in Held's paper is extremely useful. As we can see in pages 235 and 236, the reduction of residual penetration of shaped charge jets plateaus between explosive layer thicknesses of 2-5mm with both Held's 1mm flyer plates as well as Ismail's 3mm flyer plates. From this data, we can predict that the 2.3/5.4/2.3 configuration of Kontakt-1 should achieve something close to the maximum performance possible from a symmetrical sandwich layout, considering that PVV-5A is slightly weaker than the explosives used by Held.

According to an NII Stali information placard, the dimensions of a 4S20 explosive element is 252x130x10 mm. A complete Kontakt-1 block measures 314 x 148 mm overall, including the sheet metal flaps at each end of the block for attachment bolts to pass through. There are two variants of Kontakt-1 blocks, as you can see below. Diagram taken from "Защита Танков" by V.A Grigoryan.

Stock footage and stills of a Kontakt-1 block being disassembled are available here (link). Disassembly and the removal of the explosive elements can be done with a simple wrench.

The V-shaped arrangement of the 4S20 elements inside the Kontakt-1 block was a unique Soviet development and was substantially more advanced than any other reactive armour configuration available anywhere else in the world at the time. The research paper "A numerical study on the disturbance of explosive reactive armors to jet penetration" penned by a team of Chinese researchers, gives us a good look into how Kontakt-1 would work. The research, which was funded by the Chinese Ordnance Society, involved testing reactive armour on armour plate inclined at an obliquity of 68 degrees using a 54mm shaped charge warhead with a copper liner. This oddly specific angle hints that this research was perhaps part of a Chinese evaluation of the performance of Soviet reactive armour on tanks like the T-72, which had an upper glacis plate sloped at 68 degrees. We can learn much from it as well. The paper describes the effects of a single layer of ERA placed at oblique angles of 45 degrees to 68 degrees under subheading 4.2. Here are the relevant paragraphs, given verbatim:

"4.2. Oblique penetration

The typical interaction patterns of the jet penetrating into ERA and main target at an impact angle of 68° are shown in Fig. 7. Compared with the normal penetration shown in Fig. 6, the reactive armor disturbs the jet more significantly during oblique impact. When the explosive of ERA is detonated, the outward movements of the plates cut the jet directly, thus severely disturbing the penetration process. With the formation of more jet segments as a result of the continuous interaction, the residual penetration capability is reduced significantly. It can be seen from Fig. 7 that, when the disturbed jet penetrate into the plate at a larger impact angle, its tip slides along the surface of the rear plate, resulting in bending, breaking, and scattering the jet (segments). Thus the depth of penetration into the main target is significantly reduced."

"It can be seen from Fig. 9 that the greater the impact angle is, the shallower the penetration depth is. In addition, the penetration depth is reduced significantly when the impact angle is more than 45°. The penetration depth is reduced by 55%–75% in the range from 45° to 68° (impact angle) with respect to case without ERA"

This is Fig. 9:

As you can see, a single layer of ERA with a design similar to a 4S20 explosive element (if not exactly the same) can provide a 75% decrease in penetration at 68 degrees obliquity. But Kontakt-1 is a V-shaped design. How would that fare? Let us take a look under subheading 5.2:

"5.2. Influence of impact angle

Fig. 11 shows the predicted results of main target penetration for the cases with and without 9° V-shaped ERA at various impact angles. It can be seen from Fig. 11 that the penetration capability is reduced by 60%–90% for the range of impact angles studied. Fig. 12 shows the penetration holes of the disturbed jet penetrating into the main target. It is shown that the penetration path is deviated, and the deflection increases with the increase in impact angle. The diameter of the hole, especially at the entrance, becomes larger with the increase in impact angle. Similar to the case of flat ERA described in Section 4.2, the former and the latter are probably caused by the bend of jet and the decentralization of jet, respectively."

Fig. 11 is show below:

With a V-shaped design, the pair of ERA layers, or elements, can reduce the penetration of a shaped charge by 90% at 68 degrees obliquity. According to a fact sheet from NII Stali, Kontakt-1 can reportedly reduce the penetrating effects of shaped charge jets by an average of 55% at 0 degrees obliquity, and up to 80% when angled at 60 degrees. Based on this, increasing the obliquity to 68 degrees could easily garner a 90% reduction, so we have complete justification to treat the Chinese V-shaped ERA as an exact replica of Kontakt-1. 

Furthermore, NII Stali claims that Kontakt-1 can reduce the penetration power of a typical anti-tank missile, using the Konkurs as an example (130mm diameter) by up to 86%, or 58% for a 125mm HEAT shell, or up to a whopping 92% for smaller sized warheads like the one on the 66mm LAW. These figures are consistent with the 90% reduction reported for the Chinese V-shaped ERA at a 68 degree obliquity, with the exception of the claimed reduction of only 58% for a 125mm HEAT shell. It is not exactly known why a 125mm HEAT shell would fare so much better than even an anti-tank missile with a much large shaped charge diameter (the 125mm HEAT shell has a thick casing, so the actual diameter of the shaped charge inside it is only 105mm). A plausible explanation is that the thick-walled spike tip/probe partially protects the tail of the jet when the reactive armour block is detonated and the flyer plates are propelled into the path of the jet.

The details of why the explosive elements are arranged with an angle is examined under subheading 5.3. The research shows that the maximum performance of the ERA can be obtained if the two elements are arranged parallel to each other, but if a shaped charge impacts at 0 degrees obliquity to ERA with such an arrangement, the effect will be absolutely minimal. Since practical experience shows that tanks are not always hit where it is toughest, the V-shape of the experimental ERA would give it better performance in low obliquity hits. Where a simpler single cell ERA may be of minimal value at low obliquity, a V-shaped ERA like Kontakt-1 may still perform its duties with an acceptable loss in performance. But why 9 degrees specifically? The paper explains that varying the angle between the ERA layers does not significantly change the performance of the reactive armour. Here is the relevant excerpt:

"However, the variation of penetration depth with increase of V-angle is quite small. It is observed that the penetration depth is reduced by 85%–90% for all the studied V-angles. Therefore it is demonstrated that the reduction of the penetration depth is not sensitive to V-angles investigated in this paper."

Note that the researchers tested angles of 0, 5, 9, 13, 17, and 21 degrees. 

Here are X-ray photos and simulations of the passage of a shaped charge jet through the V-shaped ERA at a 0 degree obliquity. Even at 0 degrees, the disruptive effect of the ERA is substantial.

Now that we have covered the working mechanism of Kontakt-1, it is important to note that it is not only used to protect the tank from frontal attack. The addition of Kontakt-1 blocks is also important for a different reason, which is that the crew now becomes much better protected from tube-launched or air-dropped shaped charge bomblets and submunitions, though the hatches are not protected. It does not matter very much that ERA blocks have a much smaller obliquity relative to a vertically descending bomblet when mounted on the turret roof, which is almost but not flat, because all small-sized HEDP bomblets have very low armour penetration. Even if penetration is achieved somehow, the after-armour effects from the highly degraded cumulative jet will be pitiful at best. The only disadvantage is that there are numerous gaps between the Kontakt-1 blocks, so the roof of the T-72 cannot be considered immune to such attacks.

During the the First Chechen War, many tanks had their 4S22 explosives stolen and sold on the black market due to the poor economic conditions of Russia at the time and the extremely poor living conditions of Army servicemen. Of course, the Kontakt-1 bricks are not filled during peacetime and they are only filled during preparations for tank maneuvers, but many tanks in Chechnya were left without explosive fillings partially due to the haste of the preparations (hundreds of bricks on each tank makes it a tedious chore) and due to theft. As a result of a combination of these unfortunate circumstances, many tanks rode into Grozny with Kontakt-1 bricks, but with no explosives inside. Similar cases of theft have been reported recently in Ukraine.


The T-72B and the series it spawned represented a very significant step in the evolution of the T-72, with the introduction of bulging armour in the turret as well as in the hull later on. Bulging armour is a type of non-energetic reactive armour (NERA), meaning that it is a reactive armour having the effect of actively degrading a penetrating projectile rather than only passively eroding it or absorbing its energy. This will be explained in an expository section below. The T-72B is also notable for being the first T-72 to incorporate an ERA package as part of its original factory configuration. That is, all T-72Bs were built with Kontakt-1 installed.


The glacis array of the T-72B represents the first major update since the original type found on the T-72 Ural. The thickness of the armour remained practically the same after the first update in 1976 with the introduction of the Ural-1. The spaced armour array of the T-72B may have significantly better protection from KE projectiles than the NERA arrays used in its NATO adversaries, but significantly worse shaped charge protection. Nonetheless, this was fully compensated by the use of Kontakt-1 reactive armour.

The illustration below was prepared by Otvaga and Tank-Net user Wiedzmin, and is confirmed to be correct using photographic evidence.

Arrays 4, 5, and 6 refer to the glacis configuration of the T-72B models obr. 1983, obr. 1985 and obr. 1989 respectively. Here is a listing of each layer, translated from the original claims.

Obr. 1983 / Transitional

60mm RHA + 15mm Air Space + 15mm HHS + 15mm Air Space + 15mm HHS + 15mm Air Space + 15mm HHS + 15mm Air Space + 50mm RHA (215mm Total)

 Obr. 1985 

60mm RHA + 10mm Air Space + 10mm HHS + 10mm Air Space + 10mm HHS + 10mm Air Space + 20mm RHA + 10mm Air Space + 20mm RHA + 10mm Air Space + 50mm RHA (220mm Total)

Obr. 1989

60mm RHA + 35mm NERA (5mm Rubber + 3mm RHA + 19mm Air Space + 3mm RHA + 5mm Rubber) + 60mm RHA + 10mm Anti-Radiation Layer + 50mm RHA (215mm Total)

The first and second T-72B versions incorporated complex spaced steel armour in different configurations, but a pair of bulging plates was finally implemented in the 1989 model. Before we go into detail, be reminded that the T-72B is always outfitted with Kontakt-1, and the 1989 variant is always outfitted with Kontakt-5. Only a few T-72Bs went into service without Kontakt-1, and it appears that these variants were exclusively used for Victory Day parades. It is very likely that they were special parade models that never had Kontakt-1 installed, although it would be very straightforward to add it on at a later time. Note that the T-72B entered service in 1984 and began serial production in 1985, so the most relevant variants are the obr. 1985 and obr. 1989. The pre-production model (Obr. 1983) has been described as a transitional model of the T-72A, and was only produced in very small numbers.

BTK-1Sh steel is possibly used in lieu of the usual 42 SM medium hardness RHA steel for the thicker plates of the upper glacis array, and also possibly for the side hull armour as well. It is known that BTK-1Sh was used in the hulls of late production T-64A tanks since 1973 - 1975, and in the hulls of T-64B tanks since the beginning of production in 1976. The T-80 tank also makes extensive use of this steel in the hull and the cast turret of the T-80U contains thick plates of the steel within the armour cavities.

BTK-1Sh steel is a high strength steel produced by electroslag remelting (ESR), giving it higher hardness than normal medium hardness steel while maintaining a good level of ductility to prevent brittle failure when struck. According to Andrei Tarasenko, the steel BTK-1Sh is used in the turret of the T-72B, so it is very likely that it was used in the hull of the T-72B as well, not to mention that BTK-1ShBis also known to be used in the hull of the T-80 tank according to this document. In general, BTK-1Sh is recognized as a general purpose high strength steel, suitable for welding (according to the aforementioned document, which dealt with the weldability of the steel) and for manufacture in thick plates of up to 85mm, or perhaps more. Depending on the thickness of the plate, the hardness of the steel ranges from 400 to 450 BHN. Tarasenko asserts that the resistance of BTK-1Sh is around 5-10% more compared to RHA steel against subcaliber projectiles at angles of 68 to 70 degree, but the type of subcaliber projectile is not specified. It is likely that the "subcaliber projectile" refers to monobloc tungsten alloy long rod penetrators.

That said, there is no real confirmation that the T-72B uses BTK-1Sh in the hull. Nevertheless, the implementation of this improved steel by 1985 is to be expected, seeing as BTK-1Sh has been used in the production of welded hulls since the early 1970's.

The steel used for the high hardness spaced plates is unclear. It is possible that normal 42 SM medium hardness steel was used for the walls of the array while BTK-1Sh is used for the spaced steel plates, but it is also very possible that high hardness steels were used for the spaced steel plates. If so, then it is highly likely that BT-70Sh steel was used, as it is treated a hardness of around 534 BHN when produced in thin plates. In fact, the patent for BT-70Sh specifically mentions that the range of thicknesses for BT-70Sh is from 15mm to 25mm. This matches perfectly with the thicknesses of the spaced steel plates in the upper glacis array of the T-72B. The relevant passage from the patent is presented below:

"Техническим результатом настоящего изобретения является получение листового проката толщиной 15-25 мм, обладающего высокой противопульной стойкостью в сочетании с пониженной склонностью к образованию вторичных осколков, повышенными характеристиками прочности и твердости при достаточной пластичности и вязкости, что позволит увеличить надежность защитных конструкций."

The translation:

"The technical result of the present invention is the production of sheet steel with a thickness of 15-25 mm, having high ballistic resistance combined with a reduced propensity to form secondary fragments, increased strength and hardness characteristics with sufficient ductility and viscosity, which will increase the reliability of the protective structures."

BT-70Sh is also manufactured using ESR technology, and is suitable for welding, as proven by its usage in the BMP-2 infantry fighting vehicle. However, the spaced steel plates of the armour arrays described for the Obr. 1983 and Obr. 1985 variants are not secured to the side hull plates by welding but are suspended by spacers, as we will see later. This means that welding is not an issue, so high hardness steels with poor weldability can be used.

Without clear answers regarding what steel is used in the T-72B, it is more likely that normal 42 SM steel was used for the hull and for the thicker plates of the array while the thinner spaced steel plates are made from BTK-1Sh.

Obr. 1983

The photo above shows a destroyed T-72 from the first Chechen war. The glacis array of a different destroyed T-72 is visible down at the bottom half of the left side of the photo - it is the ramp on which the tracks are laid upon. Apparently, this array is used by a transitional variant of the T-72A built in around 1983. Note that the spaced steel plates are not welded to the side hull walls which the thick front and back plates are welded to. Rather, the spaced steel plates are held by spacers, presumably with the intention of ensuring proper spacing between the plates.

We will not be examining this array in much detail, because the T-72B obr. 1983 is not a common variant and does not represent the T-72B as a whole. Our analysis will be focused on the Obr. 1985 variant. That said, there are some preliminary observations we can make of the Obr. 1983 glacis array that hold true for all of the other variants.

This array design is a good example of a multi-layered spaced armour array comprised of multiple thin steel plates. Some of the protection value of the array may come from the interference of a shaped charge jet or a long rod projectile by the "lips" formed at the edges of the perforated plates, which are deflected from the neighbouring plate and into the path of the penetrator. The photo comes courtesy of Militarysta from the Polish forum.

According to Militarysta, it was noted that only slightly better results were observed at high angles of obliquity, and that an improvement can be gained by packing more spaced plates in a smaller space. It is inferred that the additional protection offered by the intersection of the "lips" with the body of the cumulative jet is rather low, and would be an inefficient method of employing spaced armour, especially for the T-72B, as there are only three (!) spaced plates in the array, and the ratio of plate thickness to air gap size is one to one. The back surface of the heavy 60mm front plate and the front surface of the 50mm back plate would also have some effect, but overall, the contribution of the "lip" effect is obviously very minor.

Indeed, if the array in the T-72B obr. 1983 truly focuses on relying on the use of these "lips" to disrupt the cumulative jet, then it is very likely that the armour would be worse than that of the uparmoured T-72A upper glacis. Recall that the combined thickness of the steel in the array amounts to only 155mm, while the rest of the array is filled with nothing but air. An up-armoured T-72A upper glacis contains 126mm of steel, and 105mm of glass textolite, which was identified as the most optimal composite sandwich filler according to NII Stali. The design of the armour of the T-72B obr. 1983 does not hold up even if we make the conservative assumption that the spaced plate array maintains the same resistance to shaped charges as the T-72A but has improved ballistic resistance against long rod penetrators.

Therefore, the theory that the spaced armour for the T-72B obr. 1983 relies on "lips" for increased protection is a completely insufficient explanation, and does not justify the change from a glass textolite filler to a spaced plate array. Rather, the effectiveness of the spaced armour against KE threats is derived from the mechanisms of projectile deflection, destabilization, yawing, and foreshortening, among others. According to Andrei Tarasenko, the first version of the T-72B hull armour (presumably referring to the obr. 1983 variant) was 20% more effective than a homogenous plate by mass, so the mass efficiency coefficient is 1.2. The obr. 1983 array has a mass equivalent to 413.8mm of steel, so by using the mass efficiency coefficient of 1.2, the effective thickness should be 496mm RHA against long rod APFSDS. Against HEAT, the mass efficiency of spaced armour is typically not significantly more than 10% greater than monolithic homogeneous armour of the same mass, meaning that the average coefficient is typically around 1.1. There seems to be no specific information on the performance of the T-72B obr.1983 hull armour array against HEAT attack, but if we operate under the basic assumption that the armour must be at least equal to the original T-64A design which was equivalent to 450mm RHA, we find that the mass efficiency coefficient in this case is a perfectly reasonable 1.087, close to our 1.1 average. We will further examine spaced armour as a general topic in more detail in the section regarding the Obr. 1985 array. 

Obr. 1985

The photo above shows the exposed glacis armour of a damaged T-72B3, taken during the 2015 Tank Biathlon. As you may recall, the T-72B3 program refurbishes and modernizes old T-72Bs. The vast majority of T-72B in the Russian army are Obr. 1985 tanks, so it should be no surprise that the vast majority of T-72B3s will have the same base armour. 

The total thickness of this array is 220mm which is only 5mm more than the 60-105-50 array of the T-72A and 6mm less than the upgraded 16-60-105-50 array, but the thickness of steel in the array is increased from 110m-126m to 170mm. The total thickness of steel is also greater than in the Obr. 1983 array: at the 68-degree angle of the upper glacis, the physical LOS thickness of steel in the Obr. 1985 array is 454mm. This is an increase of 35.5% in mass over the original T-72 Ural armour array, 26.1% over the T-72 Ural-1 and T-72A armour array, and 12.7-20.2% over the same arrays when upgraded with the appliqué plate.

Again, it is very obvious that the spaced steel plates are not welded to the side hull armour plate by looking at the photo below. Note the jagged edges of the front and back plates and on the lower glacis plate. This is evidence of welding. The spaced steel plates, on the other hand, are clean. The spacing between the plates is maintained by metal spacing brackets similar to the type seen on the Obr. 1983 variant, but they are removed in the photo below. This explains why the space between the plates is uneven and some of the plates are in contact with each other, whereas the plates of the damaged T-72B3 seen in the photo above clearly show uniform spacing between the plates. The spacers for the hull in the photo below were presumably removed because it was about to be scrapped.

The glacis array of T-72B obr. 1985 is similar to the early obr. 1983 version, but probably more effective due to a more nuanced design. If the internal spaced steel plates are made from BT-70Sh steel while the heavy front and back plates were made from BTK-1Sh, the spaced plates will have a very high hardness of around 534 BHN and the heavy front and back plates would have a hardness of 450 BHN. If BTK-1Sh is used for the spaced plates instead of BT-70Sh and the heavy front and back plates are made from normal 42 SM steel, then the hardness of the spaced places will be around 450 BHN, while the heavy front and back plates would remain the softest at 340 BHN. Alternatively, a combination of 42 SM and BT-70Sh is also possible. 

Whatever the combination is, it should be quite effective against monobloc long rod penetrators. A simple method of estimating the protection level would be to find the areal density of the armour array and modify it with a mass efficiency factor, but there are obstacles. The problem is, even though the task of calculating the areal density is trivial (it is 3,564 kg/sq.m), finding mass efficiency figures for the T-72B is not. From areal density alone, the hull armour array of the T-72B obr. 1985 would be similar to the armour of the Leopard 2A0 (~3,500 kg/sq.m), but the actual effective armour value requires the mass efficiency to be known. For instance, since the Leopard 2A0 is known to use NERA armour, it must have a higher mass efficiency against HEAT threats compared to the spaced steel armour of the T-72B, but it is also known that early NERA designs were largely ineffective against long rod penetrators so the armour may be less effective than the T-72B against KE threats. Knowing the areal densities, the comparative thicknesses of the armour does not matter that much as it is obvious that the thicker array is simply filled with more air than the thinner array.

It is worth noting that the November issue of the famous Russian Tekhnika i Vooruzhenie 2006 (Журнал Техника и Вооружение) magazine mentions in page 14 that the protection of the 1985 edition of the T-72B is equivalent to more than 550mm against a KE projectile. With a LOS steel thickness of 454mm, the claimed figure implies a mass efficiency of at least 1.21. The veracity of this claim will be verified, and to do so, let us understand the mechanisms that enable complex oblique spaced armour to defeat modern anti-tank projectiles.

The 60mm front plate is intended to particulate shaped charge jets and to erode as well as damage long rod penetrators before they enter and interact with the internal spaced armour array. As shown earlier with the upper glacis armour of the T-72 Ural, the relatively high thickness of the front plate is meant to particulate shaped charge jets so that the efficiency of the filler is increased. The same reasoning applies for long rod penetrators. Long rod penetrators are generally capable of penetrating more armour at higher obliquity than at lower obliquity, so the steep 68 degree angle of the upper glacis is ostensibly a drawback, but that is far from the case here. Long rod penetrators are susceptible to fracturing and deforming after perforating oblique high hardness armour plates, especially at very high angles. This is due to the asymmetric buildup of stress within the rod during penetration, which is immediately released once the rod emerges from the back surface of the plate. The release of stress generally fractures the rod at the tip but sometimes fractures the entire rod as well, and the asymmetric forces also deflect the rod into a direction perpendicular to the surface of the plate. Thicker plates are more effective and more reliable at producing fractures because the longer duration of penetration causes a bigger buildup of internal stress in the rod, leading to a more severe fracture once the rod exits the back of the plate, but thinner plates can be used in this capacity as well.

The behaviour of long rods penetrators as they perforate and emerge from behind an armour plate is termed "breakout", and the period is known as the "breakout phase". These umbrella terms describe the behaviour of penetrator rods as well as the damage inflicted onto them, including yawing, tip deformation, fracturing, and so on.

The study "The Penetration Process Of Long Rods Into Thin Metallic Targets At High Obliquity" by Yaziv et al. gives us a good general understanding of the damages inflicted onto a long rod penetrator during the impact and penetration phases. The experiments and numerical simulations were conducted at target plate angles of 70 to 80 degrees, with two of the experiments having been conducted at angles of 73 and 76 degrees. The angles are slightly higher than that of the upper glacis of the T-72B, but it is still perfectly acceptable to apply the results to the armour array, since 68 degrees is reasonably close to 70 so the operating characteristics are the same. High strength steel plates with a yield strength of 1200 MPa were used in the simulations and in both experiments (A and B), matching closely with the steel known to be used in Soviet tanks.

The long rod penetrators were detailed as being 135mm long with a length to diameter ratio of 17, denoting that the diameter of the rod is 7.94mm. The plates had a thickness ranging from 7mm to 13mm - closely equivalent to the diameter of the long rod penetrator, so the thickness to penetrator diameter ratio is around 1:1. The L:D ratio of the tungsten alloy rod is slightly better than that of the 120mm DM23 APFSDS round (1983), and the rod has a tapered frustrum that matches the profile of common tungsten long rod projectiles. The impact velocity was 1407 m/s.

The tip of the rod ricochets on impact with the plate and shatters into fragments that are deflected away, and only the central part of the rod actually does the work of penetrating the plate. As the central part of the rod penetrates the plate, it is deflected downwards and begins to rotate as it emerges from the plate. These effects are measurable on armour plates at lower angles of obliquity as well, but they generally become significant at angles of more than 60 degrees. The ricocheting of the tip of a tungsten alloy long rod is extremely apparent at a target plate obliquity of 75 degrees, and increases in severity until the critical ricochet angle of the projectile is reached, whereby the entire rod ricochets off the plate and not just the tip. Depending on the L:D ratio of the long rod penetrator, it may ricochet at an angle of 82 degrees or more. For reference, this short clip shows how a long rod projectile can ricochet with catastrophic damage off a high obliquity plate (link) while only doing surface damage to the plate itself.

The downward deflection of the central part of long rod projectiles is caused by a bending moment exerted on the rod due to the non-uniform thickness of plate material above and below the rod as it travels through the sloped plate. The deflection effect occurs at any obliquity and is manifests as a fracture on the rod that promptly detaches from the rest of the body after the rod emerges from the back of the plate. A study titled "Experimental and Numerical Simulation Analysis of the Impact Process of Structured KE-Penetrators onto Semi-infinite and Oblique Plate Targets" by N. Heider et al. offers a more concise explanation of the loads experienced by a long rod projectile as it penetrates an oblique plate.

"During the perforation process the maximum bending moments occur at the tip of the projectile. This corresponds to a situation where the penetrator can be regarded as a cantilever beam with a fixed tip region and the inertial forces acting as loads leading to the typical concave bending of the projectile ... bending dominates the structural loads during the perforation process of KE projectiles."

The bending moment also introduces a lateral velocity component to the rod, and thus induces yaw. The severity of the yaw for tungsten alloy rods depends on a variety of factors, including: the momentum of the rod, the thickness to diameter ratio between the target plate and the rod, the obliquity of the target plate, and the length to diameter ratio of the rod. It is very important to note that this phenomenon could be avoided if a separate heavy alloy segment is added at the tip of the rod, so that the tip of the rod suffers most of the effects of the ricochet and sustains the bending moment during the penetration process described in the citation above, whereupon it detaches from the rest of the rod. There are a few heavy alloy long rod projectiles that feature a separate segment at the tip of the main penetrator, a few examples being BM-42 "Mango", DM53 and M829A3, the BM-42 and DM53 projectiles being the most interesting, both having three separate tungsten alloy penetrators. Because the tip segment of these projectiles is separate from the main rod, the main rod does not get bent or yawed in any way, and maintains the shape of its own tip. The small loss of penetration in a single homogeneous steel target from the use of such a tip is definitely outweighed by the increase in performance against complex composite and spaced armour such as the type found on the T-72B, M1 Abrams and Leopard 2, hence the popularity of segmented rod designs in the modern era. Moving on -

The penetration of the thin plate only has the effect of eroding 18% of the rod, whereas a much larger segment of the rod (27%) was lost from ricocheting and shattering on impact with the surface of the plate. The total amount of rod material lost from these interactions amounts to 45%, but the reduction in penetration from this effect is more than the loss of material would suggest, as the tip of the emerging rod is deformed by the downward deflection during the exiting phase as you can see below. The deformation of the tip can be expressed as another 8-9% of rod material that is deflected downwards. However, 

The velocity loss from this interaction is quite minor at only 10%, which is consistent with the analytic model presented study "Post-perforation Length and Velocity of KE Projectiles with single oblique Plates" by R. Jeanquartier and W. Odermatt. The main method of defeating the penetrator is via the huge loss of penetrator material.

The study indirectly shows that the thickness of the sloped plate has a much smaller effect on the defeat of the penetrator than the hardness and slope of the plate. As mentioned before, the ratio of plate thickness to penetrator diameter is around 1:1 or less in the study, but if the loss in penetrator material from erosion by penetrating the plate amounts to only 18% whereas a 27% loss in penetrator material can be expected from ricocheting and shattering while another 8-9% can be expected from the breakout phase, then it is apparent that increasing the T:D ratio brings much fewer benefits compared to decreasing the thickness of the plates while increasing the number of plates so that the number of impacts and breakouts is maximized. One of the benefits of a thicker plate is that the severity of deflection on the tip of a perforating rod is increased due to the increase in the lateral velocity component, so there is definitely a difference in performance between the thin 10mm spaced steel plates and the thick 20mm spaced steel plates in the T-72B obr. 1985 array beyond the mechanisms of simple armour penetration and ricocheting.

In other words, the amount of penetrator material loss from perforating a single thick spaced plate at high obliquity would be much less than the loss from perforating a pair of spaced plates of half the thickness at the same obliquity. Such an arrangement would vastly increase the mass efficiency of a spaced armour array, so that a large amount of penetrator material can be removed without requiring a large thickness of steel at the cost of requiring increased armour volume to account for the increase in the number of air gaps between multiple spaced plates.

From this, it is easy to see the advantage of multiple oblique spaced plates of high hardness steel. These results are supported by "Oblique Impact of Elongated Projectiles on Massive Targets" by Veldanov et al. and "Ricochet of a tungsten heavy alloy long-rod projectile from deformable steel plates" by Woong Lee et al., and multiple other studies. Of course, it should not be forgotten that all of the same effects apply to the heavy front plate of the array as well as the surface of the back plate. This cannot be ignored in any attempt to estimate the armour value of the array. Likewise, the 60mm heavy front plate of the array contributes to the effectiveness of the spaced plates as it removes and deforms the tip of long rod penetrators, conditioning them for defeat by the spaced plates. This was examined in detail in the section on the upper glacis armour of the T-72 Ural.

The study "Comparisons of Unitary and Jacketed Rod Penetration into Semi-Infinite and Oblique Plate Targets at System Equivalent Velocities" by J. Stubberfield et al. The topic of the study is not entirely relevant to our current investigation of spaced armour as the objective of the study was to compare monobloc and jacketed penetrators, but the study includes the highly interesting results of test firings against a homogeneous RHA block and a simple spaced armour array. The momnobloc penetrator used for the experiments was a W-Ni-Fe rod and the experimental setup is shown below.

With a rod diameter of 8.5mm, the thickness of the spaced plates normalized in terms of rod diameters is around 1.65 so it can be considered a medium thickness plate. The air gap between the two plates in LOS distance is 118mm, which is quite large. The total penetration of the monobloc rod into the two-plate spaced armour is an average of only around 66% of the penetration of the same rod into a semi-infinite plate at 0 degrees. The mass efficiency of the simple two-plate spaced armour is therefore the reciprocal of 66%, or 1.51. That is very good for such a simple spaced armour design with only two plates.

A histogram of the penetration of the monobloc tungsten rod and the jacketed rod is shown on the left, expressed as a percentage of the penetration on a semi-infinite plate. The x-axis lists the results of twelve individual shots described in the table on the right.

The condition of the monobloc rod as it exits the second spaced plate is captured in the radiograph below. As you can see, the rod is highly fragmented and the tip is completely missing, having been destroyed as it was deflected off the surfaces of the first and second spaced plates. The tip was ejected up from the surfaces of the two plates, and the amount of debris visible on the second plate is clearly more than the debris from the first plate, showing that the destruction of the original tip on the surface of the first plate reduced the efficiency of the rod during the following impact with the second plate.

On the topic of air gaps, it has to be understood that the size of the air gap between the plates has no significant effect on the integrity of the rod, because a sideways force component is generated during penetration so the rod is already deflected as it emerges from the plate. However, this does not mean that the size of air gaps in tank armour is arbitrary; one of the functions of the air gap is to allow the tip of the rod to ricochet up and away from the plate and to allow the shattered fragments of the tip to be ejected away from the penetration crater, preventing them from contributing to the depth of the penetration. Without a sufficiently large air gap, the fragments have nowhere to go. Increasing the size of the air gap also gives more time and space for the rod to rotate and yaw before it impacts the next plate, but the amount of space needed to produce a useful reduction in the penetration power of the rod is impractical for tank armour. The simple two-plate spaced array setup used in the study, for instance, measured 184.5mm in total LOS thickness excluding the 120mm air gap between it and the RHA witness block (the actual air gap is bigger due to the angles). The total thickness is therefore around 454.5mm which is around 50% more than the 300mm homogeneous semi-infinite block. Another interesting phenomenon associated with the breakout phase is the continued shortening of a tungsten alloy penetrator rod due to the residual stress on the tip from the release of internal pressure as it emerges from behind a steel plate and into an air gap. This is a potential advantage to having larger air gaps, but once again, the volumetric disadvantage generally makes it impractical to exploit this to the fullest extent. Without such a large air gap, the 1.51 mass efficiency coefficient of the simple spaced armour setup could not exist.

Furthermore, it has been noted in page 247 of "Particular Questions of Terminal Ballistics" 2006 (Частные Вопросы Конечной Баллистики) that for a multi-layered stack of steel plates at a high obliquity (>60°) with no air gaps or with negligible air gaps between the plates, the resistance of the multi-layered stack is 10-12% less than a monolithic homogeneous plate. This is consistent with a myriad of other studies regarding the penetration of complex armour and illustrates the importance of the size of the air gaps in spaced armour. For example, even a simple dual-layer spaced plate armour at a high obliquity (>60°) can be up to 8% more effective than a monolithic homogeneous plate of the same mass when the front plate has a thickness equal to 0.5-1.0 rod diameters.

To understand this phenomenon and to compare the performance of the oblique spaced steel array to a solid RHA block, we can refer to the study "Pretest Predictions of Long-Rod Interactions With Armor Technology Targets" under Subheading 3.0 "Finite-Thickness RHA Target". The penetration of a finite thickness RHA plate was tested using the same long rod projectile. Behind the finite thickness RHA plate was an RHA witness block placed at a distance of 76.4mm, identical to the spaced plate array.

In the study, the finite thickness plate fails before the penetrator reaches the back of the plate and will no longer resist the motion of the penetrator even before it fully perforates the plate. During this 20 microsecond period, the penetrator continues to be eroded despite the lack of resistance from the armour and was shown to erode by 0.2 diameters in length over a distance of 60mm. The residual rod eroded by another 0.3 diameters over the next 16.4mm of air before it struck the witness block. From this, we can surmise that the size of the air gaps in tank armour does have a supplementary effect in degrading long rod penetrators albeit at a rather low rate of 0.065 diameters per centimeter of air. Due to this glacial rate of rod erosion, very large air gaps in tank armour would be needed before a useful gain in mass efficiency will be observed. When considering the relatively small 10mm air gaps in the armour of the T-72B obr. 1985, the eroded length of a rod should not exceed 0.17 diameters per gap. There are five air gaps in the armour, so the total eroded length is approximately 0.867 diameters.

"Ricochet of a tungsten heavy alloy long-rod projectile from deformable steel plates" by W. Lee et al. provides us with additional data that the previous studies regarding high obliquity plates. Among the findings was that the strength of the oblique plate mattered especially at lower impact velocities. For this application, a high strength steel like BTK-1Sh is suitable. However, the findings regarding impact velocity require more interpretation. It is extremely straightforward if a rod were impacting a single oblique plate at a fixed impact velocity, but in the case of the T-72B glacis, the condition of the rod after it has perforated the thick front plate of the array has to be considered. 

During the penetration of a high thickness plate where the rod can achieve steady state penetration, the velocity of the tip of the rod is typically reduced to around half of the initial impact velocity in the first few microseconds after impact but the rod quickly achieves quasi-steady state penetration so the tip retains almost the same velocity throughout the penetration period, while the tail only decelerates gradually at a low rate due to elastic waves propagating from the rod tip interfacing with the target material. Immediately after perforating a plate, the velocity of the tip of the rod begins to equalize with the tail, but this would take almost a hundred microseconds for a rod with an initial impact velocity of 1400-1500 m/s, which translates to a distance traveled of around 70mm before the velocities are equalized. This must have some effect on the ricocheting of the rod on highly oblique plates, but it appears that the effect has not been investigated. Of course, a 60-70mm air gap does not exist between the heavy front plate and the first spaced plate in the array of the T-72B. 

Based on the rod tip and tail velocity data published in the study "Penetrator Strength Effect in Long-Rod Critical Ricochet Angle" by K. Daneshjou and M. Shahravi, it appears that the relatively low velocity of the tip of a rod emerging from a finite thickness plate will decrease the critical ricochet angle such that the tip of the rod will ricochet from a plate at a lower obliquity than the critical ricochet angle of the rod travelling uniformly at the same impact velocity. This decreases the minimum obliquity for partial ricochet as a side effect.

As mentioned before, the propensity of the tip of a tungsten alloy long rod penetrator to ricochet and shatter on impact with an oblique armour plate depends not only on the angle but also on the strength of the armour plate, but in the study by Daneshjou and Shahravi, it was also noted that the effect of target hardness is considerable especially at lower velocities. If the tip of a rod has not equalized its velocity with its tail after perforating a thick plate, i.e it is still in the 700<Vt<1400 m/s or 750<Vt<1500 m/s velocity range, then the relatively low velocity of the tip would increase the partial ricochet angle of the rod as well. With an air gap of only 10mm between the back of the 60mm heavy front plate and the first 10mm thin spaced plate, it definitely seems that there is not nearly enough time for the tip of any long rod penetrator to equalize its velocity with the tail before it impacts the first spaced plate. The entire rod does not need to have a reduced velocity, just the tip, because the ricocheting of the tip from a highly oblique plate depends only on the conditions of the rod at the tip and not at the tail.

The description of the high hardness plate in the paper by Yaziv et al. is consistent with armour plates with a hardness between 400 and 500 BHN, making the conclusions from the paper compatible with BTK-1Sh steel, but if BT-70Sh is used instead, then the protection level offered by the array increases accordingly. According to the patent for BT-70Sh, the maximum strength of the steel is between 1.9 to 2.0 GPa when treated to the maximum hardness of 54 HRC (543 BHN). This greatly surpasses the strength and hardness of the steel used in the simulations and experiments, and would have yielded even better results, but since there is no confirmation on the composition of the steel plates, it may be prudent to stick to the more conservative estimate.

A good example of oblique spaced steel armour arrays can be found in the study "A unified model for long-rod penetration in multiple metallic plates" by S. Chocron et al. A dense pack of six spaced plates was used for the tests, each plate being 19mm thick and sloped at 65 degrees to the vertical plane. Each plate was separated by an air gap of 25.4mm in length and the distance between the last plate and the RHA witness block was 76.2mm. Long rod projectiles were fired at the armour pack at super-ordnance velocity and hypervelocity and the depth of penetration into the witness block was recorded.

Super-ordnance velocity was defined as the range velocities of between 1.72 to 1.78 km/s which exceeds the normal muzzle velocity of 105mm and 120mm guns by around Mach 1 and Mach 0.3 respectively. This simulated hypersonic impacts. Hypervelocity penetrators with a velocity of 2.6 km/s were also tested, but the results of these tests have little relevance to us. The lower the impact velocity of the penetrator, the greater the effect of target and penetrator material strength, and the typical impact velocity range for APFSDS fired from 105-120mm guns at 1.5 km is 1400-1500 m/s, so the effect of the strength of the RHA and HHS plates in the T-72B array is undoubtedly remains a highly relevant factor in the overall protective capabilities of the armour under normal conditions.

The pretest assumption was that the spaced steel plates would offer the same resistance as the line-of-sight thickness indicated, i.e each 19mm plate was assumed to possess a resistance of 45mm of steel. However, the experiment showed that the estimated penetration depth into the witness block was 40mm less than predicted. It was surmised that repeated impacts and breakouts was the cause of the overprediction, and although it was not explicitly mentioned to be a source of penetration loss, it is worth noting that the 1.78 km/s rod was yawed by 2.34 degrees after passing through the spaced plate array, before it impacted the RHA witness plate. This is consistent with all of the other studies concerning spaced armour. The information regarding the yaw of the penetration was included in a different study, "Pretest Predictions of Long-Rod Interactions With Armor Technology Targets".

As you can see in the illustration below, the pressure spikes at the moment of impact with a spaced plate and falls rapidly as the rod passes through the physical thickness of the plate. After perforating the plate, the pressure drops down to zero as the rod travels into the air gap before spiking again as the next plate is struck. 

The inability of the penetrator rod to achieve quasi-steady state penetration through spaced plates results in a reduction in the efficiency of the rod.

According to the experiments, it was deduced that the deformed and fractured tip of the penetrator is a result of structural failure from large stresses, so it was considered to no longer be a part of the rod. For all intents and purposes, the tip was therefore considered to be incapable of contributing to the penetration of the rod, so it was discarded after the perforation of each spaced plate to simulate the detachment of the tip. To simulate the discarded tip, lengths of 1.5 D or 1.8 D were subtracted from the rod, and a loss of 1.8 D was found to generally agree with the results of the super-ordnance penetrator (1.72 km/s) but not the hypervelocity penetrator (2.6 km/s). The analytic model for the hypervelocity penetrator would require a length reduction of as much as 2.0 D to agree with the experimental results. To supplement this, it is stated on page 250 in "Particular Questions of Terminal Ballistics" 2006 (Частные Вопросы Конечной Баллистики) that the length of the rod broken off due to breakout effects for plates at a high obliquity was equal to 1.5 times the LOS thickness of the plate, regardless of the hardness of the plate.

As mentioned before, six spaced plates were used in the array, each 19mm thick and spaced an inch apart from each other (25.4mm). The witness block was spaced behind the last plate at a distance of 76.2mm and simulated a semi-infinite target. The entire array was angled at 65 degrees and the total LOS thickness was 270mm. It is extremely interesting, then, that the LOS penetration of the spaced plate array including the penetration depth into the witness block amounted to a total of only 414mm. Compared to the penetration depth of 524mm recorded for the finite thickness plate at a normal impact angle, the difference amounts 110mm RHA. Therefore, the spaced armour array is worth 110mm more armour than its own LOS thickness, implicitly indicating a mass efficiency of 1.26 for the spaced steel plate array. This is despite the fact that long rod penetrators are known to penetrate a greater thickness of sloped armour plate than unsloped armour plate.

The array of the T-72B is set at a greater angle than the experimental setup (68 degrees vs 65 degrees), but the number of impacts and breakouts is fewer; there are 7 impacts and 6 breakouts for the experimental setup and 6 impacts and 5 breakouts for the T-72B armour, and the spaced plates in the T-72B array are distributed in two different thicknesses. Without knowing exactly how much the effect scales with angle, plate thickness and the number of impacts and breakouts, it is difficult to apply the coefficient directly to the armour, but according to Andrei Tarasenko, the first version of the T-72B hull armour (presumably referring to the obr. 1983 variant) was 20% more effective than a homogenous plate by mass, so the mass efficiency coefficient is 1.2. Compared to the others, the obr. 1983 array design can only generate five impacts and four breakouts making it the worst performer among the three in terms of efficiency, so the ME of the obr. 1985 design should be between 1.2 and 1.26. Applying these coefficients to the 454mm steel mass of the array, it is calculated that the armour offers between 545mm to 572mm RHA of protection against a monobloc tungsten alloy long rod penetrator. This fits well with information stating that the armour of the T-72B is equivalent to more than 550mm RHA. For the sake of simplicity, the ME coefficient is assumed to be around 1.23 so the protection level of the armour should be equivalent to 560mm RHA, rounded up from 558.4mm.

This is reinforced by information on page 250 of "Particular Questions of Terminal Ballistics" 2006 (Частные Вопросы Конечной Баллистики) published by Bauman Moscow State Technical University on behalf of NII Stali, where an examination of spaced armour concludes that high armour obliquity is required for spaced armour to be effective against long rod monobloc penetrators, that increasing the number of plates in a spaced armour array improves its mass efficiency and that increasing the elongation of a long rod penetrator reduces the effectiveness of the spaced armour array. The relevant paragraph is shown below:

"Прирост стойкости двух- и многопреградных структур фиксированной толщины неизбежно падает увеличением длины сердечника БПС в связи с соответствующим увеличением МПР, необходимого для максимально возможного использования эффектов изгиба и разрушения сердечника при пробитии сильно наклоненных преград. На реальном возможных толщинах многопреградных структур следует рассчитывать не более чем на 15% прироста стойкости при воздействии моноблочных вольфрамовых сердечников с удлинением 15...20 (с использованием броневых сталей средней твердости) Этот прирост при больших углах встречи достигается числом преград от трех до пяти, в зависимости от соотношения МПР и общей толщины преграды. При малых углах встречи разнесение преград малоэффективно."


"The increase in the durability of dual and multi-layered structures of fixed thickness inevitably falls with the increase in the length of the APFSDS core due to the corresponding increase in the MPR (author's note: the back plate) required for the maximum possible use of the bending and fracture effects of the core when penetrating strongly inclined barriers. On the realistic possible thicknesses of multi-layered structures, one should calculate no more than 15% of the increase in resistance when exposed to monobloc tungsten cores with an elongation of 15 ... 20 (using medium-hard armor steels). This increase at large meeting angles is achieved by a number of obstacles from three to five, depending on the ratio of the MPR (author's note: the back plate) and the total thickness of the barrier. At small angles of the meeting, the separation of the barriers is ineffective."

In other words, a multi-layered spaced armour array is not more effective than a monolithic homogeneous block of armour at small angles of obliquity and the efficiency of highly angled spaced armour array will decrease when the elongation of a long rod penetrator is increased. Compared to a monolithic homogeneous block when attacked with a tungsten long rod penetrator with an aspect ratio of 15-20, a five-layer spaced armour array has a mass efficiency of around 1.15. In addition to this, using high hardness and high strength steels yields further improvements in the efficiency of the armour array:

"Увеличение прочности (твердости) стальных броневых преград при сохранении пластичности приводит к соответствующему повышению противоснарядной стойкости многопреградных структур. Использование броневых сталей повышенной и высокой твердости с пределами текучести 120. 130 МПа на структурах, характерных для лобового бронирования танков, обеспечивает дополнительный прирост стойкости 10...15%."


"Increasing the strength (hardness) of steel armored barriers while maintaining ductility leads to a corresponding increase in the anti-shelling resistance of multi-layered structures. The use of high and high hardness armor steels with yield strengths of 120...130 MPa on structures characteristic for frontal tank booking provides an additional increase in resistance of 10 ... 15%."

So using high hardness steels leads to an additional increase of 10-15%, so a mass efficiency coefficient of 1.2 is increased to between 1.32 to 1.38. Based on these facts, the mass efficiency of a five layer array (1.2) can be improved by increasing the number of layers in the array and by utilizing high hardness and high strength armour steels. With that in mind, the mass efficiency coefficient of 1.23 estimated earlier seems to be rather pedestrian. It is wise to remain conservative in our estimates, of course, but even so, it is very obvious from our current understanding of spaced armour that the common estimates of "480-530mm" seen on the Internet are far too low to be plausible except perhaps in the context of very highly elongated long rod penetrators, perhaps with an aspect ratio of above 30. 


As you may recall, the L23A1 round was previously shown to be capable of easily piercing the hull armour of the T-72 Ural-1 and T-72A from several miles away based on declassified British data, but against the armour of the T-72B obr. 1985, the situation is completely reversed. The first hint, of course, is that the line-of-sight thickness of the T-72B obr. 1985 array is almost the same as the penetration of the round at 1 km. With the current range of ME coefficients of 1.2-1.26, it is plain to see that that L23A1 has no chance of defeating the array at point blank range even for the armour estimate at the lowest end of the spectrum. Suddenly, the "crude" spaced armour appears to be one of the strongest composite armour designs ever fielded for a tank hull during the Cold War.

All of the information we have gathered is also highly relevant for the West German DM13 and DM23 APFSDS rounds due to the unique composition of the projectiles. The DM13 and DM23 do not use monobloc tungsten alloy penetrators like the American 105mm M774 and M833, but instead have a tungsten alloy rod carried in a hollow steel projectile with a steel armour penetrating tip as you can see in the photo below (DM13). For a composite long rod projectile like this, the steel armour piercing cap has a very poor L:D ratio and a hollow tip, making it a highly inefficient penetrator. This armour piercing cap should be completely eroded by the heavy 60mm front plate of the upper glacis array of the T-72B while the thin steel sheath around the long rod penetrator is peeled away during the penetration. The tungsten alloy long rod penetrator will emerge from the heavy front plate intact and interact with the internal spaced plates of the array. Having an L:D ratio of only 8:1 and 12:1 for the DM13 and DM23 respectively, the performance of the tungsten alloy rods against the spaced array would not be very good. 

1. Steel windshield and armour piercing cap 2. Steel armour piercing cap 3. Tungsten penetrator 4. Steel sheath 5. Tailfins and tracer assembly

However, a separate tip designed explicitly to counteract these effects was not included in most long rod APFSDS rounds, and indeed, there were no long rod APFSDS shells with such a feature fielded for NATO 105mm and 120mm tank guns during the 80's. In that case, a contemporary APFSDS round fired at the T-72B obr. 1985 can be expected to perforate the heavy front plate and the first two light spaced plates, whereby it is severely fractured and yaw is induced in the rod before it impacts the last two heavy spaced plates and the back plate of the array. 

The array design should be adequate for APFSDS shells appearing in the early 80's such as the 105mm DM23 at 1 km or perhaps less. This estimation is based on the knowledge that DM23 is a licence produced version of the M111 "Hetz", and that the reinforced T-72A glacis array is already sufficient to resist the M111 even at short range. Composite shells like the 120mm DM13 (1979) and DM23 (1983) will not perform well against this array, and will most likely fail to penetrate even at the average combat distance of 1 km to 1.5 km. The much more elongated DM33 round (1987) with 480mm of penetration at 2 km (according to the manufacturer, as stated in this document) or 470-480mm at 2 km (according to Swedish data) should be sufficient at combat ranges, considering that DM33 was specifically designed to overcome multi-layered spaced armour and possibly NERA armour as well. If we assume that the T-72B obr. 1985 armour array has a mass efficiency of less than 1.2-1.26 but more than 1.0 against DM33, then it may be justified to expect that it will be vulnerable at a distance of more than 2 km. However, there is some doubt that the DM33 is able to accomplish such a tremendous feat given that its enhanced capabilities against spaced armour is derived from a special tip design according to a patent drawing and not by a segmented penetrator or a sheathed penetrator. The special tip and tip of the penetrator rod will be eroded within the 60mm heavy front plate of the array (160mm LOS thickness) and there are no special provisions for defeating spaced armour incorporated into the bare monobloc rod itself.

In short, the ballistic resistance of the spaced armour array should be very high despite its ostensibly simple construction. However, the resistance of the armour to shaped charge threats heavily depends on the raw physical thickness of the steel, and is nowhere close to the level of protection offered by Non-Energetic Reactive Armour (NERA). 


After learning about how the side skirts on a tank may increase the standoff distance of a shaped charge jet and thereby increase its penetration rather than decreasing it, it seems counter-intuitive that the T-72B uses multiple spaced armour plates in the upper glacis array rather than evolving the three-layer composite sandwich from the T-72 Ural (originally from the T-64A) into a more complex multilayered steel and glass textolite array like the T-64BV or T-80BV, but as is often the case, the terminology can be misleading as not all spaced armour is the same.

From what we have seen of the original T-72 Ural and T-72A composite armour sandwiches, it is known that the heavy front plate of the array is intended to particulate a shaped charge jet before it enters the low density glass textolite filler, thus maximizing the performance of the filler. The chief concern with side skirts acting as standoff for shaped charges is that the skirting is too thin or too light to particulate a shaped charge jet, so it emerges as a continuous, undisturbed jet and gains increased penetration power as it stretches in the air gap between the skirt and the side of the tank. When a side skirt of sufficient thickness like the side skirt of the M1 Abrams is used, the jet is particulated as it emerges, meaning that it ceases to stretch and the jet splits into discrete particles as a result of the velocity gradient along the body of the jet. In other words, the front parts of the jet are faster than the middle and the tail, so the front parts leave the rest behind causing the jet to separate into individual particles of discrete velocities. If you were to take a continuous jet of a fixed length and divide it into 10 individual segments, each segment would have a different velocity but the faster portions do not separate themselves from the slower portions because the jet is accelerating forward over its entire length so that the front part of the jet is continually fed by material (usually copper) accelerating forward from the apex of the shaped charge cone, powered by the explosive charge of the shaped charge warhead. 

The heavy 60mm front plate of the T-72B is still thick enough to particulate a typical shaped charge jet, so the penetration of the jet will not increase by stretching in the air gaps of the spaced armour array. This resolves the issue of spaced armour acting as standoff for a shaped charge warhead. Nevertheless, having air instead of a low density filler gives the densely packed spaced armour array of the T-72B low mass efficiency against shaped charge jets, although it can certainly be considered efficient for its thickness. As mentioned before, a brief explanation on the NII Stali website stated that a low density nonmetallic filler sandwiched between two steel plates was the most optimum configuration at angles more than 60 degrees, and having an air gap instead of a nonmetallic filler gave the worst results of all. This might be interpreted to mean that the spaced steel array in the T-72B obr. 1985 is a step backwards, and this is quite true.

As shown in the table below, a composite sandwich with a low density filler has higher mass efficiency and thickness efficiency compared to any of the three spaced armour configurations. The composite armour sandwich described here is composed of a light 1/4" mild steel front plate and a thick 4" filling of HCR 2 (Hollow Charge Resisting), a simple low density aggregate made from asphalt (75%) and wood flour (25%). Spaced plate No.1 is composed of a single 1 1/4" RHA plate spaced 3" away from the mild steel witness block. Spaced array No.2 is composed of a 3/4" RHA front plate with a 1/2" RHA plate between it and the witness block. Spaced array No.3 is composed of a 3/4" RHA front plate and two 1/4" RHA plates behind it. All four designs have the same areal density. According to the residual penetration figures, configuration No.1 was the least effective, being a crude single-layer spaced plate. Configuration No.3 had the second best performance, or rather, second worst as it was barely an improvement over No.1. The best performing configuration was No.2, showing that spaced medium thickness plates is more effective than spaced thin plates.

Conceptually speaking, the glass textolite-based armour array of the earlier T-72 models is approximately equivalent to the HCR 2 sandwich in that it is much more effective than a spaced array of the same mass, translating to a high mass efficiency. To be more precise, the mass efficiency coefficients of the Ural and Ural-1 armour arrays are both 1.35. The mass efficiency of a spaced steel array of comparable weight would be definitely less than 1.35, but definitely higher than 1.0 due to breakout effects and other phenomena associated with spaced armour arrays. 

It is common knowledge that shaped charges are almost entirely unaffected by armour slope (up to 84 degrees obliquity, at least) and will penetrate the same thickness of homogeneous armour at any angle. This enables us to continue with the assumption that shaped charges are unaffected by armour slope for spaced plates as well, effectively enabling us to use other reference material that examine non-sloped spaced armour plates.

According to "Shaped Charge Attack of Spaced and Composite Armour", spaced armour can be effective against shaped charge jets due to the forces acting on the jet as it penetrates the target plate. The relevant passage is presented below:

"It has been shown [1,2] that the shaped charge jet tip is disrupted when it exits from a finite thickness plate. This is due to longitudinal and radial shock wave effects in the jet causing mushrooming of the jet tip or enhanced particulation. This effect has been utilised in the design of 'Whipple shields' which consist of multiple thin plates."

As you may recall, the same phenomenon was examined as part of our analysis of the T-72 Ural and Ural-1 upper glacis array. The 60mm heavy front plate of the obr. 1985 fulfills the same function as the 80mm heavy front plate in the earlier designs. This is supported by "Spaced Armor Effects of Shaped Charge Jet Penetration", where it is stated that: 

"During the process of target perforation, the jet was compressed, which increased the jet tip diameter. Upon leaving the first target plate, relief of the compressed material occurred, which led to further expansion of the jet tip."

It was also confirmed in "Shaped Charge Attack of Spaced and Composite Armour" that there is a danger of spaced armour having the opposite of the desired effect; noting that "Conversely the ultimate warhead penetration may actually increase with spacing and/or standoff as the warhead is brought closer to an optimum standoff compared to the normally fused short standoff". Other studies dealing with spaced armour have included similar remarks, and is a legitimate concern with spaced armour. As we have already learned, this can be solved by using a heavy front plate in the armour array so that the jet is particulated before it enters the spaced array, thus nullifying the effect of increased standoff.

A detailed examination of the effects of thin plates on shaped charge jets is provided by "The Shaped Charge Jet Interaction With Finite Thickness Targets". The paper examines the interaction of shaped charge jets with individual and multiple armour plates of finite thickness. Here it is mentioned that shaped charge jets continue to erode as they travel in the air gaps between spaced plates (termed "foreshortening" in the paper) in the same way as long rod penetrators continue to erode after the breakout phase. The paper gives this explanation on the effect of finite thickness targets on shaped charge jets:

" ... the problem concerning the interaction of the shaped charge jet with the target whose thickness does not exceed several jet diameters as well as with a set of such targets, spaced at some distance apart from each other, is considered. The existence of air gaps between such targets lead to additional losses of the jet length due to the erosion of its tip region upon the target perforation, which was first noticed by Brown and Finch [1] and was termed "foreshortening" (forefront shortening)."

In other words, the existence of air gaps prevents us from simply adding up the physical LOS thicknesses of the individual plates to find out the nominal RHA equivalency of the armour against shaped charges, as the figures obtained from this method will always be too low. It is necessary to understand that the publicly available RHA penetration figures for shaped charges is always obtained using semi-infinite target plates - targets where the penetration of the jet does not exceed the thickness of the plate so breakout effects are not considered at all. This is the most optimal condition for shaped charges and cannot be applied to spaced armour or composite armour.

The most relevant study on this topic is "On Modelling of Shaped Charges Jet Interaction With Spaced Plate". The paper directly deals with spaced plates at a normal impact angle as well as oblique spaced plates and is highly expository, making it a convenient resource in analyzing the armour of the T-72B obr. 1985. The conclusion of the paper supports the previous claim that spaced armour cannot be directly compared to homogeneous plates, even though shaped charge jets penetrate both types of targets via hydrodynamic interaction. Indeed, it is explicitly stated in the conclusion that equating the two types of targets leads to an overprediction of the jet velocity in the case of spaced plates, meaning that the spaced plates reduce the velocity of a shaped charge jet by a greater amount given the same plate material, the same cumulative total thickness of plate material, the same standoff, and so on. The relevant passage is given below:

"It has been shown that the hydrodynamic penetration theory can be used for getting a good estimation of shaped charge performance against homogeneous steel targets, provided to know the lateral velocities of all jet elements. But it has also been shown that such modelling is not suitable against spaced targets and overpredicts the jet residual velocities after perforating metallic plates."

The paper also examines and compares normal and oblique spaced plates of 10mm thickness angled at 60 degrees. Using the same penetration models (1D-code + eq. (5)) as the plates impacted at a normal angle but with the addition of a simple equation (eq. (6)) to account for the new relative thickness of the oblique plates, it was proven that oblique spaced plates behave and interact with shaped charge jets in the same manner as non-oblique plates, so plate obliquity for spaced arrays has no effect on shaped charge jet penetration except to increase the LOS thickness of each individual plate, unlike in the case of long rod penetrators.

Everything considered, it is clear that the spaced armour array in the T-72B obr. 1985 is less efficient against shaped charge threats compared to a composite armour array with glass textolite, but still more efficient than homogeneous steel. All together, it is likely that the anti-shaped charge properties of the spaced armour array are at least on the same level as the T-72A, not lower. 

It is reported in page 286 of "Particular Questions of Terminal Ballistics" that a multi-layered NERA armour design is 40% more effective than monolithic steel of the same weight and 10-23% more effective than multi-layered spaced steel armour of the same weight. Based on this, it can be surmised that spaced steel armour would have a mass efficiency of 1.14 to 1.27 compared to homogeneous steel armour. Applying these coefficients, it appears that the spaced armour of the T-72B obr. 1985 glacis armour ranges from 517mm to 576mm RHA. The midpoint of these figures is 550mm RHA. This would mean that the armour is only a mere 3.7% better than the T-72A glacis with a 16mm appliqué armour plate for a 13.5% increase in weight, and it would be even less efficient when compared to the five-layer steel and glass textolite designs used in the T-64BV and T-80BV. 

It is important to note that the raw physical thickness of the steel in the T-72B obr. 1985 array (454mm) already exceeds the HEAT resistance of 450mm RHA of the original 80-105-20 array of the T-64, so the armour is so heavy that it fulfills the basic requirement for protection against 115mm HEAT without even considering the advantages of the spaced plates. The armour is sufficient against 120mm HEAT such as the likes of M830 and DM12. To be more specific, DM12 is described in German literature as having a penetration of 450mm at 0 degrees and 220mm at 60 degrees, and the M830 is also attributed with the same figures. The armour would also be sufficient against practically all handheld antitank weapons as well as most older anti-tank missiles like the TOW (430mm penetration), but it has no chance of resisting the ITOW on its own (630mm penetration), and it definitely will not hold up to the Milan (650mm) or the Milan 2 (790mm), or the TOW-2 (890mm penetration), but this is based on the assumption that Kontakt-1 is not present. Due to the installation of Kontakt-1 as standard equipment on the T-72B, the upper glacis should still be completely invulnerable to all of these missiles, and any other single-charge HEAT warhead. The use of tandem warheads would negate Kontakt-1 to a large extent, so missiles like the TOW-2A would be a serious threat to the upper glacis armour.

In conclusion, the spaced steel armour of the T-72B obr. 1985 cannot be labeled as "simplistic" or "crude". The armour is designed to make full use of a complex set of mechanisms aimed primarily at destroying long rod projectiles but also with comparable or slightly better shaped charge resistance as compared to the earlier pattern of composite sandwiches. The perception of spaced steel as a crude method of protection may be valid, but only when compared to NERA armour in terms of shaped charge resistance.

Obr. 1989

According to the description provided by Wiedzmin, the glacis array of T-72B obr. 1989 can be considered to be slightly more advanced on account of its meager NERA array, but it is still rather crude. It is designed to work in conjunction with Kontakt-5, which may explain the change from simple spaced plates to thicker solid steel plates. The NERA plate installed immediately behind the front plate is comprised of two opposing bulging plates, and should be equivalent to a single NERA sandwich. Against shaped charges, the NERA works by disrupting the cumulative jet (this is further discussed in the section on the T-72B turret). Against KE threats, the NERA works based on the principle of penetrator deflection and shearing. As the long rod penetrator enters the array, it activates the first bulging plate, which bulges downward, exerting downwards force on the penetrator, and as the second bulging plate is activated, it bulges upward, exerting upwards force. This exerts a shearing force on the rod.

Stopping the rest of the attacking projectile or shaped charge jet would be the the job of the remainder of the array, 110mm thick in total. The 60mm steel plate, 10mm anti-radiation layer and 50mm steel plate sandwich behind the NERA layer could technically be considered a composite armour sandwich, as the 10mm anti-radiation layer is probably made from polyethylene. Assuming the anti-radiation layer is composed of borated UHMWPE (Ultra-High Molecular Weight Polyethylene) with a density of 1.00 g/cc (density of polyethylene produced in the Soviet Union was 0.92-0.96 g/cc), then the composite sandwich should be quite decent in principle, though it might not be the most optimal configuration given the low thickness of the anti-radiation layer.

Note that the total thickness of steel in the array is 170mm, which is the same as the Obr. 1985 array. Except for the NERA plate and the anti-radiation layer, the mass of the Obr. 1989 array should be very close to the Obr. 1985 array.

The photo below confirms that the array described by Wiedzmin is indeed used on late T-72B models. Credit for all photos goes to

As you can see, the exposed upper glacis array matches the description perfectly. There are three solid plates, and two gaps of appropriate sizes between the three plates. Also, the rearmost plate appears to be slightly thinner than the middle plate, and this matches the thicknesses given by Wiedzmin. The tank in the photo is clearly a T-72B and most likely a T-72BA, because the tracks are UMSh tracks and not RMSh tracks, as shown in the photo below.

It is known that very late T-72A models had the T-72B turret and early spaced armour upper glacis array, but these models still used RMSh tracks. Furthermore, the exposed glacis array in the photo above looks nothing like the spaced armour array, and it is not possible that the exposed array of the tank shows the earlier STEF composite sandwich because it does not have spacer plates, like in the usual T-72A array and also in the transitional T-72A models with spaced armour.

The tank cannot be a T-72B3 either, because the disembodied turret shows four Kontakt-5 panels on the right side, where a T-72B3 would have five. A T-72BA would have four panels because the IR searchlight occupies the position next to the gun mantlet. Therefore, it is not possible that that this tank is somehow a very late T-72A model upgraded to T-72B3 standard.

According to illustrations published in Sergey Suvorov's "T-90: First Serial Tank of Russia", the T-90 has the same hull armour as the T-72B obr. 1989. The illustration below shows a T-72B obr. 1989, as identified by the presence of Kontakt-5 on the hull and the manually operated anti-aircraft machine gun. The drawing of the hull array appears to match what we now understand to be the correct configuration of the armour.

The drawing below shows a T-90, as identified by the forward-facing remotely controlled anti-aircraft machine gun and the large housing for the Agava thermal imaging sight. As you can see, the hull array is identical. Since the T-90 uses the same cast turret as the T-72B, it is safe to assume that the composition of the turret armour is also totally identical between the T-90 and T-72B obr. 1989. It is also mentioned in various examples of literature on tank design that the T-90 is built with the hull and turret of the T-72B obr. 1989, so there can be little doubt that the T-90 kept the then-recent armour of the latest T-72 model.


The November issue of the famous Russian Tekhnika i Vooruzhenie 2006 (Журнал Техника и Вооружение) magazine mentions in page 14 that the protection of the 1985 edition of the T-72B is equivalent to more than 550mm against a KE projectile. This is probably a general descriptor that applies for the frontal arc of the tank including both the turret and the hull, but it is widely accepted that the turret of the T-72B is the stronger of the two, for reasons which we will see later on.

The turret of the T-72B - dubbed "Super Dolly Parton" by American observers - fully retains the usual T-72 layout but with an up-armoured frontal projection and the associated changes made to the armour profile. The two primary constituents of the turret's frontal armour are the solid cast steel cavities and the NERA array contained within them. The hollow cavity in the turret cheeks for the NERA panels is much larger than the cavity in the earlier turret design with a "Kvartz" filler. The front wall of the cavity is approximately 130mm thick at its thickest part (LOS), near the gun mantlet, thinning to 90mm as the turret cheek curves into the side of the turret (normal). The rear facing is composed of the 90mm cast steel wall of the turret cavity casting supplemented by a 45mm HHS rolled steel plate in front of it (pictured below). The HHS plate is most likely BTK-1Sh armour grade high hardness steel.

We don't actually know exactly how thick the cast steel is, but we know for a fact that the rolled plate is 45mm from the famous ARMOR magazine article. The thickness of the cast steel is estimated from the distance between the armour cavity and the gunner's primary sight aperture (it is recessed a bit into the armoured housing). Based on the photo below (one black/white segment = one inch) we can see that this LOS thickness is around 6 inches. Converting the LOS thickness to perpendicular plate, we can confidently say that the thickness of the cast steel is 85mm to 90mm. Combined with the 45 rolled steel plate, the total thickness of the plates behind the NERA array is a healthy 235mm.

The ARMOR Magazine article mentions that the plates inside the cavity are angled at 55 degrees from the gun barrel. As far as NERA armour goes, this is completely sufficient, and the relative angle increases as the turret is viewed from a sideways angle. The combined total weight of the contents of both cavities is 781 kg. 
The multi-stack bulging plate array of the turret consists of 20 modules. This type of armour can be considered a form of integrated NERA (Non-Explosive Reactive Armour).

Each NERA panel may vary greatly in length, but all of them are uniform in their thickness, each module being 30mm thick. The modules are composed of a 6mm rubber interlayer sandwiched between a 21mm steel front plate and a 3mm steel bulging plate. The maximum length of the NERA panel is 280mm. The air gaps between each panel is 22mm. The entire array is angled at 55 degrees relative to the gun barrel.

The Russian terminology for non-explosive reactive armour is the same as in the West, being labeled as a "non-explosive dynamic armour" as opposed to "explosive dynamic armour" which is better known as explosive reactive armour in English speaking countries. The operating principles of the two types of "dynamic armour" are recognized as the same, except that one is simply more energetic than the other. The drawing on the left below describes the action of ERA, and the drawing on the right below shows the action of NERA.

Both types of reactive armour were implemented in the T-72B, the first type to do so in the history of tank design.

The placement of the plates means that four to six plates will intersect with the direct line of fire of a projectile when the turret is being shot at head-on, more plates if the shot lands at the center of the turret cheek and less at the ends. Only two or three plates will be in the path of a projectile when the turret cheek is struck at an angle of 35 degrees. This is superior to the arrangement of the NERA plates in the front hull of the M1 Abrams, which places a maximum of four plates in the line of fire. Behind that is a spacer, which appears to provide almost no armour value as it is only there to brace the NERA plates and provide proper spacing. At best, it is a perforated steel plate which would offer much more substantial protection, but the large size of the holes depicted in the diagram make this highly unlikely. Behind the spacer is the main armour, which is a rolled steel plate that is estimated to be about 160mm thick.

Overall it can be seen that the armour of the M1 Abrams has rather large air gaps, making the large thickness of armour largely meaningless when comparing protection capabilities. The armour of the T-72B is thinner, but also much heavier.

It is interesting to note that the T-72M1M export model tank also uses NERA armour in the turret, although the turret design is not the same as the T-72S which is the export variant of the T-72B with the same turret. The photo below shows a cross section of the armour as described in "Kampfpanzer: Heute und Morgen" by Rolf Hilmes. This is apparently from a single Iraqi tank found after Operation Iraqi Freedom.

How NERA Works:

NERA was first proposed by Dr. Manfred Held in 1973 in a research paper, after inventing reactive armour in 1969 . The Wikipedia page on reactive armour has this to say about NERA:

"NERA and NxRA operate similarly to explosive reactive armour, but without the explosive liner. Two metal plates sandwich an inert liner, such as rubber.[3] When struck by a shaped charge's metal jet, some of the impact energy is dissipated into the inert liner layer, and the resulting high pressure causes a localized bending or bulging of the plates in the area of the impact. As the plates bulge, the point of jet impact shifts with the plate bulging, increasing the effective thickness of the armour. This is almost the same as the second mechanism that explosive reactive armour uses, but it uses energy from the shaped charge jet rather than from explosives.[4]"

The description of how the inert interlayer is energized by the impact of a shaped charge jet is simplified, but accurate. To be more specific, the source of energy is the shock waves travelling through the inert liner between the two metal plates sandwiching it. Here is a relevant passage from the paper "3D Numerical Simulation Of Non-Energetic Reactive Armor", quoted verbatim:

"The protective mechanism of bulging armors is slightly different than explosive reactive armors. When a shaped charge jet hits the inert intermediate layer, a shock wave interactions through the interlayer results in bulging of the metallic layers [Yaziv, Friling and Kivity , 1995], [Gov, Kivity, Yaziv, 1992], [Mayseless et al., 1993]"

The Wikipedia article's explanation of "increasing the effective thickness of the armour" is, sadly, only a half-truth at best. The research paper quoted above gives a short but concise explanation that the moving plates interact with the shaped charge jet and distort it. There is no mention of increasing the effective thickness of the armour in the paper, or in almost every other paper or journal article on NERA published in the last few decades. Dr. Held's early patents for explosive reactive armour describes the mechanism of the reduction in the penetration of a shaped charge jet as a product of the disruption of the jet. Patent 5811712 from 1975, for example, makes this very clear in the following excerpts:

"... the destruction of a hollow charge spike takes place in such a way that the spike is chopped up over large portions of its length, the individual particles of the spike being additionally diverted. The spike, of which the penetrative capacity in a homogeneous wall of steel is otherwise too high, then loses its boring power and remains in a divergent crater in armour plating following disruptor walls of this kind."

"The military effect of the invention resides quite generally in the efficient disruption and destruction of even elongated hollow charge spikes with a very high energy content and with a high velocity gradient, by the intervention of moving parts of the layer or wall in the total length of the spike ..."

"As shown by the present example, the invention makes it possible for the following charge spike, over its entire length, and despite its considerable velocity gradient, to be combated by moving walls and layers, so that by introducing material over a cutting (oblique) path into the traject of the spike the latter is completely disrupted and finally destroyed, or deprived of its boring action."

However, United States Patent 4,368,660 filed by Dr. Held in 1980 under assignment by MBB GmbH mentions the "consumption" of a shaped charge jet as an additional penetration reduction factor. Reading the "Summary of the invention" section of the patent, we see that Dr. Held describes the action of the flyer plates of the reactive armour having the purpose of "cutting up" or "consuming" or "spending" the shaped charge jet, which is referred to as a "thorn". The term "consumption" had so far never been used before when describing the action of flyer plates against shaped charge jets. Based on this, and the remarks of various scientists and academicians, it appears that Held was the first to identify the increase in effective armour thickness as a factor in the reduction of penetration. 

In 2004, Dr. Held published "Dynamic Plate Thickness of ERA Sandwiches against Shaped Charge Jets" in Volume 29 of Propellants, Explosives, Pyrotechnics, issue No. 4. Held examines the mechanism behind the generation of dynamic plate thickness and concludes that the disruption and destruction of shaped charge jets is still the main method of jet defeat by reactive armours. It is important to note that Held defined "dynamic plate thickness" as the virtual plate thickness that intersects with the path of the shaped charge jet. Held made no attempt to explain how the jet is degraded by the intersection, so his definition of "dynamic plate thickness" is merely arbitrary.

It is well known that the intersection of a moving plate obliquely against a shaped charge jet results in the loss of plate material and jet material alike through erosion. However, classifying the interaction as the penetration of the moving plate is misleading. In actuality, the moving plate is penetrating the shaped charge jet as much as the jet is penetrating the plate, so the mechanism cannot be described as simple armour penetration. The most important distinction is that the tip of the cumulative jet will almost always be on the other side of the plate before the plate even begins to move, due to the immense speed of the jet tip, so it is not the tip of the jet impacting the edges of the plate as the plate moves obliquely against it, but the midsection of the jet body. This cannot be described as hydrodynamic armour penetration. Rather, the interaction causes jet particulation, meaning that the single continuous jet is divided into smaller segments, each with their own discrete velocities. The result is that the armour plate behind the NERA plate will be impacted consecutively by two forms of shaped charge jets; a disembodied continuous jet (jet tip), and a smattering of particulated jet segments.

This is why a shaped charge jet does not penetrate smoothly into armour plate after passing through a NERA plate. Instead, shallow craters are created on a large area of the surface of the plate from the impact of the particulated jet, and some end up on the inside the deepest tunnel, which is invariably made by the disembodied jet tip. Jet particles that do not impact the tunnel made by the disembodied jet tip do not contribute to the final depth of penetration of the target plate. This is best seen in the four photographs below, taken from "Study on Rubber Composite Armor Anti‐Shaped Charge Jet Penetration". The craters were produced by a shaped charge jet disturbed by a rubber NERA sandwich plate at four different obliquities.

The greatest reduction in penetration was achieved when the rubber NERA plate was angled at 60 degrees. It is interesting to note that even at 0 degrees, the NERA plate caused some particulation to occur, as evidenced by the pockmarks around the tunnel created by the otherwise untouched shaped charge jet. In this case, the NERA plate acted as simple spaced armour, causing some of the tip of the jet to particulate due to the compression of the jet while it passes through the NERA plate, and subsequent decompression as it exits. At 30 and 45 degrees, the degree of particulation increased drastically, as evident from the much larger surface area covered with pockmarks, but the jet still appears to remain somewhat unperturbed. At 60 degrees, the jet is badly disturbed by the NERA plate and is split into a number of segments. Lateral forces from the bulging plate gives the segments a sideways velocity component, causing them to impact some distance away from the main tunnel created by the disembodied jet tip.

Although Held's study "Dynamic Plate Thickness of ERA Sandwiches against Shaped Charge Jets" deals with reactive armour, his findings can be applied to NERA to some extent. We cannot equate ERA to NERA directly in this context, of course, because the behaviour of bulging plates is simply not the same as flyer plates. The most significant differences are in the kinematics of the plates and their geometry while they are in motion. Held's calculations are based on the assumption that the flyer plate is flat and maintains a constant velocity throughout its interaction with the jet. A bulging plate, on the other hand, takes quite a lot of time to accelerate to peak velocity, and the shape of the bulging plate is curved rather than flat. Thus, the paper can be read to gain an understanding of the general mechanism of dynamic plate thickness only.

Referring to the graph above, we see that as the plate velocity increases, the dynamic thickness increases. For a rear plate (in-pursuit), there is an exponential increase in dynamic thickness against plate velocity, whereas for a front plate (head-on), the rate of increase is almost linear. Note that the velocities required to achieve a high dynamic plate thickness are well beyond the range achievable by NERA plates, so we can confidently infer that dynamic plate thickness is a very minor factor in the reduction of a shaped charge jet during its interaction with NERA bulging armour. 

For an explanation of how we can know the bulging plate velocities attainable by NERA plates, read the three research papers below:

The first paper investigates the deformation characteristics of the rubber interlayer and its ability to displace (bulge) the steel plates sandwiching it, with experiments conducted using a 3/5/3 bulging armour arrangement. The second and third papers examine the mechanisms behind the transfer of energy into the inert interlayer material of a NERA sandwich. All of the papers deal with the impact of shaped charge jets and the transfer of the jet energy into the NERA interlayer at a normal impact angle, but Rosenberg states that the motion of bulging plates is not sensitive to obliquity since the main source of propulsion is the energy transferred into the interlayer. Yadav's paper states that the amount of energy transferred into the interlayer depends on the duration of contact between the shaped charge jet and the interlayer during penetration, and on the velocity of the jet - the higher the better. Rosenberg's paper is the most convenient for us, as the bulging plate velocities for 3mm steel in-pursuit plates have already been modeled for us. Rosenberg's simulations use a 10mm plexiglass interlayer, but also investigate the effect of varying thicknesses. 

Since plexiglass is less dense than rubber, it can be assumed that the peak bulging plate velocity of the T-72B NERA at H=13.2 will be higher than the value stated in the graph. Only the peak bulging plate velocity matters to us because that is the region in contact with the shaped charge jet as it passes through. 

Rosenberg goes on to state in page 304 that NERA plates bulge faster with thinner back plates (in-pursuit) than with thinner front plates (head-on). He goes on to recommend an asymmetric NERA plate design for optimum performance. The context is that NERA plates with thicker front plates and thinner back plates will have superior overall performance, so this is not direct advocacy of the design of the NERA in the T-72B turret, but it is still strongly suggested that such a design would be advantageous as it would cause the back plate (in-pursuit) to bulge at a higher velocity. Based on Rosenberg's data on plexiglass interlayers, a reasonable guess of the peak bulging velocity of the rubber-based bulging plate of a T-72B NERA plate should be between 0.5 km/s and 0.55 km/s.

The graph below, taken from Held's "Dynamic Plate Thickness of ERA Sandwiches against Shaped Charge Jets", shows that the dynamic plate increases exponentially with increasing NERA plate obliquity. At 55 degrees, an in-pursuit flyer plate travelling at 0.4 km/s generates barely more than 50mm of dynamic plate thickness and an in-pursuit flyer plate travelling at 0.8 km/s generates only around 175mm of dynamic plate thickness. Based on our guess that the bulging plates in the T-72B turret have a peak bulging velocity of between 0.5 km/s and 0.55 km/s, the dynamic plate thickness offered by the bulging plate should be between 60mm and 80mm, but only if we treat the bulging plate as a flyer plate. Since it isn't, then the actual dynamic thickness should be much less.

Based on this information, it is clear that the primary mechanism of shaped charge jet defeat by bulging plates lies in the disruption effect of the plate. This confirms the earlier claim that the contribution of dynamic plate thickness to the anti-shaped charge capability of NERA is either negligible, misunderstood, or both.

The photos below illustrate the effect of a NERA plate on a shaped charge jet. The three photos are from three separate repetitions of the same experimental set up. Note that a substantial portion of the tip of the jet is unaffected in all three tests, and the location of the disturbed regions of the jet is consistent between the second and third photo.

The body of the jet behind the tip is disturbed due to the formation of instabilities caused by the disruption of the shape of the jet. According to "The role of Kelvin-Helmholz instabilities on shaped charge jet interaction with reactive armour plates", the disruptions experienced by the cumulative jet are Kelvin-Helmholz instabilities. Kelvin-Helmholtz instabilities are formed when there is velocity shear in the continuous flow of a fluid, namely the shaped charge jet.

There must be some space behind the NERA plate in order for it to perform at peak efficiency. This is because the perturbations to the shaped charge jet do not manifest until a small period of time has passed. Here are several X-ray photographs, taken from Dr. Manfred Held's paper "Disturbance of Shaped Charge Jets by Bulging Armour", page 194.

The photo below shows three bulging plates shot through by a high power shaped charge jet. Notice that there are keyhole-shaped cuts in the plates, and that the plates are cracked.

More energetic interlayer materials can improve the reaction speed of the NERA plates and increase the lateral energy imparted onto the jet. Rubber is the earliest and most basic material for this application, and can be considered the least sophisticated.

According to the document Multiple Cross-Wise Oriented NERA-Panels Against Shaped Charge Warheads, (the same document contains the photos above) a single NERA panel can decrease the penetration of an 84mm shaped charge warhead from 410mm to just 70mm - a reduction of 83%. Placing two NERA panels in parallel reduces the penetration by only one centimeter more, to 60mm. This seems rather odd, but is actually due to the rather simple fact that bulging armour cannot react quickly enough to intercept the tip of a shaped charge jet. Remember that the tip of a cumulative jet formed from a typical shaped charge travels at velocities of between 8 km/s and 10 km/s or more. The high velocity jet tip imparts a lot of energy into the inert interlayer, but it is simply too fast intercept and disrupt, but it is possible to cut off the rest of the jet and prevent it from doing any harm. This fact is illustrated the photo below, taken from the aforementioned document.

The very small reduction in performance offered by the second NERA plate is almost entirely due to the erosion of the jet from impacting the material of the panel (two 3mm RHA plates and one 5mm layer of rubber), not by the movement of the plates. This was because the body of the shaped charge jet had been disrupted by the first NERA plate, leaving only the disembodied jet tip to continue. The disturbed jet body could not contribute to the final penetration depth, as many of the particles were stopped on impact with the second NERA plate. More importantly, however, this tells us that NERA plates alone are not enough to stop shaped charges. The hypervelocity jet tip is too fast to be affected by the movement of the NERA plates, and can only be stopped by erosion against a solid plate. This is why it is always necessary to install an armour plate behind a NERA array to ensure full defeat of a shaped charge jet. This also shows that there are diminishing returns past a certain number of layers of NERA panels in an armour array as the back layers may not be able to contribute much to the reduction in jet penetration.

Moreover, the efficacy of NERA panels will depend on the material of the bulging plates as well. The document "Combination of Inert and Energetic Materials in Reactive Armor Against Shaped Charge Jets", gives us some perspective. A rubber-based NERA panel was also involved in their testing. However, their NERA panel could only effect a 22% reduction in penetration performance for a 64mm shaped charge warhead. Where did the 61 percentage point difference go? Well, here they used a sandwich of 8mm of rubber between two 2mm thick mild steel plates. In the first document, they used a sandwich of 5mm of rubber between two 3mm thick sheets of Domex Protect 300, which is ballistic grade steel with a hardness of 300 BHN, much stronger, tougher and harder than mild steel which has a hardness of only 145 BHN. Both examples were set at an obliquity of 60 degrees. This shows us that even though shaped charge jets achieve penetration by hydrodynamic interaction whereby both the jet and the target behave as fluids, the strength of the bulging plates still matters, and that hydrodynamic penetration is not a sufficient explanation for the interaction between bulging plates and shaped charge jets. It also hints that the extra energy imparted onto the bulging through the elastomer interlayer only increases up to a certain limit. Here, an 8mm rubber interlayer did not improve the performance of the NERA panel compared to a 5mm rubber interlayer.

It is understood, then, that all of the sandwich materials are important. The ability of a bulging plate to effectively disrupt a shaped charge jet also has some dependence on the strength and toughness of the plate itself despite the lack of strength input in the hydrodynamic penetration of jets, and this is because the bulging plate in a NERA panel mainly works by interfering with the shape of the jet rather than by consuming it or eroding it by "brute force" via dynamic thickness. A relatively minor reduction in the penetration of the cumulative jet does come from penetrating the dynamic thickness of the bulging plates as they intersect with the path of the jet, of course, although the definition of dynamic thickness appears to be arbitrary. There appears to have been no studies or experiments to measure the mass of shaped charge jets before and after their interaction with NERA panels so it is difficult to ascertain if any jet mass is eroded or lost during the interaction with bulging plates or if the mass is simply scattered by the disruption of the jet. Now, let us see what the T-72B uses.

How T-72B bulging Armour Works:

The bulging plates in the T-72B work essentially as described above, except that the one plate is much thicker and therefore much more rigid than the other, forcing the thinner plate to bulge. The time taken for the rubber interlayer will be slightly shorter, because the pressure wave from the impact of the shaped charge jet with the thick front plate will energize the rubber interlayer before the jet actually passes through it, and the thick front plate will slow down the jet somewhat before it reaches the interlayer, so the bulging plate can manage to interact with the front part of the jet. The diagram below, taken from the old NII Stali website, shows the passage of the shaped charge jet through the NERA plate in three successive stages.

The first stage shows the shaped charge jet penetrating the front plate, creating a bulge in the rear surface of the plate. The second stage shows the destruction of the rear surface, causing an expansion of the rubber interlayer and the subsequent bulging of the thin rear plate. The third stage shows that by the time the shaped charge jet passes through the rubber interlayer and the thin plate, the thin plate has already begun to move perpendicularly to the front plate.

The unidirectional NERA plate might propel its single bulging plate more violently, since all of the energy absorbed into the inert sandwich layer is used to propel only one plate and not two. Still, the effect of a single bulging plate will be less effective than two plates taken together, because there is one fewer plate to disrupt the cumulative jet. However, this might be compensated by emphasizing more violent expansion in a certain direction, as shown below:

(a) "Backwards moving" means that the plate bulges against the direction of travel of the jet. This is known as an "in retreat" or "head-on" type NERA.
(b) "Forwards moving" means that the plate bulges in the same direction as the direction of travel of the jet. This is known as an "in pursuit" type NERA.

The pictures above are not of an actual simulation of cumulative jet hitting a NERA panel. The plates pictured were moved by explosives which were detonated before the jet reached the plate, but they achieve the same effect in its essence. The photos above shed light on an extremely important phenomenon, which is integral to the operation of the armour of the T-72B. In the turret, the NERA panels are all of the "in pursuit" type. This maximizes their performance, effectively reversing any penalties potentially incurred by the unidirectional design, or at least neutralizing the disadvantages.

The reason for the increased effectiveness of in-pursuit plates over head-on plates is explained on page 59 in "Interactions Between High-Velocity Penetrators and Moving Armour Components". Here is the relevant passage:

"The severe scattering of an SC jet is due to instabilities of the same kind as can be found in two fluids in contact moving in parallel with different tangential velocities (Kelvin-Helmholtz instabilities). Although this kind of instability is seldom observed in solid materials, the very high velocity and relatively low material strength of the jet, in combination with the high contact pressure and the motion of the plate allow instabilities to occur in spite of the stabilizing effect of the material strength. It is recognized in fluid mechanics that an accelerating flow is more stable than a decelerating flow, and the negative pressure gradient due to obliquity of a backwards moving plate accelerates the flow in the jet direction while the positive pressure gradient in the case of a forwards moving plate decelerates the flow in the jet direction."

The bulging armour design on the turret of the T-72B cannot be compared directly to its NATO counterparts like the Abrams. The Abrams uses conventional bidirectional bulging plates. Defeat of the tip of the cumulative jet in all cases is achieved by the erosion of the jet against the NERA plate material itself and by relying on the thick steel plating of the main armour. This is not optimum against long rod projectiles, or any KE projectile, really, as thin NERA plates will do very little against such threats and the bulging effects of the NERA armour will only do so much to the projectile before it impacts the main armour. Therefore, we can say that the NERA armour in the Abrams is capable of handling kinetic energy threats to an extremely limited extent, and that it is optimized for shaped charges.

The NERA plates in the T-72B turret have thick, high hardness steel front plates acting as spaced armour (in the same manner as the glacis array, which we have already discussed) working with bulging plates to defeat the projectile before it reaches the main armour, which is itself additionally reinforced. The substitution of bidirectional bulging plates for a unidirectional bulging plate with a thick armour plate could be a deliberate compromise to boost protection from KE threats, but having a thick plate in front of a bulging plate also improves the performance of the bulging plate against shaped charges. 

In page 286 of "Particular Questions of Terminal Ballistics" 2006 (Частные Вопросы Конечной Баллистики) published by Bauman Moscow State Technical University on behalf of NII Stali, an optimal distribution of thicknesses of the five basic elements of a NERA sandwich (the front plate, interlayer, and the back plate) with a rubber interlayer was formulated based on data accumulated from testing and simulating armour of this type. The study posits that the most rational distribution of thicknesses is as follows:

  • Steel front plate - thicknesses equal to 0.2-0.5 times the caliber of the HEAT warhead.
  • Rubber interlayer - thickness equal to 1.0-2.0 times the diameter of the shaped charge jet.
  • Thin steel bulging plate - thickness equal to 1.5 times the diameter of the shaped charge jet.
  • Size of the air gap behind the thin steel bulging plate - 0.4 times the caliber of the HEAT warhead.
  • The optimal angle of the NERA panel - 60 to 70 degrees.

The jet diameter of a typical shoulder-fired anti-tank HEAT warhead is 2.5-3.5mm. This includes warheads with a diameter of 60-90mm, fired from weapons such as the Carl Gustav and the RPG-7. A typical guided anti-tank missile with a shaped charge warhead caliber of 127-152mm produces a jet with a diameter of approximately 3.0-6.0mm. Based on this, it appears that the NERA panels in the armour of the T-72B are designed with an optimum ratio of thicknesses and provided with an appropriately sized air gap between each panel, but the panels are not placed at the optimum angle. Furthermore, the actual thicknesses of the components (21-6-3) indicates that the NERA design is optimized to handle a HEAT warhead of only moderate size, between 80mm to 100mm. However, the turret of the tank is not only required to defeat anti-tank munitions from a direct frontal attack, but also from a side angle of up to 35 degrees. Due to the obliquity of the NERA panels, aiming at the turret at an side angle increases the relative angle of the panels, up to 85 degrees when viewing the turret at a side angle of 30 degrees. Between the side angles of 0 and 30 degrees, the NERA panels inside the armour cavities of the T-72B turret will be within a 55-85 degree range of angles which is very reasonable.

Needless to say, this optimized design only applies for a steel NERA sandwich using rubber as an interlayer (though the optimum angle applies for all designs). It has been found that more energetic materials can be used for the interlayer instead of rubber, and better materials for the front and back plates can be used instead of steel. According to the same paper, the increase in mass efficiency for shaped charge protection achieved by multi-layered armour using bulging armour is up to 40% compared to homogeneous steel armour of medium hardness, and the armour is more effective than a spaced steel armour array of the same weight by 10-23%. From these figures, it can be said that the mass efficiency of the armour of the T-72B (depending on the angle of attack) can be as high as 40%, depending on the side angle. Based on the claim that NERA armour is 10-23% more effective than spaced steel armour of the same weight, it can be surmised that spaced steel armour would have a mass efficiency of 1.14 to 1.27 compared to homogeneous steel armour whereas the same armour array with NERA has a mass efficiency of 1.4, so even if NERA were absent, the spaced steel plates alone provide more protection than their weight suggests.

Mass efficiency coefficients of up to 20.0 have been achieved with other NERA designs using bidirectional bulging plates and better interlayer materials, but it is important to remember that these reported figures are only for the NERA panel itself and usually have a setup with more than a meter of air space between the panel and a witness plate, unlike the Russian figures here where the NERA panels are part of a multi-layered armour array with limited air gap spaces. In such cases, much of the efficiency is lost after averaging out the numbers from including thick steel plates into the array. Therefore, a direct comparison between NERA designs examined in various scientific studies and the NERA armour of the T-72B is not valid.

NERA armour can work with both both long rod projectiles and shaped charge jets, but the mechanism of defeat is not exactly the same.

When faced with shaped charges, the bulging armour works in the same way as typical NERA plates. As the first bulging plate bulges, the midsection of the jet (the tip is far too fast to be affected) are put under lateral stresses, thus interrupting its shape. Disruption of the flow of the jet causes it to disintegrate into individual particles, and the disruption of the flow also results in Kelvin-Helmholtz instabilities forming in the jet. A sample of the NERA plates used in the T-72B turret can be seen doing exactly this in the X-ray photograph below, taken from The plate is angled at 68 degrees.

We can see that large disrupted portions in the jet, like troughs in a sine graph, appear quite often down the length of a jet, indicating the the jet is highly disrupted. It is unfortunate that the photo is so closely focused on the NERA panel, because we cannot see the tip of the jet and its length and condition - that is the most important observation we could make from an X-ray photo like this. Besides that, it is quite clear that the disturbances in the jet only appear after travelling a certain distance behind the bulging plate, which is completely consistent with Dr. Held's findings in "Disturbance of Shaped Charge Jets by Bulging Armour". Interestingly, we can see the base of the cumulative jet at the far left of the photo, indicating that the shaped charge was detonated at a relatively short standoff distance. Overall, it would appear that the NERA configuration was successful, but we must take into consideration that the plate is oriented at 68 degrees obliquity, and this is somewhat steeper than most experimental samples demonstrated in the research papers cited, not to mention the fact that the NERA plates are angled at only 55° degrees in the T-72B turret.

Other experiments on NERA were carried out in Soviet Russia as well. Here we see a Soviet NERA design using a sandwich of 4mm steel, 3mm rubber and 6mm steel. The disruption of the jet is very massive as well, perhaps even more than the in-pursuit bulging plate design.

However, this does not mean that the armour in the T-72B turret is ineffective. There are a variety of factors that vastly increase the performance of the NERA plates. As mentioned before, the unidirectional bulging plate of the T-72B NERA can be highly beneficial. The research paper "Study on Rubber Composite Armor Anti-Shaped Charge Jet Penetration" examines the effects of interlayer thickness in bulging armour with a rubber interlayer. It is stated on page 701 that "The interference between the back plate and the jet was neglected because the B plate gave a relatively smooth deflection of the jet without characteristic instabilities, whereas the jet was severely scattered by the F plate". The authors defined the back plate as the front bulging plate, and the front plate as the rear bulging plate. See the diagram below, taken from page 696.

The paper "Shaped Charge Optimisation against Bulging Targets" authored by Dr. Held shows that as the velocity of a shaped charge jet tip decreases, the effectiveness of bulging armour increases. This is succinctly illustrated in the diagrams below.

The velocity of the shaped charge jets was adjusted by varying the thickness of the shaped charge liner without changing the the diameter or the cone angle, which remain at 96mm and 60° respectively. The target was a 10mm steel plate in front of a 2/15/4 bulging armour plate. Shaped charges with liner thicknesses of 1mm, 2mm, 3mm and 4mm were tested. As the thickness of the shape charge liner increases, the lower the jet tip velocity. Jet tip dimeter, however, was unaffected. All warheads were detonated at a standoff of 2 CDs, except for the 2mm liner warhead, which was detonated at 6 CD. This skewed the results slightly, but the trend is very clear:

The shaped charges consistently exhibited more symptoms of disturbance as the liner thickness increases, except for the 2mm liner, but again, this exception exists only because the warhead was detonated at a greater standoff distance so that it attained higher velocity by stretching. The 2mm liner jet was also observed to be thinner than the other three, all of which had the same diameter despite having different liner thicknesses, but this was attributed once again to the increased standoff distance of the 2mm liner shaped charge.

Reducing the velocity of shaped charge jets greatly degrades its performance against bulging armour. The jet tip velocities and the liner thicknesses are given on page 368 in the graph. They are as follows:

Liner thickness, mmJet tip velocity, km/s

This is relevant to the T-72B because the thick steel armour in front of the turret cheek cavities will slow down a shaped charge jet drastically as it is penetrated, and thus improve the performance of the NERA plates by the time the jet emerges from the back of the turret cheek and into the cavity. While the bulging armour used in the test is not directly equivalent to the bulging armour configuration of the T-72B, these results are still perfectly applicable since all bulging armour designs work on the same basic principles.

The design of the bulging plates in the T-72B have another advantage because of its thick steel plate. A typical NERA array with multiple thin plates would easily reduce the penetration of a small shaped charge to nearly nothing by the time it reaches the main armour at the very back, but the tip of the cumulative jet will pass through each and every NERA plate on its way there, since it is too fast to be affected by any one of the NERA plates, and the NERA plates themselves offer too little resistance, since they are (usually) made from some plastic or elastomer sandwiched between two thin metal sheets. Because of this, there will be a hole in the second plate, third plate, fourth plate and every other plate behind it all the way to the main armour if attacked by a serious large caliber anti-tank missile. This would presumably make the NERA array of an early M1 Abrams highly vulnerable to tandem warheads.

Some tandem warheads have precursor shaped charge with a shallow cone angle like the type found in the PG-7VR, Panzerfaust 3-T, and in many other designs, including guided missiles like the Kombat. The precursor shaped charge in a tandem warhead would simply fail to penetrate all the way through, or not penetrate much at all in the case of tandem warheads that work on the principle of bypassing the reactive armour rather than destroying it. Case in point: patents for tandem warheads like Patent US5415105 A by Dynamit Nobel Aktiengesellschaft (manufacturer of Pzf. 3-T) have outright stated that:

"When firing against ERA-boxes, such boxes were penetrated by the preliminary charge so that the jet from the main charge could flow almost undisturbed through the hole in the box generated by the preliminary charge."

And the Dynamit Nobel official website says this about the Panzerfaust 3-T:

"The warhead of the Pzf 3-T is designed in such a way that the first of both shaped charges immediately penetrates the add-on armour without initiating the explosive contained therein. Less than one millisecond later, the main charge of the tandem warhead ignites and thereby immobilises the vehicle. The shooter therefore is not exposed to fragments thrown back from a reactive protection element."

In such tandem warhead designs, a hole is created without detonation of the ERA block due to the low energy of the shaped charge jet, owing to the shallow cone angle of the precursor warhead which produces a large diameter, low velocity jet. A large diameter, low velocity jet has less energy and spreads the force of impact more widely over the ERA block, thus preventing its detonation. There are proposals to use non-metal shaped charge liners to further enhance this quality, but it appears that copper or brass liners for precursor warheads are still the norm.

Besides this, other tandem warheads may have a high penetration precursor shaped charge. Such designs may protect the primary shaped charge from being damaged by the flyer plates of the ERA block by extending the delay of the detonation of the primary shaped charge so that the flyer plates have flown clear of the path of the primary shaped charge jet. This is described in detail in Russian Patent 2062439. The TOW-2A, for example, relies on detonating the ERA block to clear a path for the primary warhead, as you can see by the high angle liner for the precursor shaped charge in the diagram below (from official U.S government document, acquired by armamentresearch).

In any case, the thick front wall of the turret cavities of the T-72B turret protects the NERA array within from the influence of tandem warheads, though the same cannot be said of any externally mounted reactive armour blocks. However, the power of the primary charge of the tandem warhead of modern anti-tank missiles may still be a challenge for many parts of the T-72B turret base armour.

Also, recall that there are diminishing returns when multiple NERA plates are installed in an armour array, so it may not be advantageous to install 6, 7 or 8 NERA plates in a composite armour array but place a relatively thin main armour plate behind it. Such an array would be incredibly effective against individual shaped charges of all sizes, but incredibly ineffective against a KE penetrator.

Besides the effects of bulging armour, we must also not ignore the fact that the thick 21mm front plates make a substantial contribute as both spaced armour and to aid in increasing the effectiveness of the bulging plate by decreasing the jet tip velocity. Reading "Spaced Armor Effects on Shaped Charge Jet Penetration" by researchers from the Nanjing University of Science and Technology, we learn that the space in spaced armour may actually increase the penetration of the shaped charge jet if the air gap corresponds to the optimal stand off distance of the shaped charge. Beyond such unlucky coincidences, increasing the size of the air gap is not as beneficial as compared to increases in the thickness of the spaced plates.

Here is the conclusion of the paper, verbatim:

(DOP = Depth Of Penetration)

"The effect of the distance and plate thickness of spaced armor on penetration was analyzed. For a spaced armor plate with a given size, DOP decreased with the increase in the distance between the first and second plate. However, within a certain stand-off range, DOP did not decrease with an increase in distance mainly because of jet stretch, which created increasing penetration on the penetration vs. stand-off curve. When the distance was constant, DOP decreased with an increase in spaced armor plate thickness."

The paper also details the changing physical condition of the shaped charge jet as it impacts and exits the spaced plates. It is noted that "During the process of target perforation, the jet was compressed, which increased the jet tip diameter. Upon leaving the first target plate, relief of the compressed material occurred, which led to further expansion of the jet tip". Needless to say, an increase in the jet tip diameter and its partial particulation are not very beneficial to the penetration power of the shaped charge jet, but not only that; as noted beforehand, Dr. Held's research showed that "robust" jets with larger diameters but lower velocities performed more poorly against bulging armour.

The RHA plates used in the experiment were 10mm thick, angled at 69 degrees - analogous to the spaced steel plates in the T-72B glacis. It is stated that the original jet velocity at the point of formation was 6.5 km/s, decreasing to 5.3 km/s as it exited the first plate, further decreasing to 4.8 km/s as it exited the second plate. In other words, the first plate decreased the velocity of the jet by 18.46%, and the second plate by 9.4%. The smaller reduction offered by the second plate is likely due to the stretching of the jet - the first plate was not thick enough to particulate the jet and halt stretching. Since we have already established that the heavy front wall of the turret cavity (up to 130mm LOS thickness) will substantially decrease the velocity of a shaped charge jet before it even impacts the thick front wall of the NERA plate, it is clear that the jet will be particulated, slow, and therefore highly vulnerable to the bulging plates in the turret cavity when it finally reaches them.

In addition, the shaped charge jet will most likely be disrupted and particulated as it leaves the first NERA plate, leaving only a section of the jet tip travelling at hypervelocity to continue through the array. The jet tip will probably escape the bulging effect of any subsequent NERA plate past the first or perhaps the second plate in an array of typical NERA sandwiches, but the thick steel wall of the NERA plates in the T-72B turret may reduce the velocity of the jet tip before it impacts the next bulging plate. Reducing the velocity of the jet tip may enable the bulging plate to disrupt the tail part of the jet tip segment, which will probably not result in a big reduction in penetration, but in theory, there should at least be some small contribution. The thick walls also help stop the jet tip by acting as spaced armour.


NERA works in a similar way against long rod projectiles. In this context, the Soviet style NERA is clearly more suitable than a traditional sandwich configuration, thanks to the heavy front plate if nothing else. One conceivable advantage of the Soviet NERA design is that the heavy front plate enables energy to be transmitted to the rubber interlayer before the projectile impacts the rubber itself, and this may be a major source of energy for the interlayer. Typical NERA sandwiches with plastic interlayers may find themselves neatly perforated without substantial energetic expansion.

The movement of the bulging plates in the T-72B turret may induce lateral movement and produce internal stresses in a long rod penetrator. The addition of a sideways velocity component in a long rod penetrator can lead to yaw.

According to "The Relation between Initial Yaw and Long Rod Projectile Shape after Penetrating an Oblique Thin Plate" authored by Israeli researchers, even one degree of yaw before striking a thin angled plate would significantly reduce that projectile's penetration potential against any armour behind that plate as a result of the deformation of that projectile.

The two x-ray photos above show the same high elongation tungsten alloy rod interacting with an oblique armour steel plate with 1 degree of yaw, and no yaw. The rod with no yaw lost its tip as the result of breakout effects, but survived mostly unscathed. The rod with 1 degree of yaw, on the other hand, is seen visibly bent and the tip has been violently dislodged. It is visible just above the new tip of the remnant rod. The combination of oblique spaced armour and NERA in the turret of the T-72B may work in that direction.

The greater the yaw, the greater the negative effect. The hard steel strike plate (45mm) behind the NERA array is angled in the opposite direction to the angle of the NERA panels, so that as the long rod penetrator passes through each panel is becomes increasingly deflecting away (both due to deflection from the bulging plates and due to the natural tendency of long rod penetrators to tunnel inwards into the plate), the relative angle between the rod and the strike plate continually increases. Be reminded that there are at least five bulging modules in the projectile's flight path if the turret is shot head-on. Each individual bulging module works with the next module directly behind it to place the penetrator under great stress, causing it to bend, and perhaps fracture as it passes through the multi-layer array.

According to German tank expert author and lecturer Rolf Hilmes, one method to augment the efficacy of NERA armour against kinetic threats is to incorporate a heavy armour plate in front of the NERA array, so that the penetrator is shattered or fractured before it enters the array. This is the function of the heavy cast steel front plate of the turret cheeks. In later iterations of the T-72B, this effect is augmented by Kontakt-5 reactive armour, so that the NERA array in the turret is highly amplified.

If and when the projectile has gone through all of the NERA panels, it will meet the hardened rolled steel plate backing. Angled at 55 degrees to the horizontal axis, the 45mm plate measures 78.45mm. However, the function of the plate is much more significant than its mere thickness suggests, since the projectile that will be striking it will no longer have an optimal shaping, meaning that this plate could function to totally outright shatter the already fractured and damaged penetrator. The dissimilar hardnesses of the 45mm steel plate and the 90mm cast steel wall behind it turns it into a DHA (Dual Hardness Armour) pairing, making it inherently stronger and more resilient than a single monolithic steel plate of the same thickness, so that the mass efficiency is more than 1.0. The softer 90mm cast steel wall behind the hard steel plate will also produce less spall because it is more ductile. This, in addition to the anti-radiation lining acting as a spall liner, means that the beyond-armour effects of a projectile or a shaped charge would be greatly diminished whether it perforates the armour or not.

Note that bulging armour shouldn't be specially affected by projectiles with impressive length/diameter ratios by any great amount. In fact, it's quite possible that greater aspect ratios will actually increase the effectiveness of the array compared to a penetrator of the same mass but lower aspect ratio as it would make bending and fracturing it easier, because the stiffness is decreased while the material properties of the rod remain the same. Snapping of the rod is possible because of the forward momentum of the projectile, which naturally resists a change in the direction of motion. Pressure builds up in the rod due to the large forces opposing each other, and if there is a weakened point in the rod, a plastic hinge point where a segment of the rod may break off is formed. Of course, a stubbier rod would also penetrate less armour overall, so the solution lies within increasingly elongated rods with multiple segments rather than shorter and fatter rods.

Also, a rather important point related to the effectiveness of the NERA array in the turret is its ability to perform when hit at less than optimal angles, especially considering the regularity in which tanks are hit from the flank. The answer is that the NERA plates would work even better at steeper angles, as it would be if the turret was struck from the side, although that is not to say that the tank is better protected from the side. Not at all; against shaped charges, the array would still have more to lose than gain since fewer bulging modules would be in the path of the shaped charge jet. It is in this situation that the high hardness steel front plates of the NERA plates again become particularly useful as the thickness of the steel plates will further increase due to the steeper angle. From an angle of 30 degrees to the side of the turret, a pair of 21mm front plates would yield a total LOS thickness of 240mm in thickness, having a slope angle of 80 degrees. Besides increasing the LOS thickness of the armour array, the high obliquity of the spaced plates at this turret angle ensures high resistance to long rod projectiles because the angle of 80 degrees is very close to the critical ricochet angle of very high elongation long rod projectiles in the 1,400 m/s to 1,500 m/s velocity range, and is the critical ricochet angle for almost all 105mm APFSDS shells. The effect of high obliquity on long rod penetrators has already been thoroughly explored in the previous section regarding the upper glacis armour of the T-72B, so there is no need for further exposition here. Still, it is important to note that the cast steel front armour of the cavities conditions long rods for defeat by the internal spaced NERA armour in the same way as the heavy front plate of the glacis array.

From what has been demonstrated through the rationalizations earlier, it is apparent the effect of a partial ricochet on a long rod penetrator at an angle very close to the critical ricochet angle is catastrophic, and effectively reduces the penetration of the residual rod to only a small fraction of its original capacity due to severe deformation, yaw, and fracturing. This perfectly illustrated in the picture below, taken from the study "Analysis of Critical Ricochet Angle Using Two Space Discretization Methods". The tungsten alloy rods used for the experiments and simulations was 7mm in diameter and 75mm in length, and the impact velocity was 1,000 m/s. The obliquity of the plate in this particular example is 76 degrees, and the plate was 6.25mm in thickness (less than one rod diameter).

As we have already learned from our earlier examination of the spaced armour in the upper glacis of the T-72B obr. 1985, the obliquity has a much greater effect on a penetrating long rod projectile than the layman may assume. In short, we can say with confidence that the T-72B is virtually impenetrable from a frontal 70-degree arc by contemporary 105mm APFSDS unless the weak gun mantlet was hit, and the resistance to 120mm APFSDS is also extremely high.


As with all composite and spaced armours, the complex operation of the T-72B's armour does not truly allow an expression of its protection value in the simple terms of homogeneous RHA plates, especially with regards to KE threats. Generally speaking, T-72B is immune to practically all single charge HEAT missiles like the HOT missile across the frontal arc due to the combination of its NERA armour and Kontakt-1, with the only exception being the devastating 300mm diameter shaped charge of the WDU-20/B warhead on a Maverick missile, although this is meaningless since the 136 kg HE-Frag warhead of the WDU-24/B variant would be just as effective at destroying the tank.

A very basic estimation of the total steel thickness of the turret cheek against shaped charges and long rod penetrators can be done by adding up the LOS steel thickness of all the plates. First, we add up the 130mm cast front wall with the 157mm cast steel rear wall and 78.5mm high hardness backing plate, and then we add the LOS thicknesses of four 21mm high hardness steel plates. The areal density of this armour array is approximately 4,160 kg/sq.m from a 0 degree frontal view at the turret cheek directly in front of the crew positions. The areal density decreases towards the gun mantlet and increases towards the edge of the turret.

As we have seen earlier, high hardness steels actually perform worse than normal RHA steel against shaped charges but perform much better than normal RHA steel against long rod penetrators, but if we ignore every benefit that the nuances of the armour array brings and only use simple modifiers for the cast and high hardness steels (0.9 for cast armour, 1,1 for high hardness armour), then the relative mass of the turret cheek should be equivalent to around 521mm RHA against KE and 530mm against HEAT in pure thickness alone, ignoring the mass of the 3mm bulging plates of the NERA panels. This is reasonably close to the claimed protection value of 550mm against a KE projectile (Tekhnika i Vooruzhenie Magazine, November 2006 issue, p.14) considering that we are simply adding up the thickness in pure steel and ignoring the benefits of the spaced armour configuration, the NERA armour, and the different hardness of the steel plates inside the array. Since jet disruption is the primary mechanism of bulging plates, the actual protection offered by the T-72B turret cheek must be much higher than this. Based on the earlier discussion on NERA, the mass efficiency coefficient of Russian "multi-layered armour" incorporating NERA should be around 1.4. Treating the turret armour of the T-72B as such, we can multiply the mass of the armour - 551mm in terms of steel including four 3mm bulging plates - by 1.4 to obtain 771mm, so it appears that the armour is worth 771mm RHA against shaped charge warheads. Coincidentally, this is close to the figures sometimes attributed to the armour of the M1 Abrams. The mass of the rubber interlayer of the NERA panels is ignored as it is insignificant compared to the margin of error of our estimation.

Besides the base armour alone, Kontakt-1 is installed as standard equipment on practically all T-72B tanks and we know that the reduction in penetration offered by Kontakt-1 at 0 degrees obliquity is 55%, so we can multiply 551mm with the reciprocal of 0.55 to get 1,000mm RHA (we are ignoring that the surface of the T-72B turret is not actually flat). This represents the effective armour thickness of the armour if we treat it as a monolithic armour block. Taking the previous estimate with NERA effects accounted for, the actual effective thickness would be 771mm RHA multiplied by the reciprocal of 0.55, which is equal to 1,400mm RHA. With the base armour alone, the T-72B turret is on par with the turret of the M1 Abrams and the Leopard 2A0, although it is less efficient due to the higher density of the armour. The armour of the Leopard 2A0, for example, is thicker but lighter, and also more efficient. With an areal density of only ~3,500 kg/sq.m, the commonly estimated protection value of 700-800mm RHA against shaped charges implies a mass efficiency of 1.55-1.77, higher than the 1.4 of the T-72B. Efficiency aside, all three tank turrets are capable of resisting a contemporary anti-tank missile (exemplified by a Milan) in a 60 degree frontal arc and can be considered roughly equal, ignoring the Kontakt-1 on the T-72B. With Kontakt-1, the T-72B easily outstrips any other NATO tank of the time in terms of shaped charge resistance.

Against KE threats, it would be extremely difficult to imagine that the "more than 550mm" claim was incorrect considering that the hull armour of the T-72B obr. 1985 is equivalent to around 560mm RHA against a monobloc long rod penetrator with a high aspect ratio. In typical Russian fashion, the 550mm RHA figure against KE threats almost definitely refers to the protection at a ~30 degree side angle and not at 0 degrees, but since the behaviour of NERA and spaced armour is anisotropic, it is not possible to simply divide 550mm by the cosine of 30 degrees to obtain the effective protection level from a head-on frontal view so an accurate estimation cannot be made this way. Figuring out the mass efficiency of the armour is also complicated by the use of cast steel along with the rolled high hardness steel plating and the spaced NERA panels implemented in the armour design. With so many factors to consider at the same time, it is practically impossible to come up with a reasonably correct estimation as the margin of error is simply too high. A reasonable guess is that the armour has a similar effectiveness to the thinner but much more efficient T-80U turret. Considering that the T-72B turret armour array has a weight equal to 530mm of steel and measures 650-700mm in LOS thickness, a ballpark figure of 600mm RHA appears to be completely plausible as it not only implies a rather modest mass efficiency of only 1.089, but it is also not much higher than the claimed ">550mm" figure.

Based on available information, both the turret cheeks and the upper glacis armour would be able to handle M833 (1983) very well, as the M833 is less impressive than the M829 but still slightly better than 120mm DM23. For reference, M833 has a 24mm diameter, 427mm long DU penetrator, travelling at 1495 m/s. Seeing as it would have been the most common ammunition available to M60A3 and M1 Abrams tanks, this is rather important. Latecomers like the M900 (introduction in 1989 to 1990) would still be worse than its more powerful 120mm counterparts like the M829A1 as it has a lower muzzle velocity (1500 m/s) even compared to the relatively slow M829A1 and it does not have a superior aspect ratio. For reference, the M900 has a 23mm diameter, 603mm long DU penetrator. The penetration of the German 120mm DM23 and DM33 tungsten long rod projectiles appear to be insufficient at combat ranges, being less than the physical thickness of the steel in the turret array. M829A1 has the best performance among all other 120mm tank gun rounds at the time of its introduction and for the remainder of the Cold War. It is guaranteed to defeat the armour at combat distances.

The turret cheek cavities offer a great deal of modularity and repairability. The NERA panels are simply inserted into the turret cavity one by one. In the field, replacing the bulging armour is a simple matter of cutting off the top at the weld lines (very distinctly seen in the picture below), putting new panels in, and replacing the top. This makes battle damage very easy to repair, and it also simplifies the installation of upgraded panels in the future.

Aside from that, it must be noted that despite the huge leap in protection relative to the previous T-72 models, the T-72B turret is still as simple to produce as its predecessors since no new technologies were needed to cast the turret and the workmanship required to process the cast turret does not demand any new skills or any retraining. The sheer commodity of steel and rubber makes it very easy and inexpensive to fabricate the NERA panels of the T-72B and the bonding of the rubber to the steel plates is done using only glue. The highly economic design is undeniably an important asset during wartime as it would have ensured a very high volume of production even in the hardest times. Indeed, it is worth noting that the peak of T-72 production in Uralvagonzavod was in 1985 - the year the T-72B obr. 1985 entered mass production.


Coming from the factory, all T-72B models except the obr. 1989 are outfitted with a set of 227 blocks of Kontakt-1 covering the most of the hull and the forward arc of the turret as well as the turret roof. As mentioned before regarding the ERA armour on the T-72A, each block can reduce the penetration of a shaped charge warhead by an average of 55% at 0 degrees, and by up to 80% when angled at 60 degrees. NII Stali claims that it can reduce the penetration power of a typical anti-tank missile like the Konkurs (130mm diameter) by up to 86%, or 58% for a 125mm HEAT shell, or up to a whopping 92% for lower velocity shaped charge warheads like the one on the 66mm LAW.

According to NII Stali, the percentage of the tank surface covered by Kontakt-1 is as follows:

Turret Hull Front Hull Sides

The weight of the Kontakt-1 blocks over the three individual surfaces are as follows:

Turret Hull Front Hull Sides
422 kg
288 kg
300 kg

The total weight of the Kontakt-1 set for the T-72B is 1,310 kg, 110 kg more than on the T-72A. There are 46 blocks on each side skirt, 63 blocks on the upper and lower glacis plates, and 72 blocks on the frontal arc of the turret and turret roof.

Unlike the T-72A, the arrangement of blocks on the T-72B provides better coverage at the cost of somewhat reduced effectiveness. As mentioned earlier, the blocks on the T-72A are mounted on special metal frames to form a wedge shape around the circumference of the turret cheeks, allowing the reactive armour to perform up to its maximum potential at a high obliquity. However, this arrangement left the turret ring and much of the mantlet area unprotected, a problem which can be considered solved in the T-72B, even though the blocks are less effective due to a lack of sloping. The Kontakt-1 blocks on the turret ring of the T-72B are mounted on rails and are easy to remove. They are often removed when not needed during peacetime so that the driver is not obstructed when his head is out of the hatch while driving. The photo below shows Kontakt-1 on the turret ring and on top of the gun, but nowhere else.

As stated before, Kontakt-1 blocks are often left unfilled when the tank is used on exercises and when it is placed into long term storage. Sometimes, the blocks are taken off altogether.


Kontakt-5 is classified as integrated reactive armour as opposed to add-on reactive armour like Kontakt-1. Being somewhat heavier and more powerful than Kontakt-1 per block, it was not possible to simply bolt the Kontakt-5 reactive armour panels onto the tank, thus necessitating the installation of the panels onto the base armour by welding. The only way to remove them is to cut off the plates at the weld seams, so it is only possible to remove the panels if the tank is at a depot or if a BREM-1 recovery vehicle is available. If a panel is spent, a new one is simply welded in its place. A complete set of Kontakt-5 weighs 1.5 tons, most of it from the heavy steel plating. A T-72B equipped with Kontakt-5 would weigh around 46 tons dry, and a combat-loaded T-72B obr. 1989 should weight between 48-49 tons based on the verified information from Uralvagonzavod that the T-72B3 weighs 48.8 tons. NII Stali claims that the total area of the tank protected by Kontakt-5 from a frontal view is 55%. The hull is 45% covered when viewed from a sideways angle of 20 degrees. The turret is 45% covered at a sideways angle of 35 degrees. Atrocious as it may seem, these figures still do not tell the whole story; a large part of the unprotected area is at the turret ring area and gun mantlet, which is the center mass of the tank. This problem was partially solved in the T-72B3 modernization where another three Kontakt-5 panels was installed on the turret, but even so, the coverage is not particularly good as you can see from the photo below. While it may have been a novel application for explosive reactive armour, Kontakt-5 was certainly not a panacea.

Although Kontakt-5 is also used on the T-80U, there are actually a few distinct variants of Kontakt-5 that all differ in the exact construction but operate on the same basic unifying principle. The Kontakt-5 reactive armour package used on the T-72B obr. 1989 is unique to the T-72.

There is sufficient information in the public domain for us to simulate the interaction between Kontakt-5 and many modern long rod projectiles. Equipped with the theoretical models designed by Dr. Manfred Held and H.S Yadav, among others, it would be rather simple. However, that is not the aim of this examination. Instead, the aim is to gain an accurate understanding of Kontakt-5, its many intricacies, and the paths taken by Soviet engineers more than 30 years ago.


Kontakt-5 was designed to use 4S22 explosive elements, as opposed to 4S20 which was used in Kontakt-1. 4S22 is an improvement over 4S20 in every way. Chemically, the PVV-12M plastic explosive used in 4S22 is composed of 85% RDX and 15% inert phlegmatizing agent, similar to 4S20. However, 4S22 retains its ductility at a slightly expanded temperature range of -50°C to +50°C, and 4S22 has a higher flash point of 300°C, making it more resistant to napalm.

The manual for the tank and NII Stali both state that the total number of 4S22 explosive elements installed in the T-72B obr. 1989 is 240 pieces. Out of 1.5 tons, the total mass of the 4S22 explosive elements is only 329 kg so the heavy welded flyer plates of the reactive armour panels comprise the majority of the mass.

According to an NII Stali information placard shared by Alexey Khlopotov, 4S22 is identical to the 4S20 explosive element in dimensions, measuring in at 252x130x10 mm. The mass of the complete explosive element is 1.37 kg, while the mass of the explosive charge alone is 0.28 kg. The PVV-12M explosive charge has a similar composition as PVV-5A but is denser and more powerful. PVV-12M has a density of 1.5 g/cc and a detonation velocity of 7.76 km/s. Because PVV-12M has a higher detonation velocity compared to PVV-5A and a bigger mass, 4S22 has an explosive power equivalent to 0.33 kg of TNT. The thickness of the sandwich layers are assumed to be the same as the 4S20; a pair of 2.3mm steel plates sandwiching a 5.4mm plastic explosive interlayer. Using the Gurney equation for symmetrical sandwiches, the velocity of the plates of the 4S22 element at the moment of detonation should be around 1.258 km/s.

The anti-shaped charge capabilities of 4S22 on its own was demonstrated on a TV Zvezda show called "Военная приемка" ("Military Acceptance"), in episode "Т-90. Бункер на колесах" (T-90: Bunker on Tracks). The screenshot below, taken at the 18:40 mark of the show, shows the experimental set up used for the demonstration. The 60 kg armour plate used as the target is claimed to be equivalent to the steel used in the T-90 tank, and the so-called "dynamic element" is claimed to weigh 1.37 kg, which means that it can only be 4S22. The shaped charge is similar to the one previously used at the 18:17 mark of the show, which was shown to be capable of penetrating around 200mm of the same type of armour plate in LOS thickness. The targets are angled at 60 degrees.

As you can see in the screenshot below, an imprint of the forwards (in-pursuit) flyer plate is left on the armour plate and the penetration of the shaped charge jet is reduced to practically nothing. The fragments of the particulated jet only gouge the plate and crater the surface.

Evidently, a shaped charge with around 200mm of penetration into RHA can reduced to just a few millimeters by 4S22 at a 60 degree obliquity. This is only a demonstration, however, and not necessarily a scientific one. This is definitely not a demonstration of how much Kontakt-5 can reduce the penetration of a shaped charge, as the 4S22 elements are arranged differently in Kontakt-5.


Part of the T-72B3 obr. 2016 modernization involved the replacement of the 4S22 elements in the built-in Kontakt-5 panels with the somewhat newer 4S23 elements originally developed for the Relikt reactive armour system. 

A more detailed examination of 4S23 and its function in Kontakt-5 will be available in the future.


Kontakt-5 is much more complex than commonly thought. The typical description of Kontakt-5 paints it as a head-on flyer plate design using a thick and slow flyer plate, and that the design is very inefficient as a consequence. Other descriptions mention the high thickness of the flyer plate as a positive thing, as it would "feed more armour into the path of the penetrator" as it passes through the armour, but we already know that that is largely incorrect. In reality, Kontakt-5 propels a total of three flyer plates head-on towards a projectile in a timed sequence to enable the ERA to resist both shaped charge jets and KE penetrators with minimal compromises.

Dr. Manfred Held conducted exhaustive studies on impact initiation, and his works in this subcategory of ballistics are relevant to us now in our examination of Kontakt-5. It is known that when a projectile or shaped charge jet passes through a barrier placed over an explosive charge, a highly energetic burst of spall and fragments is generated at the back surface of the plate and travels towards the explosive charge, thereby initiating detonation. According to a summary on page 8 in "The Legacy of Manfred Held with Critique", Dr. Held observed that an explosive charge directly in contact with a barrier was less easily initiated by a jet impact than a one with an air gap between the barrier and the charge. One of the explanations is that an explosive charge placed in contact with the barrier is exposed in a smaller area than the charge with an air gap between it and the barrier, as the air gap considerably increases the spall cone angle and therefore the area of the explosive charge exposed to the spray of spall and fragments emerging from the back surface of the barrier. This is supported by a later study titled "High Explosive Initiation Behavior by Shaped Charge Jet Impacts", where it was reported that an explosive charges with a gap between it and the steel barrier will detonate in the impact initiation mode, whereas an explosive charge in contact with the steel barrier detonates in the penetration initiation mode. This essentially means that when an air gap is present, the explosive charge detonates promptly whereas the lack of an air gap requires the shaped charge jet to penetrate far into the charge to initiate detonation. In practical terms, we can safely say that having an air gap decreases the reaction time of Kontakt-5 to shaped charge jets, and thus improves its effectiveness at disrupting the shaped charge jet.

These results apply for both bare explosive charges as well as cased charges (charges encased in a steel container), so it is applicable to 4S22 explosive elements. Russian publications have mentioned that Kontakt-5 relies on this phenomenon to achieve detonation when impacted by long rod penetrators. The explanation is that spall is readily produced when a thin and brittle plate of high hardness is struck by a projectile as well as during the penetration process. This is validated by infamous Russian expert and pessimist Mikhail Rastopshin, a former NII Stali scientist, who revealed in an article penned in 2005 that the flyer plate of Kontakt-5 has a high hardness and is very brittle. According to Rastopshin, this facilitates the generation of spall and fragments upon impact and penetration by a long rod penetrator, thus ensuring reliable and quick detonation of the explosive elements. However, this does not mean that Kontakt-5 relies exclusively on the spall from its heavy flyer plate to initiate its explosive content.

"Test Setup For Instrumented Initiation Tests" by Dr. Held deals with the effects of projectile mass, projectile velocity and barrier thickness on the initiation threshold of encased explosive charges. Held found that adding a barrier in front of the case explosive charge increased the initiation threshold for projectile velocity compared to a plain cased charge, and increasing the thickness of the barrier increased the velocity threshold. This is completely unsurprising, because there is a pressure threshold that needs to be met or exceeded for an explosive charge to detonate, and the spalling and fragmentation of a barrier would transfer only a portion of the energy of a long rod projectile to the explosive charge. The impact of the projectile itself would invariably generate higher pressure for thin long rod projectiles.

This essentially means that the front flyer plate of Kontakt-5 would be detrimental to the reliability of detonation when compared to exposed 4S22 elements. The presence of an air gap in the design of Kontakt-5 would reduce the velocity threshold necessary to initiate detonation due to the increased spall cone angle, but the net effect would still be an increase in the velocity threshold. However, this would not matter if the velocity threshold is within the range of striking velocities for modern APFSDS ammunition. Needless to say, the specific velocity threshold varies between modern long rod projectiles due to the different characteristics of different rounds, so giving a fixed number to represent all long rod penetrators would be misleading. We can only estimate that this threshold encompasses the striking velocities of typical long rod projectiles at combat ranges of 1.5 to 2 km.

In short, if the velocity and mass threshold of the projectile is sufficient, the air gap between the heavy front plate and the explosive elements in each Kontakt-5 module will have the effect of shortening the reaction time of the system against shaped charge jets and facilitates the action of Kontakt-5 against long rod projectiles. For shaped charge jets, the quicker reaction time enables the flyer plates to intercept the tip of the jet and more of the body, thus preventing much of the hypervelocity tip from continuing into the main armour. This explains the very high reduction in shaped charge warhead performance against Kontakt-5 of up to 80% despite having fewer flyer plates and the use of head-on flyer plates rather than a mix of head-on and in-pursuit flyer plates like Kontakt-1. 

The confined nature of the Kontakt-5 panels also improves their chance of detonation for a wider range of shaped charge jet velocities. Held observed that confined explosives have a lower threshold between detonation and reaction and between reaction and no reaction for significantly lower shaped charge jet velocities. The confinement would be partially from the built-in steel case of the 4S22 element itself and from the thick walls of the Kontakt-5 panels. If the velocity threshold for detonation from spall and fragmentation is not attained, the explosive elements can still be initiated by the direct impact of the projectile.

According to "A numerical study on the detonation behaviour of double reactive cassettes by impacts of projectiles with different nose shapes", the detonation of a double stack of explosive cassettes (elements) by high velocity long rod steel penetrators can be prevented by changing the shape of the nose. The paper is highly relevant to our study on Kontakt-5 as the double stack of explosives modeled in the same arrangement as the 4S22 explosive elements in Kontakt-5, and the mechanisms that dictate the initiation of the explosive charge are explained in full. 

Steel rods were used in the simulations detailed in the study. The striking velocity of the rods was 1800 m/s. The explosive elements used in the simulations were roughly analogous to 4S22. The Composition B filler in the explosive elements in the study had a thickness of 7mm, and were encased in steel walls 2mm thick. Needless to say, the Composition B explosive charge used in the study is not directly comparable to PVV-12M, as PVV-12M has a much higher phlegmatizer content and is therefore much less sensitive to impact, but detonating 4S22 elements by direct impact from long rod projectiles is still completely plausible.

The main method of initiating detonation is by shock. The study "The Shock-to-Detonation Transition in Explosives - an Overview" gives a concise explanation of this phenomenon. In short, the impact of an object on the surface of an explosive charge produces a shockwave. The shockwave accelerates deeper into the explosive charge and detonation occurs after the shockwave has travelled a certain depth into the charge, and this depth is called the run distance to detonation. The run-to-detonation differs between explosives, but as a rule, the thickness of the charge must be equal to or greater than the run-distance of the explosive in order for the charge to detonate by this method.

It also found that reducing the velocity of the flat-nosed rod from 1800 m/s to 1700 m/s effectively prevented the initiation of a run-up detonation, but detonation was still achieved from the reflection of the shockwave of impact from the backplate of the explosive element and the build-up of pressure from the compression of the explosive material against the backplate. This effect is undoubtedly reinforced in Kontakt-5 by the placement of the 4S22 elements flush against the surface of the glacis plate, and by the fully contained nature of each Kontakt-5 panel. As detailed in the summary of the paper, the backplate effect is independent of the run-up detonation, and is therefore also independent to the spall effect. It was possible to avoid this effect with hemispherical noses as the build-up of pressure was followed by the displacement of the pressurized explosive material away from the path of the rod, so the pressure was insufficient to initiate detonation. The fact that a small reduction in velocity from 1800 m/s to 1700 m/s was enough to prevent detonation of the Composition B charge by the conventional run-to-detonation method indicates that PVV-12M would most definitely not be initiated even at higher striking velocities due to its low sensitivity. This essentially leaves the shockwave reflection effect solely responsible for initiating Kontakt-5 in the case of a failure to detonate from the spall effect.

There is less unpredictability with shaped charges, as the incredibly high pressure imparted onto an explosive charge upon impact and during penetration by a typical shaped charge jet practically guarantees detonation under any condition. The spall effect from the heavy flyer plate of Kontakt-5 merely reduces the reaction time of the system.

Although Rastopshin is entirely correct in his suggestion that Western long rod projectiles may defeat Kontakt-5 via their relatively low striking velocity, decreasing the velocity of the projectile has a negative effect on its penetration power, and this limits its ability to defeat the base armour. As such, the only truly viable methods of defeating Kontakt-5 are to have a special tip or to have a segmented penetrator. Indeed, modern APFSDS shells fielded by the major leaders in the field - Germany, Israel, U.S.A - are fired at their optimal velocities to maximize their penetrative power, rather than at velocities that are low enough to bypass reactive armour. The best example of this is the introduction of the Rh 120 L55 cannon in Germany to increase the muzzle velocity of existing 120mm APFSDS ammunition when fired from upgraded Leopard 2 tanks. Even the M829A3 - which is rumoured to be aimed at defeating Kontakt-5 via its low velocity of 1670 m/s - almost certainly has a relatively low muzzle velocity because it is closer to the optimum velocity for its particular alloy of depleted uranium. The graph below, created by Willi Odermatt (a well known scientist specializing in terminal ballistics), shows the relationship between the penetration depth of a generic long rod penetrator for generic alloys of depleted uranium, tungsten alloy and steel. As you can see, the optimum velocity for depleted uranium penetrators is generally lower than tungsten alloy.

From this, it is apparent that the defeat of Kontakt-5 by low velocity impact is currently not being pursued by Western militaries, and is not a feasible solution for the future. Indeed, the renowned German military expert, lecturer and author Rolf Hilmes stated that DM53 has a three-part penetrator and is specially designed to deal with composite and reactive armour, and it is reported that the DM53 is optimized to be fired from the L55 cannon, which allows it to attain a muzzle velocity of 1752 m/s. While this is not the optimum velocity for tungsten alloy long rod penetrators, that is only because the optimum velocity is unattainable with current generation tank guns.

Nevertheless, it is clear that the detonation of the explosive elements in Kontakt-5 is not always guaranteed. Special nose shapes on APFSDS projectiles may be able to reduce the pressure exerted on the explosive charge or prohibit the complete detonation of the charge. Russian engineers were fully aware of this fact, as proven by evidence of numerous experiments conducted in the USSR aimed at penetrating explosive elements without detonating them.

During the development of Kontakt-5, Soviet engineers spared no expense to find ways to overcome their own brainchild. According to Rastopshin, experiments have confirmed that long rod projectiles travelling at low velocities do not cause detonation of reactive armour from barrier spall. Another effort was aimed at modifying existing high velocity APFSDS rounds to defeat the armour. One of the successful solutions took the form of a protruding steel probe of small diameter installed on the tip of a specially modified 3BM-22 "Zakolka" shell.

As part of our analysis, we will once again refer to "Test Setup For Instrumented Initiation Tests" by Manfred Held. The paper deals with the effects of projectile mass, projectile velocity and barrier thickness on the initiation threshold of encased explosive charges. From his findings, we can surmise that the function of the small steel probe was to avoid the detonation of the explosive elements by presenting a small impact area, whereby the relative mass of the projectile impacting the front plate of the Kontakt-5 panel is minimal, thus preventing detonation from the spall effect. The subsequent detonation of the explosive elements from the direct impact of the rod itself might also be prevented in the same manner if the small steel probe could survive the penetration of the front plate. This is merely speculation, of course, so please do not take this explanation as concrete fact.

As the double stack of 4S22 explosive elements is pinned to the backplate of the reactive armour module, a reliable detonation of the explosive elements should be expected from a long rod projectiles with flat noses. Examples of such projectiles include the 3BM-32 "Vant", 3BM-42 "Mango", DM13 (120), DM23 (120), DM33 (120), DM23 (105), DM33 (105) and many more, including older projectiles. Information on the behaviour of projectiles with stepped tips like the M111 and the M829A2 is not easily found in the public domain, but it is known that stepped tips are used to dampen the shockwave travelling down the rod at the moment of impact to reduce the severity of the damage to the rod. How this affects the reliability of detonation is not known, but it is probably safe to assume that it is negligible, since the military-scientific industry of the USSR had access to captured M111 shells during the development of Kontakt-5.

From what we now understand of Kontakt-5 and the methods of overcoming it, it should be immensely clear that there are no modern long rod projectiles currently in use that are specifically aimed at defeating Kontakt-5 by low velocity or by low contact area, and this generalization includes some of the most modern ammunition such as M829A3 and DM53. M829A3 overcomes Kontakt-5 primarily via a two-part segmented penetrator with a steel segment at the tip, and DM53 overcomes Kontakt-5 via a segmented penetrator as well, albeit with three-parts. It should also be clear now that the M829A2 has no special provision for defeating Kontakt-5, despite widespread rumours that it was designed as a special countermeasure to Kontakt-5. A simple comparison of muzzle velocities between the four members of the M829 series confirms this: the M829 travels at 1750 m/s at the muzzle, while the M829A1 has a greatly reduced velocity of 1575 m/s, and the M829A3 is only slightly slower at 1555 m/s. The M829A2 had the second highest muzzle velocity at 1675 m/s, behind only the original M829.

With the initiation of the reactive armour being all but unavoidable, the objective of M829A2 was to minimize the damage taken from the flyer plate of Kontakt-5. To that end, M829A2 was made from a new depleted uranium alloy that was more ductile and possessed higher yield strength, making it more resistant to bending and fracturing. That said, this approach cannot be described as a special provision to deal with Kontakt-5. The need to improve the ductility and yield strength of heavy alloy long rod penetrators had always been a requirement, and the use of metal jackets over long rod penetrators like on the Soviet "Mango" projectile is a consequence of an inability to create a sufficiently ductile and strong alloy. Having already discussed the protection mechanisms employed by the various forms of composite armour employed on the T-72 throughout its history, it is plain to see that a long rod projectile that is more resistant to bending and fracturing would also be perform better against NERA and oblique spaced armour.


It has been shown by Dr. Manfred Held that the primary mechanism of long rod projectile defeat by heavy reactive armour is the transfer of momentum from the flyer plates to the projectile. The desired effect is the deflection of the projectile from its original direction of travel and in the disruption of the shape of the projectile, whether it be by fracturing it, shattering it, bending it, introducing yaw or by cutting it into fragments. In order to achieve this, a sufficiently thick and heavy flyer plate must be used against the long rod projectile, and the emphasis is on the mass of the flyer plate and not the velocity. 

Kontakt-5 modules on the hull rely solely on the action of head-on flyer plates to defeat attacking projectiles, whereas the modules on the turret and on the side hull are designed to send flyer plates in both directions. We have already examined the peculiarities of forward moving (in-pursuit) and backward moving (head-on) flyer plates, and from what we know, it is quite clear that head-on flyer plates are much less efficient than in-pursuit plates. There are a few general rules of improving the performance of flyer plates; increasing the mass of the plate; increasing the velocity of the plate; and increasing the angle of the plate, and any combination of the three.

The efficiency of the modules on the hull are increased through a combination of all three methods to a certain extent, but not without a few negative consequences. The high angling of the Kontakt-5 modules for the hull is guaranteed by the good 68 degree slope of the upper glacis, so there were no compromises that needed to be made here. However, the heavy flyer plates of the hull modules are conspicuously thicker than the plates on the turret modules, and this led to an increase in the mass of the plates. To accelerate this heavy mass to a high velocity, as many as twelve 4S22 explosive elements are used in each module. The twelve explosive charges have a combined explosive power equivalent to 3.36 kg of TNT. The blast has a small contribution in the reduction of the penetration of a shaped charge jet, of course, but the side effect is that external equipment on the tank may be destroyed or damaged and personnel both inside and outside the tank may suffer injuries. There is even a possibility of a flyer plate impacting the gun barrel under the right conditions, but I digress.

Upon detonation, the thin front plates of the first and second explosive elements are propelled at high velocity head-on against the direction of travel of the jet. Estimating the velocity of the thin flyer plates of the 4S22 elements on their own is straightforward, but predicting the velocity of the heavy flyer plates requires a few more steps, because there is momentum transfer from the thin flyer plates of the 4S22 elements to the heavy plate. A simplified model of Kontakt-5 will be used for our calculations.

Due to the very high velocity of the thin flyer plates of the 4S22 elements, it is assumed that they will fuse to the surface of the heavy flyer plate upon impact, so we can classify it as an inelastic collision. The small gap between the 4S22 elements and the heavy flyer plate enable the thin flyer plates to is inconsequential due to the closed nature of Kontakt-5 panels. This enables us to conveniently add the mass of the thin flyer plates to the mass of the heavy flyer plate without other considerations to determine the final velocity of the heavy plate. This assumption is validated in "Momentum Transfer in Indirect Explosive Drive". The heavy flyer plate will be driven by the expansion of gasses only after it impacted by the thin flyer plates. Due to the conservation of energy, the momentum of the thin flyer plates and the heavy flyer plate cannot be calculated as separate entities and then added together, because there is a finite source of energy. Therefore, the mass of the heavy flyer plate alone cannot be plugged into a Gurney equation to obtain its velocity. We must add the mass of the thin flyer plates to the heavy flyer plate, and treat the resultant final mass as a single entity.

According to Russian academicians and experts, the thickness of the heavy flyer plates is 15mm. This is confirmed by the drawings from the T-72B obr. 1989 technical manual. Beyond that, we can figure out the mass of the heavy flyer plate with minimal guesswork by simply adding up the widths and lengths of the 4S22 elements behind each plate to get the approximate surface area, and then multiply that with the approximate thickness of the heavy flyer plate. The simplest modules to calculate are the modules along the top row on the upper glacis. Since there are six 4S22 elements behind the plate, the plate has a surface area of at least 0.19656 sq.m. There is a small gap between the explosive elements and the partitions onto which the flyer plate is welded, and if we make the assumption that these small details add just under 1 cm to the length and width of the plate, then we can round off the surface area to 0.2 sq.m. Using these figures, we get 23.55 kg.

The mass of the 4S22 thin flyer plates is readily determined by simply subtracting the mass of the explosive charge (0.28 kg) from the total mass of the explosive element (1.37 kg), and then dividing that by two. Six flyer plates gives us 3.27 kg, and another twelve flyer plates gives us 6.54 kg for a total of 9.81 kg. Therefore, the final mass of the heavy flyer plate is 33.36 kg.

The mass of the explosive charge will be 0.28 kg multiplied by twelve, giving us 3.36 kg. This is rather low compared to the final mass of the heavy flyer plate after factoring in the mass of the thin flyer plates. It is explained in "Flyer Plate Motion by Thin Sheet of Explosive" by H.S Yadav that at very low C/M ratio, the calculated velocity of the flyer plates using the Gurney model is at variance with experimental results. This is supported by "Gurney Energy of Explosives: Estimation of the Velocity and Impulse Imparted to Driven Metal", which mentions that the recommended restrictions for the Gurney model is 0.2 < M/C < 10 (p. 11). Fortunately, the M/C ratio of the final mass of the heavy flyer plate to the explosive charge is 9.92, placing it just within the stated restrictions. Thus, the results from the Gurney model can be considered reasonably accurate.

It is stated on page 11 in "Gurney Energy of Explosives: Estimation of the Velocity and Impulse Imparted to Driven Metal" that a small gap between the flyer plate and the explosive charge will result in very little decrease in plate velocity, so for all intents and purposes, we will assume that the heavy flyer plate is in contact with the explosive charge, and this means that the Gurney model is applicable. The loss in plate velocity will be considered negligible, but it will be represented in our calculations by rounding down our result to the nearest ten.

Now that we have ascertained all of our variables, we can use the Gurney equation for an infinitely tamped sandwich in our calculations. The reasoning is that even though the 60mm backing plate (the front plate of the base upper glacis armour) is only four times the thickness of the 15mm heavy flyer plate, the backing plate is affixed to a rigid structure - the hull - and the Kontakt-5 panel only occupies a relatively small area of the upper glacis, so the base armour plate does not experience any acceleration from the blast. Since the velocity of the backing plate would be zero, it has the same behaviour as an infinitely thick tamper plate, and it will be treated as such.

Plugging our figures into the Gurney equation, we get 807.47 m/s. Rounding it down to account for the air gap between the flyer plate and the explosive charge, we get 800 m/s.

Despite our precautions, these are still only approximations. The Gurney model used for this calculation is for an unenclosed explosive charge, so the true velocity of the heavy flyer plate could be slightly higher because no energy is lost from the system. An enclosed system like Kontakt-5 would force all of the propulsive energy of the explosive charge to be focused on the single flyer plate.


Due to the high strength of long rod penetrators compared to shaped charge jets, the interaction between it and the heavy flyer plate of Kontakt-5 is generally not the same. One similarity is that both long rod penetrators and shaped charge jets are only eroded while penetrating the heavy flyer plate prior to the detonation of the explosive charge. During the movement of the flyer plate, the interaction can no longer be described as erosion, so by definition, the notion that reactive armour places "more material in the path of the projectile to penetrate" is immediately demonstrated to be false. Rather, during the stage where the flyer plate moves laterally against the penetrator, whether it be a rod or a shaped charge jet, the interaction is better described as the sliding of the plate against the penetrator. Throughout the sliding action, lateral forces are imparted on the penetrator, and the resistive force cuts a crater into the plate.

The operating mechanism of flyer plates against long rod penetrators is summarized on pages 60-61 in "Interactions Between High-Velocity Penetrators and Moving Armour Components". The effect of heavy reactive armour on a long rod penetrator during the penetration of the front plate and before the detonation of the explosive charge is essentially the same as simple spaced armour.

"KH-instabilities do not occur in the case of an LRP interacting with reactive armour. In this case, the high strength of the projectile material and the low projectile velocity relative to that of an SC jet prevent the generation of instabilities. Instead, the abrupt change in pressure at the exit of the plate gives rise to fracture of the projectile."

So before the explosive charge even reacts, the projectile is fractured as a result of perforating the heavy front plate - or rather, the tip is fractured. The mechanisms of spaced plates has already been examined in the section on the T-72B obr. 1985. Navigate to that section for more information. There is a possibility that the heavy flyer plate may also partially condition the penetrator to facilitate more reliable detonation by blunting the tip. As we have seen from the studies presented earlier, long rod penetrators with a flat tip will detonate explosive elements most consistently; but I digress. The vast majority of the effect of Kontakt-5 comes from the motion of the flyer plate against the penetrator, and the paper clarifies the ramifications of the head-on direction and the thinness of the flyer plate.

"The positive pressure gradient and longer interaction time make forwards moving plates more effective than backwards moving plates. Besides from the direction of motion of the plate, the most significant plate parameter for effectively disturbing the projectile is the thickness. Increased plate thickness results in substantial increases in rotation, translation, bending, length reduction and fragmentation of the projectile.

For fractures to occur in the projectile, the plate velocity has to be relatively high, 300 m/s for a plate thickness of one projectile diameter and 200 m/s for a plate thickness of two projectile diameters (only forwards moving plates). Lower projectile velocity results in longer interaction time which increases the effect of the moving plate on the projectile. The experiments also indicated increased effect at higher projectile velocity which has not been explained in these studies."

As mentioned in the passage, the longer interaction time obtained from a forwards moving (in-pursuit) flyer plate is beneficial and vice versa. Not only is less force imparted on the rod, but less of the rod is affected. For Kontakt-5, this means that only the front part of a long rod penetrator will be affected. For this reason and many others, it is apparent that it is not the most efficient arrangement. It would be more efficient to have a single in-pursuit flyer plate of increased thickness, but this is not feasible for tank armour due to the limited space, and the best compromise would be bi-directional flyer plates, which would be a simple description of Relikt. As such, the focus is on maximizing the effectiveness of the head-on flyer plate used in Kontakt-5.

According to "The Break-up Tendency of Long Rod Projectiles", the ductility of high strength tungsten alloy rods appears to not have much bearing on the bending of the rod from interacting with oblique flyer plates, but brittle high strength rods were unsurprisingly more prone to fracturing or fragmenting than more ductile high strength rods. Also, it is noted in the conclusion of the paper that the tendency for a long rod projectile to shatter increases with the strength, thickness and obliquity of the flyer plate, which is quite obvious, but more interestingly, it is stated that high velocity is beneficial for head-on flyer plates while high velocity is disadvantageous for in-pursuit flyer plates. In addition to that, it is stated that the velocity of the flyer plate and projectile have a bigger influence on the tendency of the projectile to shatter than the other parameters, which includes the thickness of the plate. As a rule, increasing the velocity of the flyer plate is more advantageous than increasing the thickness, but having a thick flyer plate travelling at high velocity would obviously be the best of both worlds. Considering the limitations of the practical application of heavy reactive armour on tanks, the high velocity of the flyer plate of Kontakt-5 appears to be the correct choice.

Although it is considered a heavy flyer plate, the 15mm front plate of Kontakt-5 actually has a rather low thickness when contrasted with the diameter of the long rod projectiles likely to be used against it. Even adding on the thickness of the thin flyer plates of the 4S22 explosive elements, the final thickness is only around 21mm to 22mm thick, which is still slightly less than the diameter of a typical heavy alloy long rod penetrator. Nevertheless, the combination of high velocity (800 m/s) and relatively high thickness would make the flyer plate of Kontakt-5 very effective at bending and fracturing a long rod penetrator. The famous photo below appears to demonstrate the effect of a head-on flyer plate against a long rod penetrator:

The penetrator in the photo was moving from right to left. The fragments appear to be a mix of pieces from the fragmented penetrator and spall from the flyer plate, which seems to be the black blob to the right of the rod. The downward curl of the damaged rod is clear evidence that the plate that intercepted it was travelling head-on towards the plate at an oblique angle, representing the flyer plate of Kontakt-5. It is evident that the plate was propelled independently of the rod and intercepted the rod at a predetermined point where the photograph was taken, so it does not fully represent the mechanism of Kontakt-5 where the penetrator impacts the plate and initiates detonation. The heavy fracturing experienced by the rod in the photograph is probably a consequence of the imperfect tungsten alloys available in the USSR at the time (early 80's).

Beyond fracturing and bending the rod, the interaction also induces yaw into the long rod penetrator. According to "Experimental and Numerical Simulation Analysis of The Impact Process of Structured KE Penetrators Onto Semi-Infinite and Oblique Plate Targets" even 1 degree of yaw can cause a tungsten alloy long rod penetrator to break apart in half after passing through a thick oblique spaced plate. The study is particularly useful for us as it explores the effects of a positive yaw angle on the rod, which is compatible with a scenario where a long rod penetrator passes through Kontakt-5. The heavy flyer plate of Kontakt-5 would impart lateral forces on the rod in an upwards direction, and thus generate positive yaw. However, the oblique plate used in the study was angled at only 60 degrees. It is known that increasing the angle of an oblique plate exacerbates the damage experienced by a yawed long rod penetrator. A long rod penetrator with 1 degree of yaw impacting a thick spaced plate angled at 68 degrees would be highly destructive towards the rod.

The picture below deals with the modelling of fragmentation and spall from the perforation of an oblique steel plate by a tungsten long rod penetrator as well as the damage sustained by the plate and rod. It is taken from the study "Behind Armor Debris Computations With Finite Elements and Meshless Particles". As you can see, a long rod penetrator perforating a thin oblique plate at 60 degrees will suffer minimal fracturing and come out largely unscathed except that it loses a chunk of its tip, but when the same penetrator strikes an oblique plate with 1 degree of yaw, the rod is broken into two large pieces when it exits the plate and the rod remnants are yawed at -1.5 degrees. When the penetrator strikes the same plate but with 2.4 degrees of yaw, the rod is broken into three large pieces when it exits and the remnants are yawed at -0.7 degrees. When we include the loss of the rod tip, then the rod has broken into three and four pieces for the two cases respectively.

The method of defeat against shaped charges is the same as any reactive armour or non-energetic reactive armour (NERA), and that has already been covered in the earlier review of Kontakt-1 and the armour of the T-72B. The efficiency of the Kontakt-5 design is not high against shaped charge jets, but the system is still quite effective by virtue of brute force. As explained before, Kontakt-5 sends three flyer plates head-on towards the shaped charge jet: two thin plates from the casing of the 4S22 explosive elements (one 2.3mm plate and one 4.6mm plate), and the single heavy flyer plate, 15mm thick, but the majority of the effect stems from the heavy flyer plate. Taken together, Kontakt-5 can be as effective as Kontakt-1 against shaped charges.

A quantitative analysis of Kontakt-5 has not been done yet, although it is definitely possible to do so now, having been equipped with a full understanding of the mechanisms at play. If any researcher or amateur enthusiast would like to assist me in conducting quantifying the effect of Kontakt-5 on generic heavy alloy long rod penetrators, you are welcome to contact me (see "Contacts" page).

NII Stali claims that the reduction in penetration for subcaliber shells (long rod penetrators) to be equivalent to 250mm RHA, but describing it as a solid figure is both illogical and misleading. All other materials published by NII Stali state instead that Kontakt-5 decreases the penetration of subcaliber shells by 1.2 times. Some publications describe the reactive armour as being able to reduce the penetration of a generic long rod projectile by 20% to 35%. Regardless, all given figures were deliberately left vague, as the actual effect of Kontakt-5 depends on the penetrator in question. Long rod projectiles with a very high L:D ratio will not be affected in the same manner as a short and stubby long rod projectile with a very low L:D ratio, and the strength of the rod in question makes a tremendous difference as well. Penetrators with a composite construction like the American M735 or Soviet 3BM-22 will behave even more differently. A "Журнал Техника и Вооружение" article claims that Kontakt-5 reduces the penetrating capability of cumulative jets by a minimum of 50% to a maximum of 80%, while NII Stali claims that Kontakt-5 reduces the penetration of shaped charges by 1.9 to 2 times.


Kontakt-5 blocks are mounted in a clamshell layout around the frontal arc of the turret. There are total of 120 4S22 explosive elements installed in the turret, 46 on the upper half and 32 for the lower half. The other 42 explosive elements belong to the ERA blocks on the roof, two elements per block.

There are three different sizes of blocks used for the front of the turret, and the manual gives instructions on how to arrange the 4S22 explosive elements inside each type. All of the blocks have two layers of 4S22 elements arranged crosswise.

  • The most numerous type are the blocks for the lower half of the clam shell; 8 of these square blocks are installed on the turret. Each block contains four 4S22 explosive elements. The first layer of 4S22 is laid into the slot horizontally, and the second layer is laid vertically.
  • The second most numerous type are the blocks for the upper half of the clam shell, marked (1) on the diagram; 7 of these rectangular blocks are installed. Each block contains six 4S22 explosive elements. The first two elements of the first layer are laid vertically, and then another cell is added horizontally on top. The second layer is laid horizontally.
  • The third type is a squarish block - marked (2) on the diagram - for the upper half of the clam shell. There is only one example of the third type on the turret; it's the first block to the left hand side of the cannon. The first layer is laid vertically, and the second layer is laid horizontally.

Unlike the panels on the hull, the Kontakt-5 panels on the turret are bolted onto the turret and not welded. This makes it extremely easy to remove the panels if necessary.

As the photo below shows, the cover plate at the end of each reactive armour panel can be unbolted to remove the explosive elements inside. The explosive elements are typically salvaged from destroyed tanks like the one in the photo below and sold on the black market. Photo from


Kontakt-5 blocks cover a little over two thirds of the upper glacis. There are total of 84 4S22 explosive elements installed on the upper glacis; 48 in the top row of panels and 36 in the bottom row. The top row of panels all have the same rectangular design, housing twelve 4S22 elements each. The panels at the corners of the bottom row have a narrower rectangular design with only eight 4S22 elements each. The two panels at the center of the bottom row have an unusual L-shaped design with ten 4S22 elements each.

The photo below (Photo credit to Bellingcat) shows the Kontakt-5 blocks on the upper glacis of a catastrophically destroyed T-72B3 in Ukraine. The access panels for all eight reactive armour panels have been removed, and you can see the explosive elements within.

And here is a photo of an intact tank.

Inserting explosive elements into the two Kontakt-5 panels at the bottom corners is done by unbolting the cover plate at the bottom, laying the 4S22 elements into a tray and sliding it into the panel, as demonstrated in the two screenshots below (Screenshots taken from RT Documentary show "Tanks: Born in Russia (E9)").

Once filled, the panels are simply bolted shut. A tray is used to make it easier to keep the elements from sliding out while bolting the cover plate back on. The tray has no active role in the design of the reactive armour.


There are three Kontakt-5 panels located on either side of the hull. These are a type of explosive flyer plate. They use the same 4S22 explosive elements as the Kontakt-5 plates on the front hull and turret. The side panels provide coverage for the fighting compartment in a 35 degree frontal arc, as illustrated in the photos below:

The panels are mounted on special brackets bolted to the steel screens over the fenders. It is possible to flip the panels up to access the suspension. To do this, the panels are simply lifted upwards until the hole (shown in red in the photo below) is above the hinge, and then the panels are locked in the upward position by inserting a retaining pin through the hole.

The side panels contain six 4S22 explosive elements each. There are three reactive armour panels on each side of the hull, each with six 4S22 explosive elements for a total of 36 elements on both sides of the tank. Each element is laid flush onto the sheet steel tray, and held in place by rubber studs embedded onto the front plate for spacing. This creates an air gap between the front plate and the explosive elements in the same manner as the Kontakt-5 plates on the front hull and turret. This is illustrated in the diagram below, which has been coloured for easy identification. The areas marked in green denotes steel, while the light blue area is the rubber spacer stud and light red marks the explosive elements. The rubber stud is marked with a (7), the explosive element is marked with a (6), and the front plate is marked with a (5).

An opened panel can be seen in the photo below. Note the six protruding studs on the front plate, corresponding to the positions of the six 4S22 explosive elements. The strip of empty space down the middle of the sheet steel tray is very obvious in the photo.

It is not clear what the strip of empty space is meant for, but it should have no negative effect on the consistency of detonation if the panel is struck from the frontal arc of the tank, because the spray cone of high energy spall and fragments generated by an impacting long rod projectile or a shaped charge jet should be wide enough to detonate at least one of the explosive elements. The only way for an attacking projectile or warhead to slip through would be if it was fired perpendicular to the side of the tank and struck the panel squarely in the middle, but in that case, the reactive armour panel would have very little effect even if it managed to detonate as some obliquity is required for the flyer plates to be effective. 

Like the reactive armour plates on the front hull and turret, the front plate of the side panels is 15mm thick. The sheet steel backplate is just under 3mm thick. However, the 2.3mm-thick steel casing of the 4S22 explosive elements also contributes to the overall thickness of material present in the panel. Since the elements lie flush to the sheet steel tray, the backplate of the system would have an effective thickness of 5.3mm, and the front plate will have a final thickness of 17.3mm after the explosive elements detonate and the front flyer plate of the explosive elements fuse to the back surface of the front plate of the panel.

It is quite obvious that the side panels are not as powerful as the ones on the hull or even the ones on the turret, despite the fact that the panels on the turret also house up to six 4S22 explosive elements, because the turret panels are of the same thickness (15mm) but are much smaller. This deficiency is not entirely counteracted by the bi-directional design of the side armour panels, as the turret panels are bi-directional as well. Rather, the overall effectiveness of the armour for the sides is heavily dependent on the very large air space between the side panels and the side of the hull. According to figures obtained from the T-72B obr. 1989 technical manual, the perpendicular space between the side of the hull to the end of the side panels is exactly 760mm. After subtracting the approximate thickness of the panel itself, the space should be around 720mm. At a 30 degree obliquity (viewing the side of the hull), the air gap is therefore 1440mm wide between the Kontakt-5 side panels and the side hull armour. 

The large air gap gives ample time and space for a long rod projectile or shaped charge jet to disintegrate before striking the main armour. This is particularly important for long rod projectiles, as the penetrator is given much more time to yaw before it strikes the main armour as compared to the reactive armour panels on the front hull and turret, where a 1-degree yaw would be expected. A heavily yawed long rod penetrator has a greatly enhanced likelihood of shattering on impact with the side hull armour.

More of these side panels can be installed if desired.


The T-72B obr. 1989 had unique hexagonal Kontakt-5 reactive armour bricks installed on the turret roof. Each brick is mounted directly to the cast steel turret roof on a pair of metal spacers. The bricks sit directly atop the anti-radiation cladding, with no air gap in between. This variant of Kontakt-5 roof armour is only one of several variants, as shown on Andrei Tarasenko's btvt.narod. A steel barrier is installed around the forward perimeter of the roof bricks to provide protection from bullets.

One of the biggest mysteries concerning the T-72B obr. 1989 is why the designers felt the need to use a new design of roof ERA blocks with interlocking geometry that would have allowed the blocks to cover the entire turret without leaving any gaps, yet install so few of the blocks that the turret roof is less covered than the original T-72B with Kontakt-1.

Unlike the Kontakt-5 panels on the rest of the tank, these bricks have an entirely unique design, strongly indicating that they are not meant for the same purpose as the others. The bricks are composed of a front plate and a thin sheet steel tray, into which four alternating layers of inert lining and 4S22 explosive elements are inserted. The diagram below, taken from the manual and modified, shows a cross section of the bricks. The areas marked in green denotes steel, while the areas in light blue mark the inert liners and light red marks the explosive elements.

However, the diagram has a small inaccuracy. The two cropped photos below show that the sheet steel trays are actually very thin. The dimensions of the components illustrated in the diagram clearly do not represent reality, but it is safe to say that the arrangement of the components are accurate.

The cropped photo below confirms that there are no protrusions from the interior surface of the front plate at all.


The cost of this performance is the danger of sympathetic detonations of the neighboring modules. As you can see in the adjacent photo, the top of the welded body blew off. Since there is no hole under that particular panel, it seems like the panel beside it inadvertently set it off, meaning that the partition between the two modules disintegrated under the pressure of the detonation of the first panel. This was long-known issue with ERA in general, and particularly with Kontakt-5 due to the large mass contained in each panel.

Presumably, one of the contributing factors would be poor welds. Perhaps sympathetic detonations could be prevented if the partitions were welded onto the hull by professional welders and not journalists.

Although it has been obsolete for over two decades, Kontakt-5 should be appreciated as a unique and ingenious solution to the problem of powerful long rod projectiles during the mid to late 1980's.


Two fuel tanks are located on the two front corners of the hull (flanking the driver), which extend from the nose of the glacis to almost up to the turret ring. These fuel tanks provide a modicum of armour. This decision was deliberate, as the fuel system is designed to use the front fuel tanks last.

Diesel fuel is can act as a form of liquid armour in enclosed spaces. The entry of a high energy shaped charge jet into an enclosed liquid medium at high velocities create a shock wave, which are reflected back into the jet. The shock wave emanates from the tip of the shaped charge jet and reflects off the walls of the fuel tank at a right angle to the original shock front. This is shown in the diagram below, taken from the NII Stali website. Due to the fast forward motion of the tip of the cumulative jet, the reflected wave will intersect with the body of the jet.

The ability of a fuel tank to sustain such an effect hinges on the rigidity of its walls. However, testing has proven that even fuel tanks with thin walls are able to stop a shaped charge jet to a limited extent. The main factor is the energy of the jet; when the energy of the shaped charge jet is high, a powerful shock wave is generated and thicker walls are required to effectively reflect the wave without rupturing. A thin-walled fuel tank will be sufficient for a weak shaped charge jet, or a powerful shaped charge jet given sufficient armour in front of the fuel tank to slow down the jet. Soviet testing of the fuel tanks in the T-34 and the T-54 showed that as long as the residual penetration of the shaped charge jet or armour piercing projectile is low, even thin-walled fuel tanks are able to completely absorb the threat without rupturing or exploding.

It wouldn't be wrong to consider the areas of the hull with fuel tanks underneath them to be essentially immune unless the warhead can overmatch the armour by a factor of more than 100 millimeters of RHA steel, though those same fuel tanks might also be a fire hazard if punctured or compromised. The fuel tanks do not have thick walls, and they are not foam-filled, and according to ex-tankers in Chechnya, they will visibly bulge and swell if penetrated by an RPG, though in those cases they were still strong enough to not burst or leak. In one incident during battles in Grozny, a T-72 was struck from the side by an RPG or SPG warhead in the driver's station. This T-72 did not have Kontakt-1 installed, but the combination of the spaced armour of the side skirt and the properties of the fuel tank managed to stop the cumulative jet from hitting the driver. Therefore, we can quite confidently say that the armour over the driver's station from the side aspect is equivalent to more than 400mm RHAe (with side hull armour and side skirt spacing factored in), which should account for its ability to resist a fairly typical PG-7VS rocket grenade. (Obviously, that is not a definitive value, considering the infinite variety of warheads available). The T-72 in that incident escaped with very minor damage.

In some other cases, like the T-72B obr. 1989 in the photos below, damage to the fuel tank may not produce any fire at all or may not be dangerous. Here, we see that the tank was pierced on the lower glacis by either a shaped charge warhead or an APFSDS shell from the front. A fuel tank is directly behind this part of the lower glacis, but the tank is clearly not burnt.


Earlier T-72 models can either lay its own smokescreen by injecting a diesel fuel into the exhaust manifold via the TDA (Thermal Smoke Apparatus), and later variants have the option of using its smoke grenade launchers. TDA is an inexpensive and extremely useful method of providing quick concealment at the cost of 10 liters of diesel per minute of continuous operation. By injecting diesel into the exhaust manifold, the hot manifold evaporates the fuel instantly, and it is ejected from the exhaust port by the exhaust gasses. Upon contacting the cool ambient air, the diesel mist condenses, forming a thick white fog. The fog obturates light in the 400-760 nm wavelength range, or in other words, the entire spectrum of visible light. This makes the TDA system a viable method of concealing the tank from anti-tank guided missiles, anti-tank guns and other tanks during daylight hours. The fog does not mask the tank from infrared searchlights like the AN/VSS-1 and AN/VSS-3A, which operate in the 785-1000 nm range, but it is possible to create denser smoke by driving the tank at a higher speeds to increase the fuel consumption rate by 10 times. High density smoke obturates light in the 400-3000 nm wavelength range, making it effective at concealing the tank from active infrared imaging systems. However, TDA cannot offer any concealment from thermal imaging devices like the AN/VSG-2 Tank Thermal Sight (TTS) installed in the M60A3 (TTS), which operates in the 7,600-11,750 nm range.

The driver should not shift gears when the TDA is in action if he wants to maintain a continuous curtain of fog, as the change in engine load will affect the volume of fog produced. It is not recommended to use the system for more than 10 minutes, and there must be an allowance of 3-5 minutes between each use. If the driver adheres to all of the guidelines, the system can be used for an infinite number of times.

Low volume smokescreen while idling

High volume smokescreen while moving
Aside from the TDA system, the T-72 also features the "Tucha" smoke grenade system. It can launch two types of caseless grenades; the 3D6 and the 3D17. A high-low propulsion system much like 40mm VOG series of grenades is used to launch the grenades. Twelve grenades are available.

Earlier T-72 versions had their smoke grenades installed on the turret cheeks.

This wasn't the wisest idea, since a direct hit on the turret cheeks could potentially remove half the tank of its ability to react by deploying a smoke screen. 


The 3D6 smoke grenade emits "normal" smoke that can only obscure the tank in the visual spectrum. This type of grenade has been rendered next to useless with the gaining popularity of thermal imaging sights in the mid-80's, now long supplanted by the 3D17 model. It is of the slow-burning type, emitting smoke from the ground-up. It travels anywhere from 200m to 350m after launch, and it takes between 7 to 12 seconds to produce a complete smokescreen 10m to 30m in width and 3m to 10m in height, depending on various environmental factors like wind speed, humidity, altitude, etc. This is not including the time taken from launch to the grenade actually hitting the ground. This is in accordance with frontal assault tactics where tanks advance and maneuver behind a continual wall of smoke generated every forward 300m until they literally overrun enemy positions. The smokescreen can last as long as 2 minutes, depending on environmental factors.


The 3D17 is an advanced IR-blocking aerosol smoke grenade. It completely obturates the passage of IR signatures or IR-based light as well as light in the visible spectrum. It is effective at concealment from FLIR sights and cameras as well as at blocking and scattering laser beams for tank rangefinders and laser-homing missiles. Unlike the 3D6, the 3D17 grenade detonates just 1 seconds after launch, allowing it to produce a complete smoke barrier in 3 seconds flat. The drawback to this is that the lingering time of the smokescreen is only about 20 seconds, depending on environmental factors. This is enough for the tank to hastily shift its position, but not much more. This grenade detonates 50m away from the tank.


Nakidka is a type of multi-wavelength infrared suppressant camouflage developed in 1971. Contrary to popular belief, the Soviet concept of warfare was centered around "deep battle" (rather than Zerg rushing), which greatly depended on "maskirovka" - the element of camouflage and deception. Implementation of "maskirovka" includes decoys, stealthy operations, concealment and surprise attacks. Nakidka plays an important role in this. It is a textile "dress" for the tank, which can neutralize the tank's IR signature (except at its exhaust outlet) and reduce its radar cross-section in addition to presenting a totally non-reflective camouflaged surface, thus drastically reducing the tank's likelihood of being detected in the visual and non-visual spectrums.

Nakidka is resistant to napalm and is unaffected by machine gun fire, though it is possible to destroy it with high-explosives. Still, the point of Nakidka is to prevent the tank from being spotted in the first place. It holds up fine against indiscriminate area weapons. A full suit of Nakidka only adds several dozens of kilograms to the tank's overall weight.


Like many tanks of the era, the T-72 has an escape hatch. The (rather small) hatch is located directly behind the driver's seat, and therefore most easily accessible by him. The gunner and commander can get to the hatch as well, but they have to be very, very flexible in order to do so unless the turret is traversed to the rear. Nevertheless, it is indispensable in certain situations, allowing crew members to escape the tank if it is flipped over or when the tank is under fire from dismounted infantry. The escape hatch cannot be used while the crew are wearing winter clothing or a bulky IP-5 rebreather. The hatch is strong enough that it does not compromise the integrity of the hull against a 6kg to 10kg anti-tank blast mine detonated under the tracks.

The hatch is far too small for anyone wearing winter clothes, and more rotund tankers will obviously find it impossible to exit through it. The hatch is fully air-tight, and drops out to open. The hatch is as thick as the rest of the hull floor, and is held in very, very firmly in place by four locks.


Ventilation is controlled from the KUV-11-5-1S ventilation and filtration management box. The ventilation system has a built-in dust ejector at the air inlet to ensure a supply of clean air under normal operating conditions.

The diagram below - taken from "Special Electrical Equipment of the T-72" published by the military department of the Omsk State University of Technology - gives us a cross section of the system. The air outlet for the ventilator in the normal operating mode is marked (21). Air is taken in by the fan, flows through the air booster, and exits through the outlet (21). A dust ejector is installed at the air inlet to ensure that clean air is supplied into the crew compartment even under highly dusty conditions.

The ventilator draws air from a port on the hull roof, located just behind the turret ring. Before crossing water obstacles, the ventilation system is deactivated and the air intake is closed to prevent water from entering the fighting compartment and to prevent damage to the electric motor.

The ventilator housing and the white pipe leading to the air intake can be seen tucked away in the rear corner of the fighting compartment in the photo below. The air outlet from the filtration system drum is indicated by a red arrow.


Soviet tank designers were very conscious of the dangers of nuclear warfare, especially artillery-fired tactical nukes. The T-72 perfectly reflected their seriousness, featuring the GO-27 NBC protection suite and the KUV-11-6-1S ventilation system with a filtration unit and the capability to generate an overpressure. A radiation lining shielded the occupants from neutrons. The photo above shows the B-1 instrument and control box, the B-2 sensor for gamma radiation detection, and the B-3 power supply unit. 

The dosimeter detects and measures gamma radiation levels. The B-1 instrument and control panel displays the radiation level in rads per hour (rad/h), and is able to measure and display the radiation level in a range between 0.2 to 150 rads per hour. The system has a measurement accuracy of ± 30%. The B-1 instrument and control panel is shown in the photo below. Photo credit to Leonid Varlamov.

The system has different reactions depending on the rate of dosage of radiation. The system is able to react instantaneously to a nuclear detonation (classified as a Type "A" radiation threat) and initiate the necessary protective measures.

  • Type "R": When the tank is exposed to gamma radiation from a radioactively contaminated site and is exposed to a dose rate of 0.85 Rads/h and above, the response time of the system does not exceed 10 seconds.
  • Type "A": In the event that the tank is exposed to a gamma ray flux with a dose rate of 4 Rads/s and, the response time of the system does not exceed 0.1 seconds.
  • Type "O": When biological or chemical contaminants are detected, the response time of the system does not exceed 40 seconds.

The reaction of the system includes visual and audio signals to alert the crew. The above photo of the B-1 instrument and control box shows three coloured incandescent lights marked "O", "P" (R in Cyrillic) and "A". When any one of the threats is reacted upon, the driver is instantly informed of the type of threat by the colour of the light.

Once a Type "A" radiation threat is detected, the system immediately activates the air filtration system and initiates the lock down protocol, which seals every gap exposing the interior of the tank to the outside environment. Gaps such as the co-axial machine gun port are sealed using steel barriers propelled into position by pyrotechnic charges. Due to the immense speed of gamma rays (very close to speed of light) and the quick reaction of the system, the tank will be hermetically sealed by the time the blast wave from the nuclear explosive arrives. This protects the crew from the blast wave itself as well as from exposure to fallout after the initial blast wave.

A Type "R" radiation threat is a much less serious situation. Type "R" threats are detected when the tank is exposed to radiation from an irradiated environment. The long reaction time of the system to this type of threat is offset by the low danger of minor irradiation.

Type "O" threats are airborne biological or chemical threats. The system detects contaminants in the air using a cyclone-based air sampler and analyzer. The air inlet for the sampler and analyzer is depicted in the diagram below. Due to the rather long reaction time, the driver is sometimes obligated to manually switch on the chemical and biological threat protection measures when entering contaminated zones, assuming that the tank is preceded by a forward reconnaissance force that included chemical troops mounted on NBC reconnaissance vehicles like the BRDM-2RKh.

The air inlet is installed just next to the driver's hatch. Photo credit to the Facebook page.

The location of the B-2 gamma radiation sensor can be seen in the photo below, taken from the STV Ground website.

The B-3 power supply unit is installed just next to the gear shift:

PKUZ-1A Digitized Protection Complex

The GO-27 system was replaced with the PKUZ-1A in the T-72B3 modernization. The PKUZ-1A was first used in the T-90A, and features improved detection and reaction time to chemical, biological and nuclear threats. The PKUZ-1A analyzes the air outside the tank using an ionizing system.

The system capable of detecting gamma rays with energies ranging from 0.66 to 1.25 MeV. The system is capable of measuring gamma radiation at dose rates of 0.1 to 500 rads/hour, making it somewhat more versatile than the GO-27. In order to measure the true level of radiation outside the tank, the radiation attenuation coefficient of the armour of the tank and the anti-radiation linings is manually inputted at the factory. This improves the accuracy of the system. Like the GO-27 system, PKUZ-1A automatically executes defensive systems and alerts the crew via visual and audio signals when an NBC threat is detected.

The PKUZ-1A system comes with a new instrument and control box. The new control panel fulfills the same function as its predecessor, but is more user friendly. The old ammeter-based radioactivity gauge was replaced by a digital LCD segment display for quicker and more precise readings. The old ammeter gauge display could not give an accurate reading if the tank was moving because the vibrations caused the indicator needle to jump around.

The new control panel can be seen at the right side of the screenshot below.



Anti-radiation measures have been among the top priorities regarding crew protection, no less important than solid armour itself, given the nuclear environment that the T-72 was expected to fight in. In accordance with this requirement, the T-72 was fitted with an interior anti-radiation lining called "Podboi" since the very beginning, but the tank had an additional anti-radiation cladding of the same type installed on the exterior of the turret and hull from 1983 onward. This was apparently a response to an announcement by U.S president Ronald Reagan in 1981 that the production of neutron bomb production would be restarted. The heavy armour of the T-72 - and tanks in general - provided very good protection from the immediate destructive effects of neutron bombs, but the powerful burst of neutron radiation could not be easily blocked. The thin roof and sides of the turret and hull were particularly vulnerable, being much thinner than the frontal armour of the tank. The external cladding, called "Nadboi", was therefore concentrated around these areas. This is most obvious with the T-72B, which had the anti-neutron cladding installed since it was formally introduced in 1985:

The lining and cladding are composite materials composed of a mixture of polyethylene and polyisobutylene - polymers with a high hydrogen content, allowing them to absorb large amounts of radiation. The lining and cladding are fitted on the tank with a special glue and pressed firmly to the tank by special metal studs with a washer affixed to the head. The polymers are impregnated with lead to increase their opacity to gamma radiation, and boron was added as a response to developments in neutron bomb technology. One of the components was borated polyethylene, a type of high density polyethylene infused with boron. According to Anderi Tarasenko, the name of the material is "boron 2EP002". Boron is known to be extremely effective at capturing neutrons thanks to its large absorption cross section, making it suitable for use as radiation shielding. Unfortunately, the high cost of boron compounds made it impractical to implement in a high concentration, so it was decided to include only a thin layer of borated material in the composite cladding. The location of the layer was such that it reportedly slashed the required boron content by half, but the reduction in radiation dosage remained at the same level as before.

The fibrous construction of the sheets and layered nature of the anti-radiation lining also makes it a suitable spall liner not dissimilar to early flak vests that used woven nylon plates. NII Stali states on their website that as a rule, spall liners are made from aramid (kevlar) or from UHMWPE (Ultra High Molecular Weight Polyethylene). This hints heavily at the dual roles of the borated polyethylene anti-radiation lining. The high thickness of the lining doesn't hurt either. The cross-sectional drawing below, taken from a CIA report on Soviet tank developments, indicates that the thickness of the "Podboi" lining is around 50mm. From the thickness alone, it is very likely that the lining is capable of absorbing some amount of spall even if it is not as effective as a more modern kevlar-based spall liner.

The flammability of the "Nadboi" cladding is unclear, but it is beyond question that it was designed to survive the heat from a nuclear explosion. The cladding would have to fulfill its purpose as neutron and gamma radiation shielding before it gets swept away by the nuclear shockwave and high speed winds (and the debris it carries) since neutrons and gamma radiation will arrive at the tank instantaneously, but the cladding needs to survive the flash of heat from the blast, since heat radiates at the speed of light. To prevent the destruction of the cladding from the heat of the nuclear blast, the outermost layer of the composite material is made from a flameproof material. Since "Nadboi" is almost always observed to be missing from burnt-out T-64, T-72 and T-80 tanks, it is obvious that the material is still flammable to some degree. This may not be entirely relevant in a combat situation, as the cladding is often burnt off by an external heat source, like burning fuel from the wrecked tank. It may be a problem if the tank is attacked with napalm or other flame weapons, but such attacks would be a very minor issue compared to more serious anti-tank weapons like recoilless rifles and guided missiles.

In the photo below, a T-64 with external "Nadboi" anti-radiation cladding displays the damage dealt by a 122mm HE-Frag artillery shell. Note the charred chunks of fabric, proving that the cladding is made from textile sheets. More importantly, the cladding has not burned off entirely. The damage is almost entirely localized to the point of impact of the artillery shell, indicating that the cladding does not burn readily when subjected to an intense flash of heat. Instead, it is much more likely that the cladding was stripped off by the blast of the shell and not burnt off.

The photo below shows how the mounting studs for Kontakt-1 reactive armour protrude through the "Nadboi" cladding. The thickness of "Nadboi" on the turret roof is around two inches, and the thickness of the cladding on the turret hatches are more than two inches thick. This is because of the low thickness of the hatch compared to the turret roof.

The circles with four holes that pockmark the surface of the cladding are the metal studs that press the cladding on the surface of the turret. When the cladding material is burnt away, these studs usually remain intact since they are welded to the turret. See the two photos below showing a burnt-out T-72 turret (credit to armour-kiev-ua).

The hatches on the turret are covered in pre-formed panels. The thickness of the "Nadboi" cladding is most obvious on the commander's hatch, seen in the photo below.

The shell casing stub ejection hatch is heavily shielded with two inches of "Nadboi" and "Podboi" on both sides. This is because the low thickness of the hatch itself is insufficient to stop gamma radiation.

As mentioned before, the lining and cladding not only function as neutron absorbers, but they perform admirably as a form of spall liner as well. According to Swedish trials of purchased ex-East German T-72M1s, it was concluded that the anti-radiation liner was perfectly capable of absorbing secondary fragments of penetrating cumulative jets. Depending on the construction, spall liners may reduce the spray cone angle of secondary fragments from a shaped charge warhead by up to 50% or more if the armour is greatly overmatched and it is possible reduce the quantity of secondary fragments by up to 80%. The NII Stali web page gives a more optimistic claim that the spray cone angle of secondary fragments (unknown type of anti-armour ammuntion) can be reduced by a factor of 3 and the quantity of fragments can be reduced by a factor of 10. The reduction in the amplitude of a shockwave from an external explosion is in the order of 4.5-5 times for a lightly armoured vehicle. The T-72 is not a "lightly armoured vehicle", of course, but the presence of a spall liner would still help improve the conditions inside the tank if explosive ordnance detonated outside. If the armour is not completely perforated, the spall liner may absorb all of the spall produced from the surface of the armour plate. Either way, the likelihood of injuring the crew or damaging the internal equipment of the tank is greatly reduced. The anti-radiation lining and cladding should have good performance on account of its substantial thickness both inside and outside the tank. In fact, this feature has helped to saved lives in at least one confirmed incident:


In this instance, the T-72 was hit in the flank by an RPG attack which also blew off a part of the port side storage bins. The crew survived and the tank only suffered from a minor puncture wound. It is interesting to note that that part of the tank is above the autoloader carousel, but loose ammunition is stowed on the wall with clips. The survival and quick repair of the tank indicates that there was no ammunition stowed in that particular area. Note that the exterior of the side hull over the fenders is also covered with the anti-radiation cladding.

The presence of the lining is a huge factor in the safety of the carousel ammunition in case of armour perforation, especially from the side, but that's not all; due to boron's large surface area-volume ratio, it does quite well at absorbing blast waves, thus mitigating some of the effects of blast damage. Additionally, the lining helps to insulate the tank and prevents condensation. This helps preserve the myriad of electric and electronic components in the tank.

However, the material does not come without some risks. It is flammable, and constitutes a minor fire hazard. The lining and cladding was partially removed in the T-72B3, as you can see in the photos below.

Why they chose to remove "Podboi" from some parts of the sides of the turret while keeping the lining on the roof and the turret cheeks is unclear, but it may be because the T-72B3 is not expected face neutron bombs in the near future. The only function of "Podboi" would then be as a spall liner in the event that the tank is hit from the front.


To prevent the spreading of internal fires in the engine and crew compartments, the 3ETs11-2 quick-acting firefighting system was installed. There are a total of fifteen TD-1 thermal sensors installed inside the tank, strategically placed in the engine compartment and crew compartment. The fire fighting system reacts regionally when a temperature difference of at least 150°C is detected in the crew compartment or engine compartment. Once a fire is detected by any one of the fifteen TD-1 sensors, the maximum response time of the system is 50 milliseconds.

The TD-1 thermal sensor consists of fifteen thermocouples wired in series. The reaction time of the TD-1 sensor does not exceed 10 seconds, meaning that it takes a maximum of 10 seconds between detecting the fire to the activation of the fire extinguishing system. The sensors do not guarantee reliable detection of fires in the 60°C to 150°C range of temperature differences due to insufficient contrast. This is due to the hardware limitations of the TD-1 sensors.

The driver can manually activate the fire extinguishers wired to the automatic firefighting system from a red control panel to his right.There is also an additional manual fire extinguisher to the driver's left foot.

A TD-1 thermal sensor is shown below.

The photo below shows a T-72A tank. Notice the large number of TD-1 sensors placed on the walls and around the floor. Note that the sensors are all concentrated near potential fire hazards; the conformal fuel tanks, loose ammunition stowage positions, and the powerful amplidyne amplifier for the turret traverse motor. There are five TD-1 sensors placed next to the rear conformal fuel tank alone.

The P11-5 control and information panel is part of the firefighting system. The panel has seven indicator lights. The three lights on the top row (3, 5, 6) are to inform the driver of the serviceability of the pyrotechnic fire extinguisher quick release valves, the light on the center left (2) indicates the presence of a fire in the fighting compartment, the light on the center (4) indicates the presence of a fire in the engine compartment, the light on the center right (7) indicates the status of the air filtration system, and the light at the bottom center (12) indicates if the OPVT mode is activated. By referring to indicator lights (2) and (4), the driver can manually discharge the fire extinguishers for either the fighting compartment or the engine compartment by pressing the buttons (1) and (15), which are located behind a hinged metal cover.

The panel is partially visible at the very top of the photo below (from Prime Portal, credit to Marek Solar).

Like most firefighting systems for armoured vehicles, the 3ETs11-2 uses freon gas as the fire extinguishing agent.

Two handheld OU-2 carbon dioxide fire extinguishers are also provided to supplement the automatic fire extinguisher system. If the TD-1 fire detectors fail to respond (usually in the case of small flames), then these will be the only firefighting tools available to the crew, if the driver opts not to manually activate the extinguishers connected to the 3ETs11-2 system.

The T-72B3 modernization replaced the 3ETs11-2 firefighting system with the newer 3ETs13 "Iney" system.

"Iney" employs a slightly more modern control system, but the fire detection and response algorithms are essentially the same as in the 3ETs11-2. The main improvement offered by "Iney" is the use of new OD-1S optical thermal sensors. Ten of the fifteen TD-1 thermal sensors of the earlier 3ETs11-2 were replaced with OD-1S optical sensors, all installed in the crew compartment to maximize crew survivability. The engine compartment is still only equipped with five TD-1 sensors in the same locations as before.

The response time of the OD-1S optical sensor does not exceed 2 ms. This is a very substantial improvement over the 50 ms response time of TD-1, and contributes towards the much quicker overall reaction time of the system.

The P11-5 control and information panel was replaced with the P708 digital control and information panel. P708 replaces the simple incandescent light bulbs for the indicator lights on the P11-5 with an LED display, which is a much more intuitive way of conveying information to the driver quickly and efficiently, but other than that, the new panel is exactly the same as the P11-5. A close look shows that besides the new LED screen, the buttons, toggles and other interactive components are exactly the same as in the P11-5.

A P708 control panel can be seen tucked away at the right side of the photo below. Photo taken from Popular Mechanics Russia.


An a self-entrenchment blade is provided at the lower front hull of the tank. It is secured by two rotating latches, which need only to be turned with a wrench by a crew member for the dozer blade to be usable. Needless to say, it is an invaluable tool for self-fortification, allowing the tank to enter hull defilade when natural cover is unavailable, or even augment existing cover with additional concealment.



With the dozer blade, the T-72 can create a soil barrier in front of itself from even ground in about 20 minutes or more and much less if on uneven ground, but depending on meteorological conditions. On snowed-over terrain, a snowbank may be created in as little as 5 minutes to help conceal the tank.


The T-72 is furnished with a plethora of stowage bins intended for the storage of various things. The most prominent ones are the two large bins located around the rear arc of the turret. These are used for storing the crews' personal effects as well and other accessories. The lids of the bins are sealed by tension latches. These latches are effective at keeping water from entering the compartments.

This is quite the improvement over the T-55 and T-62, as these older tanks were not equipped with external stowage bins on the turret but instead had loops from which bags could be suspended. As a result, the number of locations to stow day to day necessities was rather limited and could be lost if not secured properly to the stowage loops on the surface of the turret. The Israelis gave their Tiran tanks with Centurian-esque external stowage bins on the turret for this very reason.

The photo below shows the stowage bin at the very rear of the turret. There are two isolated stowage compartments in the bin. One on the right hand side (the left side in the photo) for smaller things, and the central compartment, which is large enough to stow anything you want to. The bin is hinged to the turret, as you can see in the photo below.

The photo below shows the bin hinged open to allow easier access to the engine access panel. How it stays up isn't exactly clear.

Besides the rearmost bin, there is also the side bin. It also has two isolated compartments.

The bins are too thin to deflect bullets or mortar shell fragments.

There is also bank of 4 storage bins on the port side of the hull, directly above the tracks.

The port side storage bins are usually used to store maintenance equipment and spare parts.


The T-72 was a true main battle tank, having successfully combined the three attributes that govern tank design: firepower, armour and mobility. Throughout its career during the Cold War, the T-72 had one of the world's most powerful tank guns, had excellent armour protection, and was an agile opponent in combat. Its on and off road performance almost reached the same level attained by the speedy Leopard 1 and AMX-30, outmatched the heavily armoured tottering Chieftain and Challenger tanks and greatly outpaced the sluggish M60A1 and A3, all while weighing and costing less than any of them.
The superior engine power of the T-72 and its light weight meant that it could not only traverse difficult terrain, but that it could safely cross low-capacity bridges and make good use of the thousands of tactical bridge layers in Soviet army service, even including the ones derived from the then-already-antiquated T-54. Not only is it possible for the T-72 to exploit light load masonry bridges or pontoon bridges, it is possible for a convoy of T-72s to travel over such structures without needing take turns to drive at a snail's pace.
Swedish mobility trials of T-72M1s (and MTLBs) in Northern Norrland between 1992 and 1994 yielded very positive results. The T-72s in question displayed good performance over snow as deep as 0.8 meters although it still failed at times to reliably traverse frozen ice banks, but it can be argued that that was because of the inexperience of the Swedish test drivers who might not be too familiar with the idiosyncrasies of the T-72.


The T-72 has been host to several engines over the years, starting with the V-46, evolving into the V-84, and finally the V-92. All of the T-72 engines to date are V-12 four-stroke diesels, with some limited multifuel capability. They are able to consume low octane gasoline (A-66 and A-72), diesel, and jet fuel (T-1, TS-1 and T-2). The driver can set the type of fuel by simply setting a dial located in his station. The engine does not need to be further modified beyond that, but it is highly inefficient when using anything except diesel. An ST-10-1S starter-generator is attached to the engine. It is a starter motor that doubles as the electrical generator to power the electrical systems of the tank. The operating speed of the generator is between 3,600-6,250 RPM. 14.7 kW is used to start the engine electrically.

The main method of starting the engine is via an electric starter. In cold weather, the engine can be started with compressed air, or even perhaps by towing. In exceptionally cold weather conditions, the most dependable method of starting is a combination of compressed air and the electric starter. It takes around 20 minutes to start the engine in extremely cold weather, which is much longer than the 3 minutes needed by the GTD-1000T gas turbine engine used on the T-80, but diesel piston engines have their own advantages. Usage of the compressed air starting system is avoided except when absolutely needed as it wears out the engine.

A pair of compressed air bottles are used for the engine starting system. They are placed at the very front of the hull, to the right of the driver's feet but to the left of the right hull fuel tank. The compressed air is also used for the pneumatic periscope cleaning system.

It is not known if the compressed air bottles pose a hazard when the tank armour is struck but not pierced. There is no doubt that the bottles will explode if penetrated by a shaped charge jet or by metal fragments, but the small size of the bottles make that unlikely unless a very specific part of the front hull armour is hit.

V-46-4 / V-46-6

The V-46 liquid-cooled engine is the baseline engine for the T-72 series, first appearing on the T-72 Ural and then the T-72A. It traces its roots to the V-2 which once powered the legendary T-34. True to its remarkable origin, it has a remarkable power density, far above its competitors such as the; MB 837, which powered the Leopard 1 series, AVDS-1790-2A, which powered the M60 tank series, and even the "lightweight" opposed-piston Leyland L60 series, which powered the Chieftain tank. When compared: AVDS-1790-2A - 0.324, MB 837 - 0.426, Leyland L60 - 0.535, V-46 - 0.795, the V-46 comes out on top. Overall, the V-46 and all its descendants are unquestionably robust, dependable engines in every way. A disadvantage of this engine is the amount of smoke it produces, which may expose its position to enemies equipped with thermal imagers.

Output: 780 hp
Rated speed: 2000 rpm
Idle speed: 800 rpm
Fuel Consumption: 1 g / 245 kWh or 1 g / 180hp.h
Torque back up: 9% ... 18%
Weight: 980 kg

T-72 Ural and T-72A power to weight ratio: 18.1 hp/ton

The exhaust port for this engine is characteristically long and narrow. It has very rudimentary sheet steel cooling fins on top. The fins are arranged so that as the tank drives forward, cool air rushes from one side of the fins to the other, drawing away some heat along the way.

The exhaust port connects to the exhaust manifold via a simple duct. The exhaust port is secured onto the duct via a pair of bolts and nuts on either side.

The V-46-4 is the variant which the T-72 Ural uses, while the V-46-6 is used in the T-72A. The only difference between the V-46-4 and the V-46-6 is a change in the placement of oil containersWith the V-46, both the T-72 Ural and T-72A can achieve a top speed of 60 km/h on asphalt, and set an average speed of 35 to 40 km/h on dirt roads.

V-84-1 / V-84MS

The V-84 engine differs from its predecessor mainly by an increase in output, along with an insignificant weight gain. The additional power comes from the new centrifugal gear-driven supercharger, which provides better aspiration for cleaner combustion in the cylinders. The increased power offsets the added weight of the tanks that have it installed, which includes the T-72A/B obr. 1984 and all subsequent T-72B models, allowing it to remain as nimble as its predecessors. This engine is much less smoky than the V-46 because the higher oxygen levels in the combustion chamber allowed a greater portion of the fuel particles to be consumed for more efficient consumption of energy, producing more output. One side effect of the added power is the increased heat output. Since the cooling fan for the radiator draws power directly from the engine, the increased heat is mostly eliminated, but more heat escapes from the exhaust manifolds. The engine also wears out slightly faster because of the increased power.

Output: 840 hp 
Rated speed: 2000 rpm
Idle speed: 800 rpm
Fuel Consumption: 247 g/kWh or 182 g/hph
Torque back up: 6% ... 18%
Weight: 1020 kg

T-72B, T-72B1, T-72BA power to weight ratio: 18.87 hp/ton 
T-72B3 power to weight ratio: 18.2 hp/ton

The maximum torque output of the V-84-1 is 340 Nm. This is much more than 196 Nm produced by the 5TDF of the T-64BV but much less than the 448 Nm produced by the GTD-1250 of the T-80U and the 538 Nm of the AGT-1500 of the M1 Abrams.

The exhaust port for the V-84 type engine is identical to the V-46.

Like previous variants, the T-72B has a top speed of 60 km/h on asphalt, and an average speed of 35 to 40 km/h on dirt roads. This remains mostly unchanged even with the burdensome Kontakt-5 installed. Most T-72B3s are equipped with this engine.


The V-92S2F turbocharged engine boasts an impressive power density of 1.02 hp/kg combined with high standards of reliability. The increased torque reserve greatly improves driving characteristics across rough terrain and the fuel efficiency has been substantially increased, boosting the T-72's already good fuel economy to a new high. The engine is virtually smokeless. The cylinders and pistons were updated and more robust compared to previous engines to cope with the added power.

Output: 1130 hp 
Rated speed: 2000 rpm
Idle speed: 800 rpm
Fuel Consumption: 215 g/kWh or 158 g/hph
Torque back up: 25% ... 30%
Weight: 1100 kg

T-72B3M / T-72B4 power to weight ratio: 21.73 hp/ton

Variants outfitted with the V-92S2F can be identified by the heavily modified exhaust unit, now much narrower, but fatter and with different fins. A new exhaust unit was needed due to the turbocharger.

Without the cooling fins and muffler removed, the exhaust duct itself is just a simple metal tube.

The exhaust duct is fitted over the exhaust manifold on top of the engine, which is oval shaped.

The use of the V-92S2F on the T-72B3M boosts its top speed to a blistering 75 km/h on paved roads and allows it to cruise cross-country at a speed of up to 60 km/h on dirt roads. This elevates the tank's mobility to the level of the T-80BV speed-wise, and gives it something close to parity when moving cross country thanks to the high torque reserve.

The MS-1 cyclone air filter used with all of the V-series engines is adequate for most environments. It is a two-stage filter, and requires a filter change once every 300 km traveled under extremely dusty conditions. According to Sergey Suvorov, the filter requires maintenance every 1000 km in winter, and every 500 km in the summer.

The T-72's engine deck is taken up by the engine access panel, the engine's air intake, radiator/air intake and the cooling system air outlet. All of them except the engine air intake have armoured covers to protect them from bullets and shrapnel coming from above. 

Left and right sides. Engine access panel up front, radiator/air intakes behind it (with armoured covers), and cooling system air outlet behind that (again with armoured covers)

The engine can be removed with the use of a 1-ton crane, which can be found at even the most modest depots. In the field, engine replacements are done with the help of engineering vehicles. The two photos below show the two-layered engine deck opened up to expose the engine, ready to be serviced or removed.

However, the T-72's engine is not integrated as part of a powerpack, like on the Leopard 2. Powerpacks are far more convenient to replace. It usually takes more than an hour to replace both the engine and transmission of a T-72, compared with only about 35 minutes or less for more modern vehicles, like the Leopard 2. Highly skilled teams can replace the powerpack of a Leopard 2 in less than 20 minutes.

Air intake for V-46 engine, tucked away discreetly behind the turret
Modified air intake for V-84 engine

The engine deck is cool enough that people can ride on top of it.


The liquid cooling system is of a convection type. It works with water and air, used to cool hot coolant oil that is pumped around the engine. The coolant oil first runs up to the radiator unit, where it is cooled by water flowing in a labyrinth of aluminium fins with turbulators, which is itself cooled by flowing air being sucked in by an engine-driven fan at the rear of the engine compartment. The unwanted hot air is pulled into the fan and ejected out of the rearmost outlet in an upwards direction. 

The biggest drawback of this system is that dust particles kicked up into the air from driving at high speed may be sucked up by the high velocity air stream from the cooling fan, creating a distinctive "rooster tail" dust cloud behind the tank. The radiator pack is shown in the photo below.

Reports indicate that this system may be slightly limited, sufficient for European climates at best. It was designed so that the engine will work with no loss in efficiency at an ambient temperature of up to 25° C, but the engine will begin to experience more significant reductions in performance at temperatures exceeding that. Overheating becomes a major issue in ambient temperatures of up to 50° C, which is sometimes recorded at the Thar desert in India. At temperatures above 45° C, the engine will begin to suffer huge reductions in power (up to 33% loss). At such temperatures, the tank must be stopped every 25 kilometers to allow the engine to cool to prevent excessive fatigue. The simplest solution (as practiced by tank crews all over the world) is to remove the covers, which helps to improve air intake volume through the radiators to increase the rate of cooling, but this is not sufficient on extremely hot days. At temperatures of 30° C or less, the cooling system is more than adequate for its task.

According to "Пути Снижения Затрат Мощности В Системах Танкового Диселя" ("Ways to Reduce The Power Costs in Tank Diesel Systems") by S.P Baranov and V.T Nikitin, the power consumption of the cooling system of the T-72 is 7.7% at an ambient temperature of more than 30°C. This is somewhat lower efficiency compared to the ejection-type cooling system of the T-64 but better than the fan-type cooling systems of foreign tanks like the M60A1, Leopard 1 and Leopard 2 which consume 14.4%, 14.7% and 14.5% of engine power respectively. The ejection-type cooling system of the T-64 simply takes less power than the fan-type cooling system of the T-72, but the lower overall power consumption compared to the foreign are largely explained by the higher coolant temperature and higher permissible coolant temperature for Soviet tanks due to the use of oil rather than water as a coolant. In the case of the Leopard 1 and Leopard 2, it is because water is used as a coolant, and in the case of the M60A1, it is because the AVDS-1790 series of engine is air cooled and not liquid cooled (which apparently makes it easier to maintain).

In total, the net engine power of the T-72 is 14.3% lower than the gross engine power after adding up the costs of the air intake heater and the air cleaning system. This is only slightly better than the M60A1, Leopard 1 and Leopard 2 which only have power deducted through the cooling system. Thus, the net engine power of the T-72 Ural and T-72A is 682 hp, and the net engine power of the T-72B is 712 hp. This is not much more compared to the 642 hp of the M60A1, only slightly less with the 710 hp of the Leopard 1 and much, much less than the 1260 hp of the Leopard 2. However, the modest engine power is counterbalanced by the lower weight of the T-72, which is lower than both the M60A1 and Leopard 2 for all T-72 variants. Astonishingly, the weight of a combat loaded T-72 Ural (41 tons) is even slightly lower than a combat loaded Leopard 1A1 (41.5). Of course, these power deductions are only for the engine related subsystems. The deductions from the electrical systems of the tank are not yet considered. For example, the ST-10-1S generator on the T-72 generates 10 kW, and the power comes from the engine. The electrical systems of foreign tanks are typically more power-hungry which is often reflected in better performance (quicker turret rotation speed, for example), but also results in slightly lower net power.  

Apparently, the V-92 engine series and its accompanying modifications have partially solved the overheating issue at very high ambient temperatures. Specific details are not known to the author, but it could only either be an increase in the centrifugal fan's power, or a simple modification of the water flow channels in the radiator, as Indian T-72s and T-90Ss apparently have.
The photos below show the radiator covers opened and closed, exposing the protective louvers within.

The photo below shows the engine compartment with cooling pack and engine access panel removed. 


Note the crossbar to hinge both of the aforementioned accessories. Also note the centrifugal fan at the bottom left corner. It is a simple riveted aluminium fan with a diameter of 655mm and a width of 205mm, with twenty evenly spaced vanes. It is powered by a driveshaft connected to the gearbox so that it increases or decreases its power in accordance with the engine's mechanical output, thus adjusting for the engine's heat output as well. It is strong enough to throw water out of the engine compartment like a blowhole even while the engine is idling.

Centrifugal fan

The use of a centrifugal cooling fan is one of the many conservative design features of the T-72, and in fact, the entire cooling system is fundamentally the same as the design used in the T-54. However, that does not mean that it was no longer viable by the 70's, as the design could meet the cooling requirements of the V-46 engine in most weather conditions while remaining compact, easy to maintain, and reasonably protected from top attack, although there are still a few flaws. By placing the radiator on the engine deck and exposing a large surface area, it becomes vulnerable to napalm attacks or molotov cocktails, as the cooling fan creates a suction force that can suck in burning gels and liquids through the radiator louvers, but this is compensated by the optional sheet steel covers. Closing these watertight covers prevents the ingress of burning liquids at the cost of accelerating the overheating of the engine. The cooling fan itself is well protected, since it is too small to be hit by aerial weapons and it can eject any burning liquid thrown inside it. By contrast, the cooling system of the Leopard 1 may offer better protection against incendiary attack as only the cooling fan is exposed on the engine deck whereas the radiators are not, but the radiators are on the sides of the hull, making them vulnerable to heavy machine gun fire and artillery shell splinters. The cooling fan itself is barely protected from ballistic attack, but does not need to be, since there is no coolant to leak and it can still function with a few missing blades.

In the event of damage from an air attack, maintaining or replacing the radiator is quite simple, since the entire unit can be hinged open. The radiator can be disconnected from the coolant pump quite easily, as the two components are only connected by two hoses.

The louvers that protect the radiator inlet, cooling fan outlet and engine air intake can all be shut or opened with the press of a button from the driver's station. Closing these louvers provide additional protection from aerial attack. With the louvers closed, the engine deck can have a very high resistance to hits from air-delivered cannon fire from low angles of attack. Examples include 20x110mm AP-I rounds from A-1 Skyraiders, the USAF's main ground attack plane in the early-mid stages of the Cold War, or 20x102mm AP-I rounds fired from AH-1 Cobras and in many fixed wing aircraft such as the F-4 Phantom and F-16, which may be used in the close air support role.

The photo above shows the engine access panel and armoured cover hinged open. Note the spaced armour arrangement. Since ground attack aircraft and attack helicopters almost never fly at high altitudes to deliver cannon attacks due to the risk of being seen and shot down, the armour is more than enough to deflect hits from all manner of cannon fire. A-10 pilots are trained to approach targets at an angle of attack of around 3 degrees from treetop level. Reducing the obliquity by a few more degrees will not change the fact that the engine deck is too thickly armoured to be affected even by shelling from 30mm DU rounds.


The T-72 uses a hydraulically assisted mechanical syncromesh transmission with dual planetary gearboxes and dual planetary final drives, a type of transmission that is known as a dual transmission system. This type of transmission is principally the same as one from the T-54, but better, of course. It is highly compact, rock solid, extremely reliable (practically unbreakable), and also quite precise, meaning that the driver can direct the tank between obstacles more easilyThere are seven forward gears and one reverse gear. The transmission weighs 1,870 kg.

According to the manuals for the T-72A and T-72B and the book "Main Battle Tank" (Основные Боевые Танки) published by Arsenal Press in 1993, the gear speeds (at 2,000 RPM) and gear ratios are as follows:

Gear speeds (km/h) Gear ratios
1st: 7.32
3rd: 17.16
4th: 21.47
5th: 29.51
6th: 40.81
7th: 60
R: 4.18
7th: 1.0
R: 14.3

The gear ratios of the T-72 gear box are identical to the T-64, although the transmission is very different. Results of the tests of Object 172 tanks in the Turkestan Military District in 1968 showed that the average speed of the tanks on a paved road was 43.4 to 48.7 km/h, and the maximum speed recorded was 65 km/h, presumably achieved by driving down a perfectly straight stretch of highway. The official top speed is 60 km/h.

The gear shift is the same one used in the T-64 and is much more linear than the one in the T-54.

Here is a GIF of the gear shift of a T-64 being operated.

The brakes are of a disc type, hydraulically operated. The T-72 is not capable of true neutral steering, as it can only turn on a false pivot, meaning that to turn the tank on the spot, one of the two tracks are locked in place while the other drives the tank around it. This method of steering is mechanically simple, but inferior to a true neutral steering system where both of the tracks receive power, and one of the tracks is run at the desired speed while the other is run slightly slower in the opposite direction. Besides being slower, false pivot steering creates a huge amount of friction and places more strain on the inactive track, creating tension that leads to the gradual weakening of the track. As such, it is common practice to release the built-up tension in the track by letting the tank lurch forwards periodically during the turn.

The steering tillers (or levers) are hydraulically assisted, so that steering the T-72 is very light and easy even for an inexperienced driver. The synchromesh gearing system enables the driver to steer the tank smoother than on a T-54 when he pulls on either one of the tillers, as the changing of gear ratios in the gearbox is smoother with a synchromesh system than with a typical constant mesh system. The synchromesh system is also less harsh on the gears, thus increasing the lifespan on the gearboxes. Driving the T-72 is a very pleasant experience, according to people with firsthand experience. One of the reasons besides the steering system is the low center of gravity of the tank itself. Being so low-slung, turning the tank simply feels better; after all, a double decker bus doesn't turn quite like a lowrider.

Nevertheless, the tiller system is inherently less ergonomic than a steering bar or wheel. The only advantage of the tiller system over the more complex steering bar system is that it is much easier and cheaper to manufacture, and also more durable - durability being a key factor in the decision to stick with tillers. It is no coincidence that the powertrain of the T-72 has a legendary reputation for reliability.

A little-known fact is that with the mud guards on, it is true that the T-72 can only climb vertical obstacles measuring around 0.85m in height. When they are removed, however, the T-72 can scale obstacles at least as tall as 1.2m (already taller than the tracks) or more. The GIF below shows the tank literally climbing straight up a concrete wall.  

(From 1992-1994 Swedish trials in Northern Norrland). Video credit goes to Ren Hanxue from the Swedish Tank Archives blog.


The T-72 can mount an APU, but only the command variants have one. The T-72AK and T-72BK were both equipped with an AB-1 petrol generator, producing 1kW.  


The T-72 uses full-length torsion bar suspension. Each wheel has its own torsion bar, which runs across the hull floor and to the other end of the hull. The front two torsion bar-wheel hub interfaces have reinforced bolts, since the T-72 is slightly front heavy and so they will bear the brunt of the tank's weight during forward movement, especially across pot-holed ground.

There are six 750mm roadwheels with three return rollers per each side. The roadwheels are die-cast aluminium alloy, with thick rubberized rims. The wheels weigh 180 kg each. The T-72 Ural used an 8-spoked wheel design, but all subsequent models used a 6-spoked wheel. The wheels are evenly spaced but there are quite a few differences in how the wheels are mounted to the chassis, of course.

The maximum vertical range of travel of the roadwheels is 320mm. This is the same as the T-64 and double the range of the T-54/55 and T-62 roadwheels.
The first, second and last roadwheel on both sides are augmented with hydraulic shock absorbers. The front two shock absorbers are highly beneficial as the tank crosses rough terrain, while the rearmost shock absorber is intended to assist recovery when driving through dips and bumps. The shock absorbers are of a rotary type, and are very similar to the ones installed in the T-54.

A cross section of the shock absorber is illustrated below. The device is extremely compact, and is derived from the shock absorbers from the T-54.

The photos below give a good view of the wheels.

The T-72 first came with single-pin RMSh tracks measuring 580mm in width. These tracks have rubber bushings that help reduce vibrations and thus, reduce wear and tear as well as noise levels (though still relatively high). The ground contact length is 4,270mm. A full set of 97 links weighs just over 1,723 kg for one side, and 3,446 kg for a full set of two tracks. These tracks were simpler than the dual-pin tracks of the T-64A and had better traction on rocky and sandy terrain, but had worse traction in mud and other types of terrain and were not as durable.

Old drive sprocket

Newer UMSh dual-pin tracks are available, also measuring 580mm in width. UMSh tracks were specially developed for the T-80 due to the high stresses on the suspension due to the high torque from its gas turbine engine and the tank's high average speed across rough terrain. As such, these tracks were the smoothest and most durable of the three types used by the Soviet Union's three main battle tanks. Installation of the newer tracks requires modified drive sprockets, so only the newer modifications of the T-72 have this installed, including some late production T-72B variants. The interior surface of the track pads are rubber-lined so that the rubber rims of the roadwheels roll on a rubber surface rather than a metal one. This prolongs the life of the roadwheels and also helps to reduce the vibrations transmitted to the tank from driving over rough ground. The reduction in vibrations increases the comfort of the crew and improves the accuracy of the weapons while firing on the move. Another benefit of of this track is its ability to be fitted with asphalt-friendly rubber pads. An entire set of tracks weighs just a hair under 1,800 kg, so the combined weight of a pair of tracks is just under 3,600 kg.

The photo below shows a T-72B3 with UMSh tracks and rubber track pads installed for driving on paved roads. Photo courtesy of Vitaly Kuzmin.

There is a simple mud scraper bolted on to the side of the hull in every T-72, just above the drive sprocket. The scraper helps to prevent loss of traction from excess soil on the tracks, especially sticky mud like clay.

Removing rubber pads on T-90
Throughout the T-72's evolution, it has "fattened up" somewhat, and the largest weight gain was in the T-72B upgrade. While the T-72 Ural and T-72A both weighed 41 and 41.5 tons respectively, the T-72B tipped the scales at 44.5 tons. The T-72B obr. 1989 weighs 46 tons thanks to the installation of Kontakt-5, and Kontakt-1 adds around 1.2 tons.

The T-72 Ural and T-72A exerted 0.83 kg/ of ground pressure, while the T-72B, being heavier for its thicker and denser armour, put in 0.898 kg/ of ground pressure. Compared to its immediate foreign counterparts, the T-72 had little to no advantage in soft terrain, despite being a great deal lighter than all of its adversaries. Against the Chieftain, Leopard 1 and M60A1 of its era, the T-72 Ural and T-72A fared slightly better in this respect, but the T-72B was neither better nor worse off than its more modern challengers like the Leopard 2, Challenger 1, M60A3 and the M1 Abrams. The weight discrepancy doesn't manifest in this regard, but it suddenly becomes apparent when we consider the infrastructure of Eastern Europe at the time, especially the bridges - both permanent and temporary ones - which had stricter weight restrictions. Another advantage of the light weight of the T-72 is that it is light enough to be transported on existing rail platforms, and also light enough to be compatible with the weight limit of the old MTU-55 bridge layers and TMM truck-based bridge layers, both of which were and still are present in large numbers in the Russian Army Engineers.  

In total, the suspension, powertrain, and all the associated systems have a combined weight of 8.57 tons.

If the T-72 were to be trapped in swamps, bogs or in extremely deep snow, it may escape with the help of the eponymous log.

By tying the log to track pins on both right and left tracks as illustrated below, the tracks will drag the log along and under them, thus forcing that section of the track to rise above the mud while simultaneously giving the track something more solid to drive over. This allows the tank to get out of the hairiest situations.

The unditching log was demonstrated in Sweden by an ex-GDR T-72M1 in 1991 as part of a series of tests. Colour photo available on the website.

Here is a video (link) demonstrating a tank unditching itself using the log.


Like the T-62 and T-54 preceding it, the T-72 is capable of crossing deep water obstacles. Safe fording depths are usually cited to be around 1.2m, but water obstacles measuring up to 1.8m deep may be forded for short distances if necessary. Doing so will require the air intakes to be shut off and the partial implementation of the snorkeling feature (engine draws air from fighting compartment, turret hatches are left open, but snorkel is not installed), since the water level would be above the hull. With the installation of the proprietary OPVT snorkel, fording up to a depth of 5m is possible. Pre-fording preparations are necessary in order to do so, requiring the edges of all hatches and the openings of various openings and periscopes to be coated with a thick resinous waterproofing paste, as the water pressure at such depths is simply too much for rubber seals to handle. 

The driver must then turn on the MBP-2 bilge pump. It is located to his lower left side. The bilge pump works on 300W and operates at 3,400 RPM.

Crew members are each given a closed-circuit IP-5 rebreather. The crew must put on the IP-5 before entering water as a precautionary measure.

It comprises a watertight, form fitting gas mask, a chemical respirator chamber containing potassium superoxide (KO2), and a flotation collar. The rebreather uses the chemical reaction between potassium superoxide and carbon dioxide, activated by water from the user's breath reduce the former two to oxygen and potassium carbonate. The freshly produced oxygen gas is mixed into the previously exhaled breath to replenish its oxygen content for rebreathing.


The OPVT snorkel "breathes" for all three occupants, as well as the engine. In the latter case, an air intake fan duct draws air from the crew compartment and routes it to the engine. The suction effect from the intake fan helps to circulate the air inside the fighting compartment as well. 

The normal NBC-capable ventilation system is inoperable while snorkeling, but this does not mean that the crew is vulnerable to such dangers while snorkeling; recall that the crew must don a closed cycle rebreather system before entering water. This means that the crew never has to breathe contaminated air, although the interior of the tank will be unavoidably contaminated. The OPVT snorkel is installed on the gunner's hatch, through a circular porthole, visible in this picture:


Because the hatch can be simply swung open, installing the snorkel is not difficult. The snorkel is fitted with two floating markers during training exercises to indicate the tank's position underwater to help rescue teams locate the tank if it has stopped underwater.

The two photos below shows the snorkel being installed during a river crossing exercise.

Also, the exhaust port must be replaced with a special valve bank to prevent water from entering into the exhaust manifolds.

T-72s equipped with the V-92S2 or V-92SF engines must use different valve units.

All T-72 variants are capable of snorkeling regardless of variant.


All T-72 variants have a total internal fuel capacity of 705 liters, spread across several fuel cells. Two fuel tanks are located on the forward hull on either side of the driver. Another conformal fuel tank is located directly behind the right frontal fuel tank. It also doubles as ammunition racks, and so does the crescent-shaped conformal fuel tank directly behind the autoloader carousel which holds 12 propellant charges. Another 495 liters of fuel is stored in conformal fuel cells located externally on the starboard side fenders. The total fuel capacity is 1,200 liters.

As you can see in the diagram above, the external and internal fuel systems are not interconnected. They each have their own separate fuel lines, but both connect to the same fuel pump.

Being entirely separated from each other, the driver-mechanic is able to shut off and isolate the internal and external fuel tanks from his station. Isolated fuel tanks will be disconnected from both the fuel pump and the fuel return lines, so the fuel within the tank will be left to sit. This can be beneficial in some circumstances, such as when there is an imminent threat of an internal fire spreading. By shutting off all of the internal fuel tanks, the fuel will not leak out as energetically as it is no longer being drawn by the fuel pump or maybe even stop leaking entirely, depending on the specific location of the damage to the tanks. It is also possible for the driver to shut off all internal fuel tanks, and rely on external fuel only if the situation allows it. This keeps the internal fuel tanks full and prevents the tanks from being filled with air (and oxygen), so that the chances of an internal fire are minimized. This also creates the possibility of filling the internal fuel tanks with water,and since the majority of the volatile propellant charges are stowed in conformal fuel tanks, they can become ad hoc wet stowage racks for increased safety. This was probably never done in practice, but based on how the fuel system works, it is absolutely possible.

The video below shows a T-72B in Grozny retreating with some of its fender fuel cells alight. As you can see, the tank is not disabled by the fire and is perfectly capable of moving under its own power to a safe location where the crew can put out the fire with the fire extinguishers carried inside the tank.

Two externally mounted auxiliary fuel drums can be carried on special mounts at the rear of the tank. The auxiliary fuel tank holders are hinged, and may be folded flush to the hull rear when not in use. Each drum has a 200-liter capacity, thus increasing the maximum fuel capacity of the T-72 to a total of 1,600 liters.

The drums connect directly to the fuel system, and both can be disconnected by the driver at the same time by the push of a button. 


The T-72 Ural can travel 480 km on internal fuel alone, or 700 km with external fuel drums. Thanks to improvements in fuel efficiency on the T-72B3, it can travel 550 km on internal fuel alone, or 800 km with external fuel drums despite having the same fuel capacity. As with all automobiles, fuel efficiency decreases while driving cross-country because the amount of engine power needed increases as the harshness of the terrain increases, and so does fuel consumption. 
Because of the T-72's relatively large fuel capacity and high fuel efficiency, refueling the T-72 isn't even necessary for short continuous operations. This greatly eases the logistical burden on the frontlines.


According to user testimonies compiled by the author, the driver's station can be definitively said to be the most comfortable place to be in the T-72. Anecdotes from people who have sat in the driver's seat have reported that the station is spacious enough to let someone more than six feet tall to operate the pedals with a comfortable allotment of legroom. When driving, the driver must hunch slightly forward in order to operate the steering levers, step on the pedals and look through the periscope at the same time. Besides anecdotal sources, we can once again refer to this diagram from "Human Factors and Scientific Progress in Tank Building" by M.N. Tikhonov and I.D. Kudrin as provided by Peter Samsonov, we can see that the driver of a T-72 gets 0.864 cubic meters of space. That's more than the 0.621 cubic meters afforded to the driver of a T-55. The driver enters and exits his station via an oval 530mm-wide hatch located centrally on the hull axis, underneath the cannon. The driver can only exit the tank when the cannon is elevated or oriented to one side.

The driver is provided with a single forward-facing TNPO-168V periscope to facilitate driving. It is a very wide periscope measuring 267mm across - much wider than the driver's head - with a binocular field of view of 38 degrees, and a total field of view of 138 degrees. The periscope provides 31 degrees of vision vertically - 15 below the horizontal axis and 16 above. The view from the TNPO-168V is superb (reportedly). Szabó István, a non-military Hungarian with extensive experience driving demilitarized tanks, described the TNPO-168V periscope as "so wide that it looks like a small window from inside! Forward visibility is excellent for such a periscope! No complaints here". Like all the other periscopes on the T-72, the TNPO-168V is heated through the RTC heater system.

There is a horizontal handlebar below the periscope. It is not part of the periscope. It is simply a handlebar for the driver to hold on to when shifting in or out of the hatch.

The picture below shows the view from a TNPO-168V. As you can see, forward visibility is very good indeed. It helped that the periscope itself is quite short, meaning that the distance between the eyepiece mirror and the aperture mirror is very short, so the "tunnel effect" is minimized. 

Another good example of the excellent visibility from the TNPO-168V periscope is demonstrated in this video (link).
When not in use, the TNPO-168V periscope is stowed away in its aluminium container.

For night time driving, the driver is provided with a TVNE-4B passive-active binocular periscope with a pair of Gen 2 light intensifier moduleThe periscope runs on the tank's 27V power supply system, and can run continuously for eight hours. It is typically kept in its proprietary aluminium box and stowed away by the driver until it is needed.

The periscope is much narrower than the TNPO-168V, but the TVNE-4B still provides a decent field of vision as well as an acceptable viewing distance. TVNE-4B stands for "Tank Driver's Passive Night Vision Optic, model 4 with built-in power supply". According to "Optical Night Vision Devices For Armoured Vehicles" by S.A. Belyakov and published by the Russian Ministry of Defence, TVNE-4B facilitates a viewing distance of up to 100 m in the passive mode under ambient light conditions of 0.005 lux (moonless, starlit night), and 60 m in the active mode using the FG-125 IR headlight. The active infrared mode of vision is very similar to the M24 active IR periscope, but the passive mode offers inferior range compared to the AN/VVS-2 passive periscope, which offers a viewing distance of 150 m. In terms of overall quality and functionality, the TVNE-4B is closer to the obsolete M24 than the AN/VVS-2. The TVN-5 closely mimicked the AN/VVS-2 and offered better performance, but came much later and is used only in limited numbers, so we will not be examining it in this article.

The periscope can be directly inserted into the periscope slot for the TNPO-168V without any modifications, but a plastic spacer is slipped on top of the aperture head to give it a proper fit and to prevent water and mud from ingressing through the gaps. The plastic spacer can be seen at the very top of the exploded diagram below. The periscope only has a decent horizontal field of view of 36 degrees and a vertical field of view of 33 degrees, but more importantly, the driver has depth perception. Most night vision periscopes have two eyepieces but only a single aperture, so the driver lacks stereoscopic vision and thus lacks depth perception.

Although TVNE-4B only has a 60 meter view range in the active mode using only the hull's single small IR headlight, it may also pick up infrared light from the turret's three IR spotlights. Indeed, the single FG-125 IR light located just beside the auxiliary/night sight on the turret is meant to augment the driver's viewing distance when operating in the active imaging mode and also to provide a source of light when the tank is wading across water obstacles. The periscope has 1.12x magnification. Compared to the AN/VVS-2 passive driver's periscope, the TVNE-4B has poorer performance all across the board, except that it enables stereoscopic vision.

This video shows the image intensifier of the TVNE-4B periscope in use. Here is a screenshot from the video:

From the video, the claimed 100-meter viewing range of the periscope is shown to be valid, and it appears to be more than adequate for nighttime driving. Here is another video of the TVNE-4B in use, this time in a T-90. The IR headlight on the left side of the upper glacis is used for illumination in this instance, and the range of vision seems to be worse, but this could be due to the low video quality (240p).

There is an accessory windshield that may be attached to the outside of the hatch. Its main purpose is to protect the driver from bugs and dust while driving in non combat conditions.

It is not unique to the T-72 in any way. Many other Soviet vehicles can mount these windshields.

The driver enters through a pill-shaped hatch. Two TNPA-65A viewing prisms are embedded in it, one looking in the 10 o'clock and the other in the 1 o'clock direction. Looking through them requires the driver to look upwards. This layout is generally far less convenient than the more commonly encountered bank of three viewing prisms found on the T-80 and Abrams as well as most others. The TNPA-65A periscopes are very narrow, almost slit-like. It is difficult to see very much other than the tracks and part of the road, but it would also be very hard to hit or damage them with machine gun fire. The TNPA-65A periscopes are meant to check the corners of the tank only. They are far too limited for driving during combat.

As mentioned before, the TNPA-65A periscope provides 14 degrees of binocular vision horizontally, and only 6 degrees of vertical vision.

The driver's hatch itself is 20mm thick. The rubber seals make them completely watertight down till a depth of around 1 or 2 (relative to the height of the hatch, not the turret roof). Unfortunately, the seals on the TNPO-168V periscope are not nearly as dependable. Being mostly watertight, the tank can ford streams as deep as 1.2m or deeper without the danger of excessive water ingress.

Steering the tank requires the use of two hydraulically assisted tillers, which are located on either side of the knees of the driver. Though the tiller steering system can be considered one of the more antiquated aspects of the T-72, it's worth noting that many of its rivals like the AMX-30, Chieftain and Challenger used the same system as well. However, most main battle tanks had already grown out of tillers and progressed into steering wheels even by the 1960s, like the M60 and Leopard 1 did. The Leclerc and Leopard 2 both use steering wheels and the Abrams tank uses motorcycle handlebars. The only exception is the Challenger 2, which (shockingly) still retains a tiller system as well.

The driver sits on a bucket type seat, with bolsters on each side. The seat can be adjusted for height in order to accommodate persons of a wider range of height, and also to raise the driver high enough to peek over the hatch. It has been reported to the author that the driver of a T-72 gets quite a lot of legroom, and that it is quite comfortable, more so even than the commander's and gunner's stations.

Like the commander and gunner, the driver's "air conditioning" comes in the form of a DV-3 plastic fan.

All of the driving-related indicator gauges are placed on a board to the driver's left, like in previous tanks like the T-62, T-54 and T-34. The placement isn't exactly convenient, but looking at them while driving (in any tank) isn't really very necessary anyway.

Behind is the left front hull fuel tank. The fire extinguishers for the hull's automated firefighting system is located underneath it. 

Over at his left foot there is a rather rudimentary GPK-59 gyroscopic compass for directional navigation. It is particularly useful when driving underwater when nobody in the tank has any scenery to refer to for a sense of direction. The use of gyrocompasses can perhaps be labeled as a rudimentary form of an Inertial Navigation System (INS), advanced versions of which are often present in modern combat vehicles due to their independence from outside input contrary to a GPS-based navigation system. Sadly, the T-72 has not received either in any of its iterations.




T-72 Ural Armour Against NATO, Mikhail Baryatinsky
(Михаил Барятинский, Т-72 Уральская броня против НАТО)

Vehicles and Weapons Magazine, 2006 November issue
(Журнал Техника и Вооружение 2006 год, 11)

V.A Grigoryan, Tank Armour, 2007(Защита Танков, В.А Григорян 2007)