A discussion about Tungsten Cored Ammunition (APCR/HVAP, and APDS)

I’m starting this thread to lay out the information that I have on Tungsten cored projectiles. I had done a thread similar to this on the old forums, a long time ago. But, I’ve also found and learned a lot more since then. I’ve also learned how to properly model the projectiles in programs like Fusion 360, which has helped a lot… Especially for those projectiles where there is no information about the mass of the core.

This post is still kind of a work in progress, as I may add some images and documents to it, as I have time.

I’ll start with APCR and HVAP.

In general, I refer to the classic arrowhead-shaped composite rigid shells, as APCR.

They look like this.

Click to see Arrowhead type APCR shell diagrams

76mm BR-354p Var APCR

5cm Pzgr40 Ausf A Diagram

Then there is the more regular projectile-shaped APCR that is often referred to as the HVAP design.

They tend to look like these.

Click to see HVAP type shell Diagrams

As you can see, there is some variation in design and material choice. But, the overall projectile shapes are very similar. For example, Soviet and German arrowhead APCR rounds, usually used steel bar stock as the carrier shell, and either an empty windscreen/ballistic cap or a paper/plastic filled one. Or in some cases, a magnesium-aluminum alloy cap to provide a flash upon impact.

The German HVAP-type rounds were usually steel bar stock carrier shells with bakelite core nose pads and aluminum or steel ballistic caps. And the US had two variants. The early variant, consisting of M304 and M93, has an aluminum body, ballistic cap, and core retention cap, with a steel base cap. And the later lightweight variant, consisting of M319, M332, and several prototypes, had an all-aluminum alloy carrier shell. Later Soviet APCR rounds copied the German HVAP design for the 85mm and 100mm guns… However, it is exceedingly difficult to find cutaway images of those two rounds.

Now on to APDS designs.

Originally, I had thought that there were only 3 APDS core generations and that that was the main factor affecting penetration ability. However, I think I have a greater understanding of what’s happening now, than I did previously. There are a few more nuances to it, as the double conical core design appears well before L28A1/L36A1/M392 in US APDS core designs.

So let us start with your early normal Ogival core-type APDS rounds. This would consist of rounds like 17pdr APDS Mk1, 20pdr Mk3, 76mm M331, 100mm 3BM8, 122mm 3BM7/11, 120mm L1, etc…

Click to see Ogival core APDS Diagrams

76mm M331 APDS

100mm 3BM8 APDS

During the mid 50’s the US started designing and testing APDS projectiles using double conical nose designs, as a means to try and increase high obliquity penetration. They were doing this alongside testing of APFSDS, some of which used variants of said cores in their penetrator bodies. Projectiles based on these cores generally had similar vertical penetration to Ogival cores of the same mass and diameter.

Click here to see double conical core designs from the mid 50's

Note here, earlier in the text of the source document, it states that this projectile uses the same main penetrator body as 90/40mm T320.

Then we have the next version of APDS used by NATO countries. These all use the same base double conical Tungsten carbide core. But, they now incorporate a Tungsten alloy nose pad. This is where is where vertical penetration gets sacrificed, at the expense of increased high obliquity penetration reliability. The earliest versions of these rounds didn’t include the steel cup(tilt bearing) between the core and nose pad, the later ones did. In this case, at low obliquity, the Tungsten alloy nose absorbs some of the energy from the Tungsten carbide core, that would otherwise go into raw penetration. However, that same nose helps to increase high obliquity penetration.

This version of NATO APDS covers 105mm L28, L36/M392, DM13, etc.

Click to See this Version of APDS

L28A1-M392A2 Diagram

That brings us to the final version of NATO APDS. Where the transition from a Tungsten Carbide core to a Tungsten Alloy core was made. Here, a bit more vertical penetration was lost, due to Tungsten alloy being softer than Tungsten Carbide. However, the performance at high obliquity increased considerably. Also due to the difference in how the two materials perforate armor steel. One pushes its way through, and the other erodes its way through, losing mass as it does so. There are some other minor changes, like the change to a round tip, removing the need for the tilt bearing.

This version covers 105mm L52, M728, and 120mm L15.

Click to see Final version of APDS diagram

L52-M728 Diagram

An Ansys Simulation Video that was done by SY Simulations to show the difference between Tungsten Carbide and Tungsten Alloy, especially against spaced armor plates. Velocity for both rounds is set for just under 800m.

Click to see Simulation Video

THE SUPERIOR TUNGSTEN PROJECTILE | Tungsten Carbide vs Tungsten Alloy Armour Piercing Simulation - YouTube


Hvap/APCR is massively underperforming in war thunder.

There are american and German tests for the penetration of german 20mm and 30mm hvap rounds and these were all similar to the pen ingame before the nerf. American tests showed that 30 mm hvap round could pen 101 mm (even more than ingame).

APCR is another topic. These are trash aswell and don’t reflect battle reports or tests.
There would be no use in even developing these rounds if aphe is always better.


I don’t have anything on the 30mm, but I do on the other rounds.

2C7 Perforation of Armor by German Projectiles

2C7 Perf of German Projectiles

I’ve also been considering making a suggestion to change the calculator for HVAP/APCR and apds from the current weird version of Demarre, to something based on what I’ve put below.

2C6 Perforation of Armor by Tungsten Carbide Cored projectiles

2C6 Perforation of Armor by Tungsten cored projectiles

The upside to that equation system is it can be normalized to a specific armor hardness bhn. Which, could possibly allow all penetration value calculations in the game to be normalized to the same value as well. And then it would be easy to set armor hardness modifiers due to being able to use this calculator and the L-O calculator to see the difference between the Brinell hardness values.

The downside is, it’s fairly advanced math, and I’m not sure how easily it could be implemented into the game’s system.

Could you potentially show how the graphs on the 2C6 spoiler work? I was immediately confused.

Is this thread meant to challenge the ridiculous way Tungsten cores behave currently? Like the long standing APCR being atrocious against sloped armour, and the more recent, APDS (bar 3rd gen) being useless against a few layers of steel. Or the bizarre disparity between L28 and M728 (L52) in game, where despite it being basically a material change from Carbide to Heavy alloy it loses a lot of flat penetration to the point if it weren’t for the aforementioned issue, L28A1 might be preferable.

It seems that precisely because of that material change flat penetration is sacrificed in order to achieve higher slope penetration. However, to what degree that sacrifice is I don’t have a clue.

Edit: Here’s the information from the old forum page that talked about these changes.

Unlike the hard alloys, penetrators made of heavy tungsten alloys (HTA) lose mass and significantly change their geometry when passing through an armor, while a kind of influx is formed on the nose of the penetrator, which significantly increases the collision area.

Due to the washout of the core material and a larger collision area, and other things being equal, tungsten alloy penetrators have a significantly lower penetration ability when firing at the barrier at acute angles, however, when interacting with obstacles at larger angles, the opposite trend is observed. Because of the plastic deformation during penetration, HTA penetrators change their trajectory much less than hard-alloy penetrators, and this quality becomes useful at large angles of impact with the obstacle.

Some of it is to challenge some of the notions people have about APCR/HVAP, and APDS sloped penetration performance. Especially on the Ogival core versions. The Ogival core APDS is essentially a glorified HVAP projectile, that retains velocity better due to reduced drag from having less cross-section and better aerodynamics. Which would make it appear to have better high obliquity performance at range.

The current performance of L52/M728 I believe is due to a report in the past from a Soviet document, where they tested western APDS rounds captured from one of the Israeli conflicts. The problem with the source is, from what I can tell, they used the same standards of penetration for both types of cores, which wouldn’t necessarily produce proper results. The reason the US uses protection criterion with Tungsten alloy cored projectiles, is because they erode during penetration. If they were using criteria that required a certain portion of the projectile mass to pass through the plate to be considered a “complete penetration” then the Tungsten alloy core would appear to penetrate less at low obliquity due to losing mass to erosion during penetration.


I don’t quite understand how this correlates to it having better slope performance.

Also this is probably the simplest, yet most mind blowing explanation that makes complete perfect sense. Because I remember seeing that source, and stuff like M392 seemed to have performance that correlated to other sources, but then M728 just seemed way off.

Edit - Found it.

Russian 105 mm APDS testing

russian testing 105 mm APDS

I had to remember how to use the graph myself. As there are two ways to do so. One way is choosing a velocity, and then seeing what plate thickness range a given core could penetrate 5% chance on the high end, 95% chance on the low end. And the other way is to select a plate thickness and a given core, and then use the graph to determine what the ballistic limit velocity range would be.

However, it’ll be easier for me to show with lines and such on an example image, than it would be to explain, which will take some time to prepare.

Yeah I’m not sure how L52 could penetrate so poorly in flat penetration. L15 suffers it too and Chieftain is not a tank that needs more suffering.
Also I know this is almost completely unrelated but I’m super suspicious of all T-54/T-55/Type 59s new 8 rounds per minute fire rate which is identical to M48 and Chieftains.
The documentation of the L11 guns acceptance said 10 AIMED rounds a minute was achievable, whereas the British T-55 tests from around the same time gave a static T-55 an aimed rate of fire of 1.9 rounds per minute. I mean different tests sure so different standards but gives some idea of the difference, because it’s not just “this is what it’s rate of fire was in this test” it’s “we were baffled by this slow rate of fire”. At a time when Chieftain was in service.

So there’s absolutely no justification for it IMO, T-55 is using the extreme upper bound of reasonable fire rate while Chieftain and M48 especially use a lower bound. And IK it’s a game Balance etc, but these T-54 49 went from 7.3 to 8.0 because of this change! It’s also makes the game blander when 3 equivalent cold war tanks now have identical rates of fire, when there’s very little justification for that.

1.9 seconds for the T-55s?

Here’s an example of the first method.

In this method:
1: You choose a velocity at which you want to figure out about how thick a plate the round can penetrate at that velocity.

2: Then you pick the round.

3: You trace a line from the velocity point, through the core caliber density point of the round, to the first vertical y-axis line on the main part of the graph.

4: Then you trace a horizontal line from the point where the first line meets the graph’s first y-axis line, across the probability band for the angle you’re trying to figure out. In this example, I’m looking at vertical or 0-degree penetration.

5:Trace vertical lines down from the points where the horizontal line intersects the inner and outer edges of the shaded area to the x-axis line labeled e/d, this is the plate thickness to core diameter ratio range.

6: Trace lines from the e/d intersect points, across the core diameter point for M93, to the lower line labeled plate thickness inches. The resultant thicknesses are a 5% chance to penetrate the higher thickness, and a 95% chance to penetrate the lower thickness. As the document states, there’s a 90% probability spread.

1 Like

No 1.9 rounds per minute, sorry a typo.

So for M93, at 3200 ft/s or 975 m/s, V5 would be 8(2/8) inches or 209.55 mm and V95 would be 7(1/8) inches or 180.975 mm, giving a V50 of ~195 mm.

Even though that’s ~50 m/s lower than what it does from the muzzle it’s still higher than what it does in-game… After checking TBDV3, M93 reaches 3200 ft/s at roughly 340 yards, or roughly 310 meters.

Edit: I tried 3400 ft/s for M93, which is 1036 m/s, and ended up with a V5 that was out of the graph but seems it would line up exactly with 9(1/8) inches or 231.775 mm, and a V95 of 7(6/8) inches or 231.775 mm, so V50 would be 214~ mm. 3400 ft/s is M93’s muzzle velocity if I’m not mistaken.

Ah, I believe that makes more sense than a 1.9 second reload. 😂

If you DeMarre the 935 m/s results to 1036 m/s, you get a V5 of 242mm, a V95 of 210mm and a V50 of 226mm.

The V5 results are close to the navy chart numbers for M93, which looks about 9.6 inches.

I assume you’ve seen these tables before, from that Terminal Ballistic data volume III thing (and how to read them hah). Though the more extreme angles are just estimates it’s clearly underperforming in game at 0 and 30 degrees, I wonder why this has never been fixed? Possibly because when they added Germany, USSR, USA they dolled APCR out to nearly every single tank and making it work properly might completely flip ground meta on it’s head, especially if it was doubled with making APHE more realistic.

I know 105mm T29 APCR off this British sourced exert is estimated to get 15 inches normal point blank and 11.3 at 30 degrees, but in game it gets 11.46 inches and 8.5in respectively. It gets even worse beyond that, it’s estimated penetrate 7.6 inches at 45 degrees doesn’t work in game, (i used TigerIIH 7.35inch turret front as reference) and it couldn’t penetrate until 31 degrees! Meaning that round significantly suffers in game.

I have TBDV3 as a PDF on my computer and do know how to obtain the results.

However, what I’ve read from @MiseryIndex556 and some others is that the tables on TBDV3 for M93 and M304 HVAP rounds were obtained using tests on softer armor, so the values are a bit inflated.

Still, even with that in mind US APCR in general does just underperform, and really the calculator is to blame.

As an example, let’s look at 1249 m/s M332 and the long 90 mm T44 APCR, fired at 1143 m/s. That is a 100 m/s increase for M332, yet in-game it has lower flat penetration at close range, despite the fact that it uses the exact same core as T44 (both projectiles use a 3.6 kg, 47.6 mm diameter tungsten carbide core).

Edit: I forgot to mention that 90 mm APCR rounds have a 38.1 mm core instead of the correct 47.6 mm, which means that they actually have a higher penetration value from the calculator than they should…

In game performance screenshots

90 mm T44 APCR

90 mm M332 APCR

This just shows that the calculator is just… messed up to say the least.


I thought I’d also throw these in as examples of the current tungsten carbide layer issue. As it seems a planned change very poorly implemented and not really explained to the community at all as far as I’m aware.

90mm T44 shot hugely affected by multiple layers of armour.

And even APCR being completely stopped by layers of armour with no gaps in-between them.

1 Like

Can you send me the document? I dont have the whole thing.

I’m not sure if the charts are normalized to softer armor. It makes sense if it is but I’ve read they were normalized to 237 BHN.