Armor quality of Panzer III and IV

[Avalancheon]

The hit #4 was made with the older sharp-tipped shell, btw.

Thats not what tankarchives translation says. He claims that shots #4 and #5 used 'pointed type' shells, I.E., the BR-471Bs.

That's a T-34.

Okay, I accept your correction.

It's interesting to compare the performance of soviet shell vs KT armour with US testing of the same BR-471B in 1972:

We have 150mm/50° RHA of 255-269BHN. The shot #3 with a BR-471B managed to spall the rear face of the armour, I shall consider the lowest border of protection limit for this target as 771m/s (2528fps)

The US graph indicates the thickness defeated at 2530fps to be 4+5/8" or 117mm/50° of 241-277 BHN RHA. (US Protection Criteria)

So your implying the Tiger IIs glacis offered less resistance than it should have. How can you say that for certain, given that the armor was only spalled? Isn't that a kindof gray area in terms of what you can calculate as far as ballistic resistance?

The lower glacis 100mm/55° RHA of 255-269BHN. Again the same shell.
The shot #6 at 653m/s(2140fps) results in a cracked bulge. The actual P. limit is likely to be a bit higher than this.
US chart gives up 3+1/16" or 78mm/55° against 269-311 BHN RHA

I'll leave the chart here if anyone wants to check the numbers. https://imgur.com/5uOUxBm

Shots #6 and #7 against the lower front plate used 'flat type' shells, I.E., the BR-471. Thats why they failed to penetrate.


Edit: I read somewhere that at a distance of 100 meters, the 122mm can only pierce 120mm of armor sloped at 50 degrees. If thats true, then the fact that the Tiger IIs glacis didn't resist the shots is somewhat surprising. Perhaps it can be attributed to the cold temperatures at the Kubinka testing grounds, which made the armor act brittle.

[Mobius]

The hit #4 was made with the older sharp-tipped shell, btw.

Thats not what tankarchives translation says. He claims that shots #4 and #5 used 'pointed type' shells, I.E., the BR-471Bs.

No, sharp=pointed. BR-471.

The US graph indicates the thickness defeated at 2530fps to be 4+5/8" or 117mm/50° of 241-277 BHN RHA. (US Protection Criteria)

So your implying the Tiger IIs glacis offered less resistance than it should have. How can you say that for certain, given that the armor was only spalled? Isn't that a kindof gray area in terms of what you can calculate as far as ballistic resistance?

This has no equivalence. You can't use the BR-471B graph to show anything of a BR-471 shot.

[Peasant]

Ditto what Mobius said. I know russian.

So your implying the Tiger IIs glacis offered less resistance than it should have. How can you say that for certain, given that the armor was only spalled? Isn't that a kindof gray area in terms of what you can calculate as far as ballistic resistance?

The interaction between a shell and a given armour with known physical properties will mostly return the same results, this is the key concept behind the existence of a fixed Ballistic Limit. It is possible to extrapolate the different B.limits from knowing one. Look here for more information: viewtopic.php?f=47&t=229641

[critical mass]

My reading largely concurs with the observations made by Peasant and others.
Particularely the method to check for hit#3 as unaffected (notice that the disc ejected by hit#4 is large enough to partially cover adjacent hit#3) and derive P(L) limits accordingly to be compared with a different source. (less so the nose plate hits, see below)

There are, however, caveates, I´d whish to outline:

(A) It remains unknown whether or not the 1960´s-70´s US test shoots with BR471B against 1960´s RHA were conducted at such subzero ambient temperature conditions, too. This presumption, however, is a necessarity when comparing back plate failures, which are heavily correlated with Izod toughnesses.

(B) MIL-S-1260 was developed after ww2 and with the various lectures of wartime RHA production in mind. It may not be comparable in this regard with ww2 vintage material, particularely in regard of the brittle temperature
(but so does the BR471B-if manufactured to 1950´s higher hardness standarts)

(C) the comparison is of test data vs smoothed test data interpretation. A more direct comparison would be possible if the APG DPS-647 report would be consulted, which contains the individual test data. While we know -in general- how the graphs were generated, we do not know anything about variances, which is a significant point when one has only relatively few valid datapoints to start with.

he lower glacis 100mm/55° RHA of 255-269BHN. Again the same shell.
The shot #6 at 653m/s(2140fps) results in a cracked bulge. The actual P. limit is likely to be a bit higher than this.
US chart gives up 3+1/16" or 78mm/55° against 269-311 BHN RHA

I am not convinced that shot#6 resulted in a cracked bulge. The damage to the plate was confined to the surface, 290mm x 130mm, 60mm deep. It´s a dent. I don´t see any evidence for a back bulge, implying that the limit was far from reached.

Hit #7 (also with BR471) happened to hit the same plate at 764.7m/s and only resulted in a dent with a small bulge.

---

irregardless of how one would view the effect of these points, it´s necessary to understand that plate quality comparisons cannot be generally drawn from these sort of comparisons but are rather specific cases, relevant for shear affective failures. The failure mode we discuss with BR-471B (flat nosed, ballistic capped) is not intact projectile + ductile holing but plugging with broken projectiles. That beeing said, I do not intent to state that these attempts are futile, they are instructive because they may modify our prior asumptions by creating new posteriors if done correctly.

It´s entirely possible that the relatively soft MIL S1260 RHA plate will act favourably in these specific conditions (against BR-471B) while acting less favourable in different penetration mechanics (f.e. ductile holing). Remember, the plate gives in in whatever resistence (to plastic deformation, to shear, to cracking) is depleted first for any given specific type of attack (f.e. compare the TIGER2 tests at Aberdeen which resisted 105mm T29E3 HVAP striking glacis and nose plate).

[whelm]

It´s entirely possible that the relatively soft MIL S1260 RHA plate will act favourably in these specific conditions (against BR-471B) while acting less favourable in different penetration mechanics (f.e. ductile holing). Remember, the plate gives in in whatever resistence (to plastic deformation, to shear, to cracking) is depleted first for any given specific type of attack (f.e. compare the TIGER2 tests at Aberdeen which resisted 105mm T29E3 HVAP striking glacis and nose plate).

Did any details ever come to light on the specifics on that HVAP test? distance, and striking velocity and such. HVAP as well will act quite differently (poorly) on high oblique plate so that's not quite a fair comparison to an AP round.

In the 50's for example with the new 105mm ammo program they upped the velocity of the T29 HVAP to 1280 m/s. This allowed it to penetrate 127 mm @ 55 deg at a range of 1200 yards with a BL of roughly 1128 m/s against 269-285 BHN plate.

So the older "standard propellant" slower velocity HVAP round would do about the same at point blank.

[Peasant]

Good to have you in this discussion, CM. I have seen you posting comments on tankarchives blog and clearly you know your stuff.

he lower glacis 100mm/55° RHA of 255-269BHN. Again the same shell.
The shot #6 at 653m/s(2140fps) results in a cracked bulge. The actual P. limit is likely to be a bit higher than this.
US chart gives up 3+1/16" or 78mm/55° against 269-311 BHN RHA

I am not convinced that shot#6 resulted in a cracked bulge. The damage to the plate was confined to the surface, 290mm x 130mm, 60mm deep. It´s a dent. I don´t see any evidence for a back bulge, implying that the limit was far from reached.

Hit #7 (also with BR471) happened to hit the same plate at 764.7m/s and only resulted in a dent with a small bulge.

Thanks for directing my attention to the shot #6. Actually I've made a mistake in translation. It doesnt say "cracked bulge" but rather "a bulge with a crack partially separating it from the plate" which looks like the start of a plug ejection, just like one would expect to see in a penetration of a blunt headed projectile.

Hit #7 has landed on the connection between the upper and lower glacis, so I've excluded it from this analysis.

It´s entirely possible that the relatively soft MIL S1260 RHA plate will act favourably in these specific conditions (against BR-471B) while acting less favourable in different penetration mechanics (f.e. ductile holing). Remember, the plate gives in in whatever resistence (to plastic deformation, to shear, to cracking) is depleted first for any given specific type of attack (f.e. compare the TIGER2 tests at Aberdeen which resisted 105mm T29E3 HVAP striking glacis and nose plate).

MIL-S-12560 is not "relatively soft", in fact its very similar to the WW2 late german vehicle RHA. I've posted the specs earlier, you can see for yourself.

(A) It remains unknown whether or not the 1960´s-70´s US test shoots with BR471B against 1960´s RHA were conducted at such subzero ambient temperature conditions, too. This presumption, however, is a necessarity when comparing back plate failures, which are heavily correlated with Izod toughnesses.

(B) MIL-S-1260 was developed after ww2 and with the various lectures of wartime RHA production in mind. It may not be comparable in this regard with ww2 vintage material, particularely in regard of the brittle temperature
(but so does the BR471B-if manufactured to 1950´s higher hardness standarts)

(C) the comparison is of test data vs smoothed test data interpretation. A more direct comparison would be possible if the APG DPS-647 report would be consulted, which contains the individual test data. While we know -in general- how the graphs were generated, we do not know anything about variances, which is a significant point when one has only relatively few valid datapoints to start with.

c) point is a valid criticism. I dont know enough about metallurgy to say anything about the rest.

irregardless of how one would view the effect of these points, it´s necessary to understand that plate quality comparisons cannot be generally drawn from these sort of comparisons but are rather specific cases, relevant for shear affective failures. The failure mode we discuss with BR-471B (flat nosed, ballistic capped) is not intact projectile + ductile holing but plugging with broken projectiles. That beeing said, I do not intent to state that these attempts are futile, they are instructive because they may modify our prior asumptions by creating new posteriors if done correctly.

I see what you mean, but you cannot just that say that "the armour is good because it would've performed well in a controlled setting against the type of projectiles it was optimized to defeat". Most of the Allied WW2 AP shells would've broken their nose when hitting a 50°sloped plate, resulting in a blunt penetrator similar to soviet shell.

Still regardless on how this armour would've performed at higher ambient temperatures, there is no denying that its bad feature to have in tank armour, when all it takes is to expose it to -10°C (average central european winter) for it to loose about 20% of its effective thickness. Were other WW2 armour steels just as bad?

That's all for now.

[critical mass]

There have been some questions, I´d like to adress with the memo below.

Thanks for directing my attention to the shot #6. Actually I've made a mistake in translation. It doesnt say "cracked bulge" but rather "a bulge with a crack partially separating it from the plate" which looks like the start of a plug ejection, just like one would expect to see in a penetration of a blunt headed projectile.

That´s one posibility, and one might add, when viewing from the perspective of expected plate failure modes here, a probable one.
In plugging/discing, the projectile punches out a large conical plug (or wide disc) of armor from the plate back when it hits with enough force to hole the plate large enough to allow the projectile to have a chance of penetrating the plate. If the projectile has not sufficient power to exact a sufficiantly large hole, the projectile rebounds, broken or not (compare f.e. 152mm hit#9).
In similar photos, but with a back of the plate shown, where the projectile did not quite penetrate (no through crack formed like here), there are obvious wide bulges surrounding the impact point before cracks are formed, in some cases with a ring-shaped crack surrounding the impact that, if the striking velocity had been higher, would have caused that circular area at the plate back to tear free.
http://tankarchives.blogspot.com/2013/0 ... m-gun.html
The projectile acts like a hammer, punching out the steel in front of it´s projectile nose as a cylindrical or cone-shaped ring around the projectile nose. With the blunt nosed WWII-era soviet APBC projectiles, or inferior sharp tipped AP which suffer break up, the nose tip virtually never sticks out of the plate back prior to the plug being torn entirely free of the plate (as would be the case with more ductile events).
The point remains that there was no back bulge, which is initially associated with plug formation even before cracks are formed (compare hit#16) and this suggest to me that this impact was very much below any actual Holing Limit. It´s therefore not representative for the protection ballistic limit.
How much more energy is required to punch out a plug is not clear to me from this hit alone (see more below).

MIL-S-12560 is not "relatively soft", in fact its very similar to the WW2 late german vehicle RHA. I've posted the specs earlier, you can see for yourself.

The document You referred to was revision -B to MIL-S-12560 and -that was my thought- may not be identic to the original specification. However, I have found recently that differences between revision-B and the older standart are minor and therefore happily stand corrected.

I see what you mean, but you cannot just that say that "the armour is good because it would've performed well in a controlled setting against the type of projectiles it was optimized to defeat". Most of the Allied WW2 AP shells would've broken their nose when hitting a 50°sloped plate, resulting in a blunt penetrator similar to soviet shell.

I think this, too, is a fair comment. The only ww2 projectile to have a chance of staying intact against this target (100mm @55°) is the 12.8cm Pzgr43.

Still regardless on how this armour would've performed at higher ambient temperatures, there is no denying that its bad feature to have in tank armour, when all it takes is to expose it to -10°C (average central european winter) for it to loose about 20% of its effective thickness. Were other WW2 armour steels just as bad?

I don´t think that the presumption is valid here. First of all, the difference would be 20% if the hit#6 resulted in a holing, which it did not. So the difference is smaller by an unknown fraction (more below). 2nd, I have some temperature strain rate/Charpy graphs and a difference in impact strength in order of 20% was not uncommen when room temperature was compared with -10°C (f.e. oil tempered PP792 had 7.9mkg/cm² at room temperature and 5.1 mkg/cm² at -10°C). It´s only after ww2 that standart testing temperature was lowered to -40°C in combination with metallographic studies of the fracture surface in order to reliably test for and expose poor lower temperature brittleness effects on AFV armor plate (except for the one and other research paper one jointly comes across to touch this question).

----
That beeing said I have invested the better of roughly five hours of work to examine the nose plate hits in Kubinka. All eight hits with AP and APCBC-HE (only 88mm Pzgr39) were in direct line to the target, which helps eliminating a source of variance. Because the target plate is therefore identic (100mm at 55° to the perpendicular) and presuming that the failure mode is identic, too (discing with broken up penetrators- which is likely true for all attacking projectiles,88mm pzgr39 included)), it may therefore be justified to run a regression analysis (see attachment).

I have choosen De Marre K functions as a proxy of plate resistence because this is how ARTKOM assessed armor tests in ww2 (see Krogfuss article of DeMarre in Artkom for the details). While we wouldn´t do it today (De Marre K has a significant scaling issue) it sheds light on how they did it in ww2.
What can be observed is that there testers intended to obtain a lower and higher bracket with the De Marre K scaling with T/D ratio (smaller calibre need more impact force to obtain a holing than larger ones). A number of hits can be viewed as fair hits on the plate, and a number as invalid hits too close to one another, or to a plate joint, or to a torn free plate joint. Notice that all these cases of invalid hits tend to degrade the plate resistence (discontinuities, free edge effect, communition effect, triaxial stress concentrations). Yet, not all resulted in a holing but rather in clean scoops with or without bulges.

The only holing event, which was the result of a fair hit is hit #9, a 152mm AP striking at 588.9m/s. The result was a disc thrown from the back plate and the broken up projectile rebound from the plate. the holing diameter was not large enough to let the 152mm projectile body pass (sub calibre hole). However, it certainly qualifys for A(BL). The 1st BR471 failed to make a hole at K=2820).It also shows why the 2nd BR471 was fired at K=3300, this is close enough to the upper border to represent a severe test for the armor. While the hit was very close to the joint with the 150mm upper glacis (10mm below the edge on the nose plate), which by virtue of the interlocked design and constructive angle gave peripheric support to conical plug formation, there again was only a scoop. This is somehow significant because the projectile (BR471) would first turn away from the plate, and then into the plate, which would tend to clear the actual penetration channel from the 150mm supporting glacis plate.
I personally, would have decided to repeat this hit, but the testers went on with the 100mm BS-3 which obtained a good, inverse straddle. This was probably enough to fix the trend by using the contemporary De Marre K procedure, the final 88mm hit just again confirmed the lower boundary (lower boundary with no holing events was confirmed by 122mm, 100mm and 88mm hits, making it a fairly robust baseline).

I have added a weighted, smoothed (Loess) interpolation line with probability brackets. The invalid hits are counted but weighted only for 50% individually in this graph.
However, in view of the limited data aviablility, one will need to stress that the trend described between both lines is a range of mixed results, while the trendline is the average, thus, the smoothed data interpolation.

in view of the so established trends one might notice, that for the 122mm BR471, a De Marre K=3100 would meet a 50% confidence level to give some holing events (Army(BL)) at K=3100, which equals a striking velocity =719.7m/s or 2361fps for what would be the US protective BL for this plate accordign to it´s respective smoothed interpolation. This certainly isn´t such a strong deviation from the US graph (say, approx.~10-12%).

Notice that the Army (HL) is different from a calibre sized Holing (where a lot of damage can be inflicted behind the plate, unlike shown here) and that the HL is different to the NB(L), where the complete passage of the projectile can be expected. In plugging /discing events, the delta between Holing limit and NB(L) -perforation limit- can be as much as twice as large as in ductile hole formation events (f.e. 15% instead of 8%).

hope it helps,

[Avalancheon]

Very thorough analysis. But I'm struggling to piece it all together.

Does this prove or disprove Pheasants belief that the glacis was offering a resistance equal to only a 117mm thick plate (rather than its actual 150mm thickness)?

Or was the plate in the American tests acting in a sufficiently different manner from the Tiger IIs glacis that the comparison is invalid?

[critical mass]

I was strictly referring to the 100mm nose plate and haven´t looked into glacis yet. Note also that my memo is representative for my point of view in light of the evidence at hand. I am not convinced that there are sufficient data to jump to conclusions in one direction or another.

For what´s worth, the soviet, somehow higher hardness RHA plating of 110mm strength, inclined 55° of the IS-3 pike glacis was defeated at 900m range table range by 122mm BR-471 at 40° & 320°, where the netto impact angle matches the constructive angle. Presuming that the 1944 range table´s were used here too, which seems a reasonable presumption to me, then we are dealing with a plate resistence of De Marre K=3000-3020 (ca. 746m/s striking velocity) for perforation (not holing limit!). This indicates to me, to put it mildly, that the soviet 110mm RHA plate was exhibiting not quite the same resistence to either holing or projectile passage as did the 100mm TIGER2 nose plate when fired at by the same gun/projectile.
When compared with 1960´s MIL-S-12560 RHA smoothed test data interpretation (again, caveate present is the hard test data vs smoothed curve with unknown variances), the Protection limit of a presumed 4.33" US RHA plate at 55° would have narrowly exceeded even the A-19 muzzle velocity (no holing or penetration possible here) and the P(L) at the 900m downrange´s, ca. 2450fps is rather 3.75" RHA, very much considerably below the 4.33" thickness exhibited on the IS3 pike nose. Here we have a case that the holing limit of the postwar US MIL-S-12560 plate is ca. 15% higher than the perforation limit of a soviet wartime HHA plate, it´s actual holing limit would probably have been even further below.

This leaves a distinct possibility to consider postwar metallurgic advances in US RHA armor production under MIL-SS12560 standart to be significant.

[Peasant]

There have been some questions, I´d like to adress with the memo below.

Thanks for directing my attention to the shot #6. Actually I've made a mistake in translation. It doesnt say "cracked bulge" but rather "a bulge with a crack partially separating it from the plate" which looks like the start of a plug ejection, just like one would expect to see in a penetration of a blunt headed projectile.

That´s one posibility, and one might add, when viewing from the perspective of expected plate failure modes here, a probable one.
In plugging/discing, the projectile punches out a large conical plug (or wide disc) of armor from the plate back when it hits with enough force to hole the plate large enough to allow the projectile to have a chance of penetrating the plate. If the projectile has not sufficient power to exact a sufficiantly large hole, the projectile rebounds, broken or not (compare f.e. 152mm hit#9).
In similar photos, but with a back of the plate shown, where the projectile did not quite penetrate (no through crack formed like here), there are obvious wide bulges surrounding the impact point before cracks are formed, in some cases with a ring-shaped crack surrounding the impact that, if the striking velocity had been higher, would have caused that circular area at the plate back to tear free.
http://tankarchives.blogspot.com/2013/0 ... m-gun.html
The projectile acts like a hammer, punching out the steel in front of it´s projectile nose as a cylindrical or cone-shaped ring around the projectile nose. With the blunt nosed WWII-era soviet APBC projectiles, or inferior sharp tipped AP which suffer break up, the nose tip virtually never sticks out of the plate back prior to the plug being torn entirely free of the plate (as would be the case with more ductile events).
The point remains that there was no back bulge, which is initially associated with plug formation even before cracks are formed (compare hit#16) and this suggest to me that this impact was very much below any actual Holing Limit. It´s therefore not representative for the protection ballistic limit.
How much more energy is required to punch out a plug is not clear to me from this hit alone (see more below).

This is a concept that never occurred to be before. Shows just how much I still dont understand about terminal ballistics.

I see what you mean, but you cannot just that say that "the armour is good because it would've performed well in a controlled setting against the type of projectiles it was optimized to defeat". Most of the Allied WW2 AP shells would've broken their nose when hitting a 50°sloped plate, resulting in a blunt penetrator similar to soviet shell.

I think this, too, is a fair comment. The only ww2 projectile to have a chance of staying intact against this target (100mm @55°) is the 12.8cm Pzgr43.

How do you estimate these? Do you apply a formula or do you just eyeball it? I've used one from 1946 BIOS report, but it's is only for 30° impact and I didnt have much success in extrapolating its results to other obliquities.

----
That beeing said I have invested the better of roughly five hours of work to examine the nose plate hits in Kubinka. All eight hits with AP and APCBC-HE (only 88mm Pzgr39) were in direct line to the target, which helps eliminating a source of variance. Because the target plate is therefore identic (100mm at 55° to the perpendicular) and presuming that the failure mode is identic, too (discing with broken up penetrators- which is likely true for all attacking projectiles,88mm pzgr39 included)), it may therefore be justified to run a regression analysis (see attachment).

I have choosen De Marre K functions as a proxy of plate resistence because this is how ARTKOM assessed armor tests in ww2 (see Krogfuss article of DeMarre in Artkom for the details). While we wouldn´t do it today (De Marre K has a significant scaling issue) it sheds light on how they did it in ww2.
What can be observed is that there testers intended to obtain a lower and higher bracket with the De Marre K scaling with T/D ratio (smaller calibre need more impact force to obtain a holing than larger ones).

Not a quite on-topic but: Can you somehow modify the DeMarre to compensate for this? Maybe change the Thickness/Diameter variable exponent? Can you even have a formula that describes the plate/ballistic limit for a wide range of calibers(20-155mm), at least for ideal, sharp tipped , perfectly non-deforming projectiles?

hope it helps,

Yes, it's been an interesting read.

[Avalancheon]

If anyone is interested, I also have a thread on the armor quality of the Tiger I. I've made a few interesting discoverys that are worth some attention.

viewtopic.php?f=47&t=238721

[critical mass]

I see what you mean, but you cannot just that say that "the armour is good because it would've performed well in a controlled setting against the type of projectiles it was optimized to defeat". Most of the Allied WW2 AP shells would've broken their nose when hitting a 50°sloped plate, resulting in a blunt penetrator similar to soviet shell.

I think this, too, is a fair comment. The only ww2 projectile to have a chance of staying intact against this target (100mm @55°) is the 12.8cm Pzgr43.

How do you estimate these? Do you apply a formula or do you just eyeball it? I've used one from 1946 BIOS report, but it's is only for 30° impact and I didnt have much success in extrapolating its results to other obliquities.

The less effective vintage 15cm naval shells give a T/D ratio against homogenious armor for intact projectiles which in principle would make intact 12.8cm penetration at least feasable -though by no means ceertain (hence my use of the term "chance"). If You have the BIOS 1946 report ADM213/951 You probable have seen the Gercke formula. It contains an error (take the 1943 Lilienthalreport 166 formula´s instead). This can be used to apprximate limit velocity for projectiles in whole state fit to burst [G(h) in german terminology). You can cross calibrate the formula using the 12.8cm test results (30° and 45° obliquity) mentioned in p.113f. of the same document.
The test data indicate a delta between 12.8cm broken projectile penetration and a 12.8cm whole projectile penetration:

whole:
150mm @45° whole at 885m/s pzgr43
157mm@45° whole at 1050m/s pzgr Gg
135mm@45° whole at 840m/s pzgr Gg

broken:
201.7mm@45° broken at 900m/s pzgr Gg
135.3mm@45° broken at 840m/s pzgr Gg

difference between 26.3kg and 28.3kg pattern 43 APCBC-HE is a better heat treatment and the addition of a windscreen to the latter.
The 12.8cm was tested against 25cm plates (at vertical), 20cm plates (at 30°), 15cm plates (at 45°). I do not recall any specific 60° tests with this shell but judging from 7.5cm tests, one would have choosen 10cm plates had such trials been commenced.

----
That beeing said I have invested the better of roughly five hours of work to examine the nose plate hits in Kubinka. All eight hits with AP and APCBC-HE (only 88mm Pzgr39) were in direct line to the target, which helps eliminating a source of variance. Because the target plate is therefore identic (100mm at 55° to the perpendicular) and presuming that the failure mode is identic, too (discing with broken up penetrators- which is likely true for all attacking projectiles,88mm pzgr39 included)), it may therefore be justified to run a regression analysis (see attachment).

I have choosen De Marre K functions as a proxy of plate resistence because this is how ARTKOM assessed armor tests in ww2 (see Krogfuss article of DeMarre in Artkom for the details). While we wouldn´t do it today (De Marre K has a significant scaling issue) it sheds light on how they did it in ww2.
What can be observed is that there testers intended to obtain a lower and higher bracket with the De Marre K scaling with T/D ratio (smaller calibre need more impact force to obtain a holing than larger ones).

Not a quite on-topic but: Can you somehow modify the DeMarre to compensate for this? Maybe change the Thickness/Diameter variable exponent? Can you even have a formula that describes the plate/ballistic limit for a wide range of calibers(20-155mm), at least for ideal, sharp tipped , perfectly non-deforming projectiles?

That would be possible. But one might ask what is intended with this "correction"? One might need to guard against the possibility that it would just represent another fudge factor with limited validity. It´s usually not difficult to get a function out of a string of events but it´s an entirely different problem if You change the intention from descriptive statistics to prognosis.

[Peasant]

First, I sorry for taking so long to answer, Your post puzzled me for quite a while. Let me explain:
I wasn't talking about the Gercke penetration formula but the one used to determine the maximum plate thickness/striking velocity that a given projectile design can perforate in condition fit to burst. If we just take either Lilienthalreport 166 or Gercke formula(btw they both fit the known data pretty well, so why is one better than the other?) and extrapolate them beyond the muzzle velocities of the guns used historically, it appears as the penetration keeps growing with the striking velocity. But my experience with terminal ballistics suggest that after a certain point the projectile will deform and shatter, which would put a cap on the maximum intact penetration obtainable with a given projectile.
The formula (4) gives the limit conditions for such points. For 7.5cm PzGr.39 this would be 172mm/30° with the striking velocity of around 1000m/s where any further increase of velocity would only damage projectile even more and would not lead to increase in penetration (by German definition).
Yes the curves in the chart are extending much further that this, but notice the note that says "Impact Penetration obtainable" and not "intact" as one might think from a casual inspection.
So, what do you make of this?

[critical mass]

First, I sorry for taking so long to answer, Your post puzzled me for quite a while. Let me explain:
I wasn't talking about the Gercke penetration formula but the one used to determine the maximum plate thickness/striking velocity that a given projectile design can perforate in condition fit to burst. If we just take either Lilienthalreport 166 or Gercke formula(btw they both fit the known data pretty well, so why is one better than the other?) and extrapolate them beyond the muzzle velocities of the guns used historically, it appears as the penetration keeps growing with the striking velocity. But my experience with terminal ballistics suggest that after a certain point the projectile will deform and shatter, which would put a cap on the maximum intact penetration obtainable with a given projectile.
The formula (4) gives the limit conditions for such points. For 7.5cm PzGr.39 this would be 172mm/30° with the striking velocity of around 1000m/s where any further increase of velocity would only damage projectile even more and would not lead to increase in penetration (by German definition).
Yes the curves in the chart are extending much further that this, but notice the note that says "Impact Penetration obtainable" and not "intact" as one might think from a casual inspection.
So, what do you make of this?

Peasant,

thanks for the memo. I´d rate the translation as correct but caution. "Impact penetration obtainable with good quality shells" is indeed written literally but it makes no sense in the context of a penetration graph. "impact" does not even mean to cause a perforation event, it literally only means to "physically hit" the plate and therefore is superflous in a test report about armor penetration here (well penetration can only be measured in case the plate was impacted in the first place, agreed?), if Your litteral translation is to be believed. Why one needs a good projectile for merely "hitting a test plate" appears to be unexplainable. I´d argue that "Intact" makes more sense in this context as it represents the complementary unit to the 2nd and parallel transverse spline (see attachment) which sensitively is labelled "intact penetration with about 2/3 of mass produced shells".
There is more than that.

The lower line intact penetration with about 2/3 of mass produced shells" represents late ww2, official AP service acceptance specification (proof of projectile), in which a submitted AP lot had to pass with either

2of two selected (by WaPrüf) shells filled to weight and plugged blind penetrating intactly (no projectile damage compromsing fit to burst) or,

if one projectile of both fails in staying intactly -or penetrating (but not both, as this caused rejection), another projectile from the same heat had to repeat the test and this projectile HAD TO PASS intactly, or all projectiles from the heat were to be rejected.

Cautioning that projectile to projectile variances may still cause an individual projectile to be off (tests are destructive), one need to keep in mind that, this is the projectile quality floor (minimum acceptable quality, every single heat was tested).

This is where another observation comes into beeing. These two transverse splines carry arrows. And these arrows point to a field delineated by the next line. Notice that the arrows are not juts label pointers. In these sort of documents label pointers do not carry and arrow heads (compare, f.e. the label pointer lines designating projectile identifiers). Notice also that in the centre of the fields the contour symbol of the penetration curves for each cal. changes from solid to dotted, this is related to avg. break up at 30° (matches my data nicely).

The line thus tends to indicate that only good quality AP can stay intact at higher velocity than those given by the splines. How much was actually tested for a couple of projectiles from known good quality lots and this directly touches Your question:
Th 75mm Pzgr39 was tested against 200mm RHA plate at 30° and stayed intact, whether or not it penetrated at velocities to 1090m/s. Higher velocities are speculative in absence of data for this projectile.
The 88mm Pzgr39 (from a mass produced lot made by Bochumer Verein) was also tested at Hillersleben and penetrated a 305mm RHA plate at 30° and 1280m/s, the tests extended to 1400m/s but the projectiles according to the account of Dr. Ing. Wagenknecht (chief of projectile manufacturing at Bochumer Verein) the optimum penetration was obtained at between 1250m/s and 1300m/s and at higher velocities, the 88mm Pzgr39 regularely broke up resulting in lower penetration. These tests were experimental but the projectiles were taken from a mass production lot, run of the factory, proofed and of known quality (no failures).

From the best of my knowledge on this matter, and that includes primary source dealing with actual test firings from 1942 which eventually lead to the creation of the Lilienthal report 166 penetration graphs for mass produced shells, I´d argue that what You refer to for intact penetration limit is in fact the absolute projectile quality floor and not average or the upper limit. There is a fair amount of scatter in the quality of projectiles coming from different manufacturers

Intactness and penetration is correlated but not abssoluetly the same. A certain degree of projectile damage will not change the limit velocity for a penetration event (i.e. scars or deep gauges on one side) but once the amount of projectile damage (upset/offset and breakage) changes the shape of the projectile appreciably, then there will be an increasing delta between intact and broken penetrations. Back in the year 2005 I have posted the Lilienthalreport diagramm for this effect the first time on a different board which dealt with 75mm Pzgr rot and 75mm Pzgr39 (see attachment).

hope it helps

[Peasant]

Yes, You have my thanks.
One thing I've noticed is that the formula defining the "absolute projectile quality floor" is creating a parabola curve in space of "projectile caliber/lower limit of intact penetration". This implies that the bigger caliber shells are more susceptible to damage than the smaller ones. Is it because of the limitations of the 1940's metallurgy, making it difficult to manufacture good, large caliber shells or is it a consequence of the https://en.wikipedia.org/wiki/Square–cube_law ?

Regards, P.