Elongation of Homogeneous Armor

[Thorsten Wahl]

regarding alleged poor plate

according Unterlagen zur Bestimmung Hauptkampfentfernung und Geschoßwahl Heft a Textband
Excerpt

The Upper Limit

It should be noted about the safety of horizontal protection against penetration, that results in the pentration-expectations at the relatively small impact angles can only be an indication. Low material differences frequently have a great impact, so for example, differencies in the cap shape(cap removal) and hardness (of the plate Thoddy) can affect pentration abilities in such a way, that in one case, the projectile enters the plate and in the other case, under otherwise the same conditions, ie at the same impact velocity, but with different cap, the projectile will be rejected. Also, the values, both of the angle as well as the impact speeds at which a projectile is dismissed or recently pentrate the plate, often influenced each other so closely, that they can not be separated exactly. Further factor is, that in cases of relatively low impact angles, the elusive nature of distraction has even more impact, than at larger impact angles. This effect can be found especially when several plates have to be pentrated. It is also possible that the projectile enters the plate with an angle to the direction of flight, so it has to pentrate with a much larger cross section.

As you can see it is explicitly mentioned, that a plate with a different (higher) hardness for the armored deck was choosen, wich provide in accordance with the the revised impact conditions (decapping, distraction, yaw) better performance compared to "normal" plates (with lower strength and greater elongation).

I hope the translation is ok

[Tiornu]

Dunkerque had two belt penetrations and no deck penetrations.
Nothing I know of in Jean Bart's experience indicates substandard performance of the deck armor.

[Thorsten Wahl]

The thing with "bad" French armor is mentioned in Bericht 166 of the Lilienthalgesellschaft armor Meeting Berlin 1943 "armor materials and armor pentration process"

but there was only a qualitative statement from comparative AP shots against french armor but no in dept analysis.

[RobertsonN]

According to the very detailed account in Dulin and Garzke there were no belt penetrations of Dunkerque by the Hood. One hit was, as I described, of the deck, and another passed just under the belt and exploded against the TB (very similar to the underwater hit by POW on Bismarck). As a result of the latter, the TB and cable passageway inboard were locally destroyed. Splinters damaged one boiler and severed the main steam pipe. There were numerous casualties. A wave trough was unlikely, but the ship may have had relatively little fuel in her bunkers reducing the depth of underwater protected hull.
As for Jean Bart, the data by Worth/Lundgren on NavWeapons give the penetration of deck armor of the US 16/45 as 4.5 in (British) to 5.4 in (Italian) at 24000 yds. Penetration does not reach 6 in of British armor until over 28000 yds. No data are given for French armor. The thickness of the Jean Bart's deck was 5.91 in. While the penetrations using this program are often doubtful (optimistic), they are probably on stronger ground in the case of US guns against one layer systems (the inner layer in the French layout being splinter protection only).
The well known photographs of the massive damage to the bow and stern of JB show that it was fortunate the ship was in harbor when it was hit.

[delcyros]

I think what people are referring to is a turret deck hit on DUNKERQUE 1940. As far as I remember, this turret roof was made from KC and not homogenious armour. High obliquity impacts with the blunt AP-cap in place would be less likely to deflect off against face hardened armour but I don´t remember all the details. Still a remarkable performance from a 15in APC to defeat the turret roof at rather close range. A homogenious plate would be more likely to deflect off the projectile.

Back to Wh: Thanks Dave, RobertsonN and Thorsten for you´r replies, again my question refers to thin and thick Wh plates. In the norwegian and british mechanical properties trials conducted on samples taken from TIRPITZ, it was found out that the upper, weatherdeck plates were softer than the lower, main armour deck plates. The treatment was different with regards to hardening treatment, too. This is in general agreement with Krupp´s findings that in a layered scheme a decapped AP is going to penetrate harder plates at higher difficulty.
I understand that this is also mirrored in british trials at least but US trials seem to come to very different conclusions. Do we have a source detailing the limit velocities?

[Tiornu]

The G&D account of Dunkerque's hits are incorrect in the details. Dunkerque suffered, I believe, four hits: one to the turret, two penetrating belt hits, and one that struck a largely unarmored section. Neither belt hit was below the waterline, but one was low enough to go through the deck slope before clanking into a boiler.
It is regrettably easy to misunderstand our penetration tables. Regarding the formula for deck penetration, "It is most accurate in predictions for British and German shells, while Japanese and American shells suffer slightly." So the performance against JB's 150mm deck would be greater in most cases than the figures indicate. It's also important to note that there is some disagreement on the details of JB's hits in US and French records, which I find confusing since the US assessment was based on info from the French. The grotesquely visible damage to JB was caused by bombs.
I know of two writings in the works that deal with Casablanca. I don't know if they will focus on details of interest to us here, but it would be nice.

[Dave Saxton]

Back to Wh: .........I understand that this is also mirrored in british trials at least but US trials seem to come to very different conclusions. Do we have a source detailing the limit velocities?

The only tests of un-capped shells vs homogenous plates that I know of conducted by the USN was in 1925. At that time, they found that uncapped shells did slightly better against standard hardness homogenous armour during oblique impact than did capped shells of equal mass. They did not establish the effects of higher tensile strength homogenous plate on uncapped shell. Generally the USN BoO only studied single plate penetration concepts, because they lacked the specialized equipment needed to record and quantify the data. So the effects on de-capped shell did not receive a lot of attention. We do know that they considered the upper armoured deck on the new fast battleship designs to cause yaw effects, and that they asked Carnegie Steel to study these concepts further during the war.

The US examination of German homogenous plate was of a 17" thick experimental plate recovered from Meppen. It can't be considered representive of Wh plates worked into warships by any means.

[delcyros]

It is regrettably easy to misunderstand our penetration tables. Regarding the formula for deck penetration, "It is most accurate in predictions for British and German shells, while Japanese and American shells suffer slightly."

I understand with the data presented by Thorsten here

http://warships1discussionboards.yuku.c ... d-M79APCLC

do indicate some precision of M79APCLC and facehard v6.8 at obliquities ranging from 0 to 20-25 deg. At higher obliquities, the data returned by these programs heavily underestimate the required striking velocity to achieve the NBL.
At 45 deg, f.e. M79APCLC requires only about half the striking velocity to defeat said thickness armour 320mm Wh than primary sources like the penetration graphs from the GKdos 100.
I have made the same observation with 11.13in 330kg APC against homogenious armour at 45 deg. The Unterlagen zur Bestimmung der Hauptkampfentfernung show 160mm homogenious armour to be defeated at 45 deg requiring about 500m/s striking velocity (~1660 fps) from graphs based on actual firing trials, while M79APCLC was fine with only 1060fps using the Q=1.0 and E=18% suggested from Nathan Okuns metalproperties (revised 2010). That makes for a difference of 600 fps or more than 50%! How can You call that most accurate?

Thanks in advance,
delc

[RobertsonN]

A very interesting graph. There seems to be a similar discrepancy with official British figures. Worth/Lundgren give 8.7 in British deck penetration by the British 16 in 2375 lb shell at 34000 yds. The official British figure was just 6.0 in

[delcyros]

The below graph I did using penetration graphs from the Unterlagen zur Bestimmung der Hauptkampfentfernung. It shows in thick lines curves against KCn.A. and in thinner ones such against Wh. It is referring to the old patterned, 283mm L3,7 -300kg APC (pocket battleships only).



Note the penetration through Wh homogenious grade, thin armour at 45 degrees:
20mm requires a striking velocity of 73 m/s
40mm requires a striking velocity of 127 m/s
60mm requires a striking velocity of 196 m/s
80mm requires a striking velocity of 260 m/s
100mm requires a striking velocity of 328 m/s
(further data can be found in the plot of GKdos 100, my drawing above shifts to KC instead)
120mm requires a striking velocity of 399 m/s
140mm requires a striking velocity of 468 m/s
160mm requires a striking velocity of 548 m/s

Using M79APCLC gives somehow different results, with the error getting bigger with the thickness of the plate used:

20mm (=0.79") require 67m/s (=223 fps) -reasonably close
40mm (=1.58") require 127 m/s (=413 fps) -very good match
60mm (=2.36") require 187 m/s (=619 fps) -reasonably close, underestimating by 4.8%
80mm (=3.15") require 246 m/s (=812 fps) -underestimating by 6.5%
100mm (=3.94") require 298 m/s (=986 fps) -underestimating by 10.0%
(further data for homogenious armour can be found in the original source which I am referring to now, my drawing above shifts to KC instead)
120mm (=4.72") require 343 m/s (=1136 fps) -underestimating by 16.3%
140mm (=5.51") require 383 m/s (=1263 fps) -underestimating by 22.2%
160mm (=6.29") require 415 m/s (=1371 fps) -underestimating by 32%

Only thin plates seem to be well matched on M79APCLC. Any ideas why?

[RobertsonN]

Thanks for making these data available in a very clear way to non-experts like myself. The penetration figures look realistic for the 11 in gun. While big discrepancies have been obvious for decades (they are apparent in the Dulin & Garzke books but these authors make no comment on this), the work that you and Thorsten have done in producing such comparison curves makes the matter much clearer where there is rough agreement and where big disagreement.

[Byron Angel]

..... One point I feel compelled to raise here: Nathan Okun is a physicist by education and occupation who has been studying this subject for close to forty years. I know for a fact that Nathan is intimately familiar with GKdos100. I, for one, would like to hear Nathan's side of this issue before assuming that his numbers are necessarily wrong. It has often emerged that official range tables and documents like GKdos100 incorporated extrapolated, estimated, or computed data (as opposed to validated empirical data).

My opinion.

B

[RobertsonN]

I too am a physicist by training, although one with no special expertise in this field. I also have some flexibility in my ideas: for example, I've become distinctly more favourable about the Richelieu this week, partly as a result of the arguments of others here and looking at what Dulin and Garzke has to say in fine detail.
However, if I am not mistaken, M79APCALC was described recently on NavWeapons as being developed on the basis of actual test results for a 15 lb 3 in uncapped AP shell. A 15 in shell weighed over a hundred times as much, so the extrapolation involved is enormous. And physicists generally are sceptical about big extrapolations. When historic test results are apparently available for major calibre shells why should one use a model that is based on results for tank sized shells?

[Bgile]

I too am a physicist by training, although one with no special expertise in this field. I also have some flexibility in my ideas: for example, I've become distinctly more favourable about the Richelieu this week, partly as a result of the arguments of others here and looking at what Dulin and Garzke has to say in fine detail.
However, if I am not mistaken, M79APCALC was described recently on NavWeapons as being developed on the basis of actual test results for a 15 lb 3 in uncapped AP shell. A 15 in shell weighed over a hundred times as much, so the extrapolation involved is enormous. And physicists generally are sceptical about big extrapolations. When historic test results are apparently available for major calibre shells why should one use a model that is based on results for tank sized shells?

You are completely underestimating Mr. Okun. I suggest reading some of his articles.

[delcyros]

I on my part do appreciate and admire the work Mr. Okun is doing on the complex naval guns/armour interaction dynamics.
But let me express my curiosity about these discrepiancies between primary sources and M79APCLC results.
I understand that he is trying to expand upon M79APCLC to include different nose shapes / cap attachements and more but don´t know how
much progress is made in this direction.

It could be observed that thickness of the plate (not so much plate/thickness ratio as expected) and a certain obliquity range (25-60 deg) do relate significantly to these observed differences.
Putting the data´s presented in a multiple regression analysis, I got a very high significance niveau of 0.03 (little is high here), implying that the differences are by 97.0% not explainable by a random function.