Richard Anderson wrote: ↑28 Feb 2019 16:49
Avalancheon wrote: ↑27 Feb 2019 15:17
But as for the questions about the protection (or lack thereof) offered by the Shermans armor, that is kindof relevant. Even a 'mere' medium tank should be able to offer some protection from enemy anti-tank guns. On paper, the M4A3
should be able to shrug off hits from a 75mm L48 gun from medium range. But we've found no evidence of them doing this in combat.
Sorry, somehow I missed this. There may actually be a very simple explanation. The tanks in question are actually the M4A3 75mm and 76mm (w) and the M4 and M4A3 105mm with 47 degree hull front of the "Ultimate Medium Tank M4 series" (the M4A2 with 47 degree hull front in the series were mostly shipped to the USSR). The thing is, they did not start appearing in the ETOUSA until mid-August and late September 1944 and in relatively small numbers, about 509 M4A3 75mm, of which 250 were actually M4A3E2...so not really the type in question. Another 589-odd M4A1 and M4A3 76mm (w) were shipped June-September 1944 and were issued to units July-October, but the M4A1 again did not have the 47 degree hull front - it was a cast hull design. Of course, further shipments arrived November-December (but always in quantities insufficient to build up a strong replacement tank reserve - a major bone of contention between the ETOUSA and the War Department).
Anyway, what that means is that distribution of the new tanks in the late summer and fall of 1944 was very diffuse, spread between numerous units. Except for one case, the 9th AD. For some reason, which I have never been able to track down, it was issued a complete set of new M4A1 and M4A3 76mm (w), just in time for one of its battalions to go into action with CCR against 2. Panzer and Lehr in the Ardennes, where the Panthers had little problem defeating the new hull front. Otherwise, most of the improved tanks simply don't appear in significant numbers, especially in the photographic record, until very late in 1944 and then in 1945.
Hmmm,
Avalancheon is actually asking a good question here. Based upon the official penetration curves against RHA at 45° obliquity for the 75mm Pzgr. 39 APCBC-HE as of 1943 mass production as reproduced in Lilienthalbericht 166 (1943), the 2.5inches thick RHA and 47° sloped, M4A3 glacis represents a target, which is somewhat beyond the reliable penetration realm of the 75mm KWK40 even from close range without considering an additional target angle. F.e. at 750m/s terminal velocity (that´s point blanc <50m, for the KWK40), the 45° penetration really is only in order of 63-65mm, and should be less than that at the refence 47° slope angle. Probability to penetrate would be nil, when factoring in even slight target angles. And it cannot be stated that all M4A3 glacis penetrations are caused by KWK42 or 88mm...
In response to this, I may offer an explenation. The report 166 graphs are for mod.1943 Pzgr39 production, and that was before the projectile proof obliquity was raised from 30° to 45°. For what´s worth, the Pzgr39 graph at 45° in this report is for broken up projectiles, though they do anticipate 45° intact penetration as a future requirement in the report (hence, my default position is that the graph is NOT for intact projectiles at this obliquity and date). Further, the graph shows an already smoothed interpretation line through non-existent-anymore datapoints, which originally would have exposed some noticable zags every time the hardness of the target plate abruptly changes.
The 60mm german target RHA plate at 45° is noted in the graphs at 110 kg/mm^2 ± 5 kg/mm^2 tensile strength, that translates to a medium hardness of BHN 330 ± 15. Compared to the relatively soft 240-270BHN US RHA hardness practice, this consititutes for a significant difference in target properties. While one might argue, whether or not the BHN offers more resistence beeing so hard against approx. 0.8 T/D ratio impacts, the key difference here is that the harder material will damage a projectile more completely than softer target material would, particularly at these moderately high obliquities, which homogenious armor works very effectively by base slapping and damaging the projectile as it rotates first away and then into the plate as it digs in deep enough that the back side of the plate starts to experience failure by either plastic deformation or plugging.
Photographs of tests at 45° show the onset of deformation originating in the base of the pzgr39, generally.
Now, lack of damage to the penetrating projectile, caused by softer target materials could very abruptly alter the involved penetration mechanics. If the projectile stays intact, it can penetrate by plastic deformation, or plugging with intact penetrators, both are lower energy penetrations than plugging/ discing with a broken penetrator would be.
The only relevant evidence in this question for US RHA are the 45° post war USPAG data, which give a complete, intact penetration of 75mm Pzgr39 through 5 3/16 (131.8mm) US RHA plate. Apparently, these trials confirm the aforementioned presumption in regard to projectile damage. Under this condition, the 75mm Pzgr39 generally stayed intact, whether or not it penetrated (3155fps and up) or was rejected by plate (3099fps and down). This corresponds to a De Marre K = 2565 for 3155fps / plate thickness and projectile combination. If the thickness is allowed to be reduced to 65mm, this K translates -everything else beeing equal- to a velocity = 586.4m/s, or, approx. 160m/s less than what report 166 graphs suggested. I stress that the use of discrete data here just demonstrates a variance between graphs for broken penetration and intact penetration, and by no intents, do I want people to misunderstand them as useful proxy for downrange penetration probabilities (At 1.0km the velocity is 637m/s and at 1.5km the velocity drops to 585m/s, but these values cannot be used as a relevant reference here because we don´t know the likelyhood of penetration). De Marre K is a rough tool, and variances are to be realistically reckoned with.
Even though one has to be careful from jumping to conclusions from limited samples, this seems to somewhat contribute to resolve the conflict between legend penetration and observed evidence.