Metallurgical Study Of Enemy Ordnance

Warship design and construction, terminology, navigation, hydrodynamics, stability, armor schemes, damage control, etc.
marty1
Member
Posts: 219
Joined: Tue Nov 02, 2004 6:31 am
Location: Seattle

Post by marty1 »

Hi Dave:

I think your commentary on the Wh & Ww alloy recipes not being driven by strategic alloy shortages is very important. I have often come across theories on various internet forums suggesting shortages were the cause of low nickel content in German RHA produced during the war. However if the Wh & Ww recipes were in existence well before the start of hostilities than the alloy shortage argument doesn’t really make sense.
User avatar
Dave Saxton
Supporter
Posts: 3148
Joined: Sat Nov 27, 2004 9:02 pm
Location: Rocky Mountains USA

Post by Dave Saxton »

The Wotan materials worked into Tirpitz had the amounts of Cr and Ni basically inverted compared to usual practice, with much greater Mo concentrations than expected. Here we find a Cr/Mo alloy design with reduced Ni and C content, in a pre-war application. Tirpitz was launched on April, 1, 1939. The amoured decks, and bulkheads would have been worked into it well before that.

The Germans were certianly feeling the pinch of Mo shortages by 1943. The German research record finds them scrambling to find suitable replacements for Mo. They found Vanadium to be the next best element to substitute the crucial role of Mo, in some of their high strength steels. Armour plates produced from 1942 to the end of the war, had more and more V replacing Mo. This is particularly true of the Tank amours. These Tank armours usually had Ni in only residual amounts too. Vanadium is a most excellant element to use as a grain refiner, and as a precipitation strengthing and precipitation hardening agent, in rather small amounts. However, it's effect in combonation with carbon is so profound, that it can easily lead to plates that are actually too hard. Moreover, V doesn't have the anti- temper brittleness effect that Mo does. Actually quite the opposite, particulary plates that get re-heated during cutting and welding. The late war plates, those left over, found, and tested by the allies at the end of the war, could have well been found to be too brittle, and too hard. Those plates should not be used to represent the typical norms of German homogenious naval armour, particularly pre-war productions.

One of the more interesting facts presented in the Enemy Metalurgy Review paper is that the Japanese continued to use nickel in copius amounts right up to the time the paper was published. They were not only using Nickel in Vickers recipe homogious armours, but in everything from aircraft bearing bushings, to cast iron cylinder heads. It seems kind of streange to learn this after it has been repeated so ofton, that the Axis used non-tarditional alloying only because of forced materials shortages.
marty1
Member
Posts: 219
Joined: Tue Nov 02, 2004 6:31 am
Location: Seattle

Post by marty1 »

Hi Dave:

Just continuing to add to this data base. The following represents metallurgical & ballistic testing data for a couple of German RHA plates – circa-1943. These are plates obtained from a German PzKw-III. I presume the tank was captured in N.Africa. Samples were sent to Aberdeen and Watertown Arsenal for Ballistic testing, and metallurgical testing.

Chemical Analysis
Plate A, Homogeneous Armor, 0.75-inch Thick, BHN = 375,
C = 0.52%
Mn = 0.70%
Si = 0.58%
S = 0.029%
P = 0.018%
Ni = trace
Cr = 1.39%
Mo = 0.20%
V = trace
Cu = 0.03%
Ti = 0.065%
Al = 0.03%

Plate E, Homogeneous Armor, 1.25-inch Thick, BHN = 331 to 341
C = 0.54%
Mn = 0.69%
Si = 0.49%
S = 0.027%
P = 0.016%
Ni = trace
Cr = 1.25%
Mo = 0.49%
V = trace
Cu = 0.09%
Ti = 0.07%
Al = 0.02%

Neither notched impact testing nor elongation testing appear to have been conducted on the samples.

Ballistic testing of both plates was conducted at Aberdeen. The results are summarized below:

PzKw-III Plate A vs. 20mm M75 AP
Obliquity--------------------BL(A)------------------BL(N)
0-deg------------------------1026-fps-----------------1155-fps
20-deg----------------------1679-fps------------------not conducted
30-deg----------------------1694-fps------------------2045-fps

Plate E vs. 37mm M51 APC
Obliquity--------------------BL(A)------------------BL(N)
0-deg-----------------------1340-fps-----------------Not Conducted

Similar Limit Velocity for US Armor of the same Thickness vs. Same Projectile Attack Are Approximately:

¾" US Armor RHA vs. 20mm M75 AP
Obliquity--------------------BL(A)------------------BL(N)
0-deg------------------------1314-fps-----------------1526-fps
20-deg----------------------1669-fps------------------1840-fps
30-deg----------------------1890-fps------------------2043-fps
Note: Typical Hardness of ¾” US RHA Plate ~BHN-320 to BHN-350

1-1/4" US Armor RHA vs. 37mm M51 APC
Obliquity--------------------BL(A)------------------BL(N)
0-deg-----------------------1373-fps-----------------1495-fps
Note: Typical Hardness of 1-1/4” US RHA Plate ~BHN-270 to BHN-300
User avatar
Dave Saxton
Supporter
Posts: 3148
Joined: Sat Nov 27, 2004 9:02 pm
Location: Rocky Mountains USA

Post by Dave Saxton »

These Carbon levels are very high, compared to typical naval homogenious armours. This illustrates how tank RHA reqiures slightly different properties than naval armours to deal with different threats. Carbon is of course the most important alloy used in steel for both hardness and strength. Carbon content also determines the reponse to heat treatments. Very low carbon steels don't respond to heat treatments. Were max hardness and strength are at a premium, cheap and easy to obtain carbon, gives the required results, and the properties can be more readily manipulated by heat treatments. Were the Cr levels may get reduced (pre war steels of similar type had uip to 5.5% Cr) an increase in C, can compensate to a degree too. In naval homogenious armour, excellant ductilty must be retained. Great toughness requires both great strength and ductilty.

Wh used much less C, and a more defined mix of alloying elements. By using greater concentrations of carefuly measured micro alloys, to obtain a specific fine grained micro structure, Wh was able to obtain greater strength, without unduly compromising ductility.

Welding requires a material with great toughness. Generally, steels with more than ~.30-.40% actual C, are considered unweldable, or at least extremely difficult to weld. Additionally, alloying agents (particularly Mn, and V) can intensify the effect of carbon. Usually the preheating reqiured for welding is determined by taking into account both the actual C content, and the micro alloys used. This is called the carbon eqivilency. One the common formulas used to determine C eqivilency is(all percentages by weight): C+ (Mn/06)+(Si/30)+(Ni/60)+(Cr/20)+(Mo/15)+(V/10). Weldable homogenious armours have a total C eqivilency, of no more than about .4-.6%. Incidently, Wh and St52 end up with virtually the same Ce. This is different from the simlar Schaffler nickel and Cr eqivilencies for micro structures of chromium and nickel alloys, but these can be of critical importance in determining dilution effects during welding of armour plates.

Since the 30's, more and more steels are produced from re-melting scrap steel. Scrap is also used more readily by electric arc furnace smelting. This is probably particulary true of wartime German tank armour. From the scrap, residual amounts of potential alloying elements (Ni, Cu, V, Cr...ect..) can find their way into the composition. However, Vanadium is so powerful, that what would be considered only trace amounts during WWII, can have significant effect.
marty1
Member
Posts: 219
Joined: Tue Nov 02, 2004 6:31 am
Location: Seattle

Post by marty1 »

In 1947, the British Conducted Firing Trails against Japanese 600-lbs Vickers Hardened Armour. 15-inch APC Mk XVIIB was used. Obliquity = 30-dgrees. The perforation limit velocity for the Japanese VH plate was about 160-fps better than the best British V.C. plates. Perforation limit of the Japanese 600-lbs VH plate was 1800-fps. The best British 600-lbs V.C. plate was 1639-fps (the worst British VC plate was only 1572-fps).

The Japanese 600-lbs Plate Chemical Composition is reported as:

C = 0.48
Si = 0.17
S = 0.036
P = 0.021
Mn = 0.36
Ni = 3.88
Cr = 1.86
Va = --
Mo = 0.04

Top End Lengthwise: Yield Point = 34 tsi, UTS = 48 tsi, Elongation 27.5%, Reduction of Area = 63.7%

Top End Crosswise: Yield Point = 35 tsi, UTS = 47 tsi, Elongation 27.5%, Reduction of Area = 61.5%

Top End Through Thickness: Yield Point = 32 tsi, UTS = 43 tsi, Elongation 5.5%, Reduction of Area = 10.0%

Face Hardness = BHN-575, depth of face = 27.3% of total thickness

No Izod or Charpy test data
User avatar
Dave Saxton
Supporter
Posts: 3148
Joined: Sat Nov 27, 2004 9:02 pm
Location: Rocky Mountains USA

Post by Dave Saxton »

The composition %'s closely match what the US analysists found of the Japanese VNC materials in 44, except the C% was actually a bit lower. Of course the C levels of a face hardened plate can be higher, because cemented armour is considered unweldable, and with a likely pearlite micro-stucture of the back portion, excess C can harmlessly be held in solid solution. With a carbide precipation hardening and strengthing mechanism being employed, C levels must be kept in check, if ductilty, and weldabilty are considered important priorities.

This is very interesting, as even with the higher C levels, the tensile strength is well below, that of Wh in particular, and also WWII era NCA with molybdenum.

I wonder how this compares with the modern CA plates worked into the KGV class and Vanguard?
marty1
Member
Posts: 219
Joined: Tue Nov 02, 2004 6:31 am
Location: Seattle

Post by marty1 »

Dahlgren also conducted post war ballistic trials on captured Japanese Armor. The report concludes that the Japanese Class-A Armor was inferior in ballistic quality to same gauge of US Class-A armor.

While the report encompasses a wide range of plate thicknesses, the following pertains to Japanese 13” Class-A Armor (Vickers Hardened). There were six 13” plates in the sample.


C = 0.52, Mn=0.43, P=0.029, S=0.031, Si=0.17, Ni=3.65, Cr=2.19, Mo=0.06, Cu=0.10

Average Tensile Strength = 103,110
Average % Elongation = 19.4
Average % Reduction in Area = 43.7

Average Plate Hardness from 0 to 1”, Rockwell C ~51
Average Plate Hardness from 1” to 2”, RC ~45
Average Plate Hardness from 2” to 3”, RC ~35
Average Plate Hardness from 3” to 5”, RC ~25
Average Plate Hardness from 5” to 13”, RC ~17

Charpy V-Notch @ 25-deg C = 78ft-lbs (Longitudinal)
Charpy V-Notch @ 25-deg C = 68ft-lbs (Transverse)

Ballistic Limit of the Japanese Plate vs. US 14” AP Mk16-8 @ 30-degrees is ~87% of Ord. Sk. 78841.

US Average BL for same plate thickness/obliquity/shell combination was 89.7% of Ord. Sk. 78841.
marty1
Member
Posts: 219
Joined: Tue Nov 02, 2004 6:31 am
Location: Seattle

Post by marty1 »

Running the numbers quickly for 14” APC Mk16-8 vs. 13” Class-A @ 30-degrees results in 100% of Ord. Sk. 78841 …F = 48327.

For the Japanese 13” VH Plate: 87% of F = 42044 so Limit Velocity is 1522-fps

For the US 13” Class-A Plate: 89.7% of F = 43349 so Limit Velocity is 1569-fps

Both are obviously Naval Ballistic Limits BL(N) .. 50% probability of complete penetration at the specified limit velocity.
George Elder
-
Posts: 168
Joined: Mon Oct 18, 2004 4:23 pm

Re: Birtish and US tests on Japanese armor...

Post by George Elder »

Hi Marti:

The results are so divergent that either the British got an outstanding set of samples or the US received very poor samples indeed. The results of the two series of tests are not easily reconsiled.

George
Bill Jurens
Moderator
Posts: 878
Joined: Mon Oct 18, 2004 4:21 am
Location: USA

Divergence in test results

Post by Bill Jurens »

To me, a rather large divergence in test results under these conditions would not be surprising. Testing and acceptace rules for manufacturers meant that most or all nations refined their projectiles to perform well against a certain type of plate of a certain thickness to be struck at a specific velocity. So the results under a narrowly defined set of test conditions were quite reliable and repeatable.

But, when the projectile is faced with a different type of plate that was itself designed to present maximum resistance to a different sort of projectile under different striking conditions, the results often are a bit erratic.

There is nothing highly unusual in this.

Bill Jurens
marty1
Member
Posts: 219
Joined: Tue Nov 02, 2004 6:31 am
Location: Seattle

Post by marty1 »

I gave the post-war USN and RN ballistic trials of Japanese VH armor a closer look. The USN trials are a bit more extensive in that several armor gauges and several projectile types were employed.

The British actually “recreated” a series of 480-lbs VH armor plate using the Japanese recipe for VH. These imitation plates, as well as the real-deal 600-lbs VH armor were subjected to trials at 30-deg by 15” MkXVII APC.

I’ve plotted the results of Both the USN and RN firing trial data as individual points – limit velocity as a function of e/d (e: plate thickness divided by d: projectile diameter). In addition, I have plotted as lines the USN Ord. Sk. 78841 Penetration Equation -- assuming 100% F – for each of the respective projectile types.

Of interest to note is how well the British trials match up with USN Ord. Sk. 78841 100%-F limit velocity line.

Image
Last edited by marty1 on Mon Apr 11, 2005 3:25 pm, edited 2 times in total.
Bill Jurens
Moderator
Posts: 878
Joined: Mon Oct 18, 2004 4:21 am
Location: USA

Armor tests.

Post by Bill Jurens »

Nice job, Marty!

Bill Jurens
George Elder
-
Posts: 168
Joined: Mon Oct 18, 2004 4:23 pm

A matter of predicting results.

Post by George Elder »

In terms of the divergent results, the main difficulty I can see is in predicting armor behavior. In this case, we have relatively small sample sizes and thus extrapolating the results to general notions is a risky business. Nonetheless, the general "rule" became that Japanese BB armor is not very good -- at least based on the US tests. The British tests make us take pause. In any event, based on the British tests, I cannot subscribe to the view that Japanese armor is worse than X, Y, or Z in terms of its general ability to resist Large caliber APC. As noted by Bill, differences in designing the plates to meet certain threats, variences in shell type performance, armor consistancy, etc., are all confounds, and thus it's probably best to avoid qualitative characterizations of Japanese heavy naval armor based on these tests. Of course, the Japanese had their own version of GKDos -- so this may help to standardize things a bit. But no time for that now... nor do I have that data.

George
marty1
Member
Posts: 219
Joined: Tue Nov 02, 2004 6:31 am
Location: Seattle

BL based upon only two shots

Post by marty1 »

I was looking at the USN trials of Japanese Armor yet again. It struck me that the limit velocities for the USN 14” APC and 16” APC are based upon very limited data. The attached figure represents the shot records for several of the USN and RN ballistic trials vs. Japanese VH armor. Complete penetrations and incomplete penetrations for each Projectile and Plate combinations are indicated.

One of the 14” Mk 16-8 trials is based upon only two shots. The second 14” Mk 16-8 trial is based upon only three shots. The 16” Mk 8-6 trials vs. 26” of Japanese VH armor is also only based upon two shots.

Ideally a V50% limit velocity is established by firing at least five or six shots at a given plate thickness and obliquity. More than six is better, but obviously these tests are very expensive to conduct.

Conversely the RN Trials involve a limit velocity based upon six shots. Note the complete penetration at about 1390-fps. This is subsequently followed by an incomplete penetration at a somewhat higher impact velocity…~1450-fps. The actual limit velocity is estimated to be ~1475-fps. This appears to be a pretty good estimate for this projectile and plate combination.

The USN “two shot trials” should really be taken as fairly loose estimations of limit velocity. One could easily see these limit velocities going 75-fps or 100-fps either way of the actual USN estimate.

Image
marty1
Member
Posts: 219
Joined: Tue Nov 02, 2004 6:31 am
Location: Seattle

For Example...

Post by marty1 »

For example: The British trials of 480-lbs VH @ 30-degrees vs. 15” APC resulted in a ballistic limit of ~1475-fps. The result of the 15” six shots and the resultant limit velocity is depicted in the upper most Figure.

Let’s suppose the British stopped the trials after completing of shots 1 & 2. The resultant limit velocity would than be estimated as ~1350-fps.

While 2 and 3 shot limit velocity estimates were not uncommon during WWII, post war testing specifications recommended firing five or six shots to establish limit velocity.

Image
Post Reply