There was a major bug in the logic for the Japanese uncapped 15.5cm and 20.3cm Type 91 AP projectiles. It was introduced in Version 5.8. These projectiles have cap-heads (separate, flat-based noses that are held on by especially-weakened windscreen threads) to snap off on water impact and, at high obliquity, allow the now flat-nosed projectile to dive nose-first underwater in a stable manner so as to go deep and punch through the lower hull of the target if it hit the ocean just short of the ship, bypassing the side armor. It also would assist punching through deck armor at high obliquity. The very hard cap head would make the shell penetrate thick side armor more easily at low obliquity, since they made the nose more oval or, in the 20.3cm projectile, actually pointed. On impact with face-hardened armor at low obliquity, the cap head shatters, but keeps the nose under the cap head intact during the initial hole-forming portion of the penetration. However, when the projectile tries to penetrate after making the initial hole, the poor nose shape and rather weak body (these actually were semi-armor-piercing (SAP) shells with large cavities, even though called armor-piercing (AP)) shattered the projectile, requiring the striking velocity to increase to the shatter velocity to completely penetrate. Thus, the NBL is with shatter, but the HBL is without shatter, when the cap head impacts the plate first. This logic was messed up in 5.8 when the Gruson and Terni Cemented armor logic was modified and would never allow shatter for these projectiles. I fixed it in 6.0.
I removed the NOCAP 5% bonus to the projectile quality if the AP cap was knocked off prior to hitting the plate being tested. While this is theoretically correct to remove the loss of energy in destroying the cap that is no longer there, this effect is "lost in the noise" of all other forms of error and more-or-less built into the values used in FACEHARD; that is, tests of old uncapped projectiles that later added caps have this effect included in all of the test results, so there is no add-on effect if the caps are taken back off later. It was a mistake on my part.
I could not find any difference whatsoever in the low-obliquity test results of 1890's "AP Shot" (small explosive charge or none) and "AP Shell" (3-6% explosive charge) against either Harvey or KC armor, so I lumped them together in my projectile tables. It did not matter if soft AP caps were added, since the results tracked there, too. This changes the projectile tables of all nations. The final projectile quality value I ended up with was quite close to the previously-determined 1890's Krupp AP shot value against Krupp original KC armor (KC a/A), so I have a cross-check on my results.
I modified the computation of the minimum plate thickness that can shatter a projectile, restricting that to steel projectiles only (cast iron is too brittle).
I made a major find in information tests of Harveyized Nickel-steel armor (and on some Krupp KC armor plates) with uncapped and soft-capped projectiles (except for two 21-degree capped impacts, all were at normal (right angles)). The source was the US Naval Institute PROCEEDINGS 1897 article "On the Perforation of Face-Hardened Armor" by Ensign Cleland Davis, US Navy. I wish I had found out about it before!! I "analyzed it to death", plotting my FACEHARD 5.8 results against the test results, many of which allowed a rough NBL value to be estimated. After trying various methods, I finally discovered that using the unshattered "perfect" AP projectile in FACEHARD 5.8 and comparing its expected results for a Quality 1.00 KC armor to the Harveyized Nickel-Steel test results allowed me to create a graph of relative quality using a ratio of striking velocity divided by FACEHARD's computed NBL versus plate actual thickness (3-18" plates were used in the tests). The plate quality as measured as a ratio of velocities drops steeply in a linear manner from 3" to 8", then much more slowly, also linearly, until 18" is reached. Some tests were below the NBL and I put an UP ARROW on them, while others were above the NBL and they had a DOWN ARROW, with a few that were right on (or very close to) the NBL. The lines I drew passed between these "teeth" for all but 4 plates, one of which was experimental (do not know in what way) and two of which were tapered plates (thickness varied with vertical distance from bottom edge), one tapered plate being well below average and one well above average. Only one "regular" flat plate was extra-strong and had a too-high NBL compared to this line. The capped and uncapped results track one-another perfectly from 8" plates and up. For the 3-8" plate range, shatter causes the uncapped projectile resistance to increase faster with decreasing plate thickness than the rate for capped projectiles which do not shatter. The difference between shattered and unshattered penetration ability of AP projectiles against Harvey armor is much smaller than that with KC armor, since the damage to the nose due to shatter is smaller and the effects rely more on the collapse of the projectile nose and body by cracking and poor nose shape after the shatter is finished destroying the nose tip than the total destruction of the major portion of the projectile nose and, usually, upper body by shatter with deep-faced KC armor. Indeed, there were a few impacts where uncapped projectiles did not shatter, though they usually had some other form of damage due to the impact (see (6), below). Since most later forged-steel AP projectiles were stronger than the ones used here, they would have the same damage resistance against Harvey armor, too. I had to add a rather extensive logic process to handle all of this, which I tested extensively, though it is still possible that there are some bugs left.
The US armor (and later projectile) firm of Midvale made four uncapped projectiles in these tests and only one had a nose-tip broken, though all suffered major nose and upper body compression and flattening ("nose upset" damage), whether they completely penetrated or not. The uncapped shells of other manufacturers almost always broke up if they failed to penetrate and broke a frequently even if they did completely penetrate, indicating a form of shatter was occurring to them. These unique Midvale extra-tough projectiles show where the later "Midvale Unbreakable 1916" nearly-invulnerable WWI-era soft-capped APC projectiles came from. Looking at later US projectile designs, this improvement was accomplished by adding an outer hardened case to hold the tough part rigid; what is called "sheath hardening" like a coconut or Russian Dolls (one inside the other, hard on the outer, large ones and soft and tough in the smaller, inner ones) over the entire nose and upper body, with the internal tough part extending from the lower nose (the upper nose was hardened to a high level like other AP shells) and running down the middle of the shell, around the internal explosive cavity, to the base, and extending outward to the surface near the driving band to make the shell less liable to breakage there. This outer sheath was gradually thickened as the distance to the nose decreased until it ran all the way across the upper half of the nose, forming a evenly-hardened arch outside the softer, tough steel near the centerline starting in the lower nose. No matter in which direction the impact forces went or how far down the projectile side (above the driving band) they occurred, they were met with a hard outer case supported by a tough inner core. In WWII, this technique was perfected and created the best AP projectiles over the widest range of impact conditions of any in the world. I added special logic for these tough 1890's AP shells against Harvey armor and they can be selected in the US projectile table, uncapped or soft-capped.
I cleaned up the various displays, improved the description wording, fixed some minor display errors, simplified the projectile damage printout (it was very redundant), and added additional messages to explain where the results came from. A little more tweaking was needed in 6.2. I am trying to get the thing perfect.
Added RUSSIA to the possible nations, with the circa-M1907 & up APC shells using the British 13.5-15" GREENBOY post-Jutland APC test data, since tests by the British of a 12" version of the Russian ammo seemed to act exactly like those improved WWI British shells. I filled out the rest of the Russian table with estimated projectiles similar to other European navies. I have pictures of many of them and they seem to be of similar designs. The Russians used "tough" (high-nickel-content and hardened) AP caps similar to Skoda's AP caps from the test results, which match high-quality hard caps. A friend of mine supplied me with more data on the ID and explosive filler contents of several Russian projectiles, so I updated this list as to these descriptive things, though the selection options are otherwise the same.
I corrected the logic for THINCHILL armor (Harvey and US post-WWI Bethlehem Thin Chill Class A armor with a very thin face) and COMPOUND armor (steel-faced wrought iron armor) as to how much of a projectile penetration quality bonus they give to projectiles due to their weak damage-causing abilities. The COMPOUND armor now gives no bonus to cast iron projectiles, a 10% bonus to the weak 1890's steel AP/APC Shot and Shells, and a 20% bonus to everything else. THINCHILL armor gives no bonus to anything but the stronger steel AP/CP/CPC/SAP/SAPC/APC shells introduced after around 1900 or later in some nations, to which it gives a 10% bonus. I had not updated this properly when adding Harvey armor, which is a form of THINCHILL armor, and re-evaluating the old 1890's AP/APC Shot and Shells as I did. The program has a lot of tiny corners and crevasses that have to be fixed when you make a major upgrade, as I did with Harvey armor and those old shot and shells. Lots of loose ends. "The Devil is in the details." It works now.
The thin Harvey plates (under 8" thick) increase slowly in resistance as they get thinner and thinner (thinnest plate is 3" thick). Shatter resistance increase (against uncapped weaker projectiles) is faster than the non-shatter increase (soft-capped and stronger uncapped projectiles). To handle this divergence properly I had to actually increase the Shatter Thickness Multiplier (1.00 for all thick Harvey Armor plates). The method I had tried in FACEHARD 6.0-6.2 did not work properly for all tests in this lower thickness range, so I finally did a brute-force Shatter Thickness Multiplier linear change formula giving an increase in its value with decreasing plate thickness and this worked for all projectiles, no matter what size they were, and almost all of the thinner plates (not all, but getting all plates to agree on anything is rare, given the statistical problems with quality control of the plates and projectiles). Lots of minor loose ends to clean up when you make such a radical change to a computer program!
I made the error of not updating ALL of the possible projectiles to allow correct Harvey Armor computations. Those that used the DEFAULT projectiles would work, but not most of the rest. I thus had to go over every single projectile in all of the various Nation tables and correct them to handle Harvey Armor correctly. In the process I rebuilt the Italy projectile table, since it mostly tracked the British projectile table and I had extensively changed that. Now Italy is correct, too.
There were a few errors in the handling of Common projectiles:
- All of them have to have the LTCASE value set to 1 or 2 unless they have a cavity size under 4% (the only Common shell in FACEHARD with LTCASE = 0 is the US WWII 6" Mark 27 "Special" Common, which is actually a US Army-type small-cavity APBC projectile with a Hood to hold the windscreen on).
- Another error was that these projectiles ALWAYS have SHATRES = 1 unless they have a full AP cap, no matter what size their cavity is, including the 6" Mark 27.
- Finally, Hoods have some effect on protecting the nose of projectiles when KC and Grüson Chilled Cast Iron armors are hit, but Compound Armor and Harvey Armor do not shatter these modern steel projectile noses in the first place, so this nose breakage logic bypasses them. The existence of a Hood has no effect on the rest of the damage logic.
There were a two more errors in the handling of Common projectiles: (A) I reset the LTCASE = 1 minimum to over 4.5% - US Navy 8" Mark 17 Special Common (with Hood) has effective filler after penetrating older KC-type armor (Krupp Pocket-Battleship-type KC n/A) at normal and it had a filler of 4.4%. Depth into nose region is the real measure (LTCASE > 0 must go far into nose). (B) More problems with nose-only shatter were fixed (complicated problem), with hoods working against the less-tough KC-type armor (SOFTSHAT = 0) just like soft caps work against the high toughness KC-type armors (SOFTSHAT > 0). Made sure this and Curved-Plate Rule did not mess each other up (Curved-Plate Rule is for impacts over 45 degrees obliquity against barbettes and similar plates). Also, fixed logic problem with British Deforming Projectiles (AP/CPBC/SPAC) concerning maximum HBL. It is never supposed to exceed the shatter HBL, but it did, so I had to fix it (I hope it works; I did a lot of tests, but I obviously got the logic wrong the last time; this is a real "non-trivial" logic process).
I have three areas to compute projectile breakage (nose and/or body): Under the HBL, between the HBL and the NBL, and over the NBL. There are also several ways of computing this damage, depending on the weakness of the projectile. One of the methods is to determine if the projectile breaks when it completely penetrates and if it does when it barely penetrates at the NBL and stops if the striking velocity is raised high enough above the NBL. This last condition is due to the projectile hitting at such a rapid rate that damage that happened when it was only moving slowly after penetrating stops happening when the impact is over so fast that the projectile does not have time to react (this is never true with shatter, but is true with most other kinds of projectile damage). This evaluation requires logic that speeds up the passage through the plate as the striking velocity increases. If the projectile barely penetrates at the NBL or hits at below the NBL, the time-to-penetrate is essentially infinite by this measure, so this logic CANNOT be used. However, I had an error in my logic where I used "greater than or equal to" the NBL for this evaluation, messing up the result at the NBL itself and with other logic that used this test condition. I really wanted "only greater than" the NBL. I therefore corrected the program at this point by replacing a ">=" with a ">". Now this is a TINY change, but the error had BIG effects at times. I am wringing out the last few bugs, but it is getting more and more difficult to do it as they become more and more specialized as to when they occur! Much like getting the last little bit of catsup out of the bottle...
Another error related to steel Hooded or Soft-Capped Common Projectiles. The Body Breakage Flag BRK was being set to 4 (not left at zero (No-Op)) even when Compound or Harveyized Armors were hit, which was not supposed to be true for this particular value of BRK. I changed an "OR" to an "AND" to fix the problem. Another tiny error.
Added a message as to when you need to know the individual weights of the nose coverings, rather than their total, when the program is doing its body weight logic. This only occurs if just the windscreen was knocked off by a previous impact, since then the Hood or AP Cap is still there, so that the weight of the projectile is less, but still more than the body weight. Also, the Japanese shells with Cap Heads, lose those, too, so this can affect penetration more than just a simple weight loss.
Fixed a few printout typos (minimum Striking Velocity is 1, not 0.01).
Added Hard caps to French 1909 and after SAPC projectiles (4% filler) which the French used instead of APC until the 38cm WWII APC was introduced. These caps only hardened the very tip of the nose (only the conical face portion of the cap was hardened down to the corner edge where the cap side skirt began), with the major portion of the cap volume between the face and the projectile nose, down to the bottom edge of the cap's skirt, remaining of the Tough variety - hard caps are always of a nickel or chromium-nickel armor-steel type, though sometimes with unusual levels of nickel or other alloy, since mild steel cannot be hardened properly to the high levels used in the face layers of hard AP caps (circa 500 Brinell and up, with a few going up to 680 Brinell. These thin-hard-face capped SAPC projectiles, with new windscreens and perhaps new fillers and fuzes, remained in use in French WWI-era battleships through the end of WWII. I use as my template for these French thin-hard-faced SAPC shells the WWI Skoda capped APC/SAPC with its tough cap (listed as hard in my Version 6.8, but with a reduced-function cap), since I do not have test results for these French shells.
Added "Tough" caps to the WWI-era selection. These seem to be high-nickel-content steels much like homogeneous, ductile nickel-steel armor, but sometimes with up to 25% nickel content (Austro-Hungarian Skoda post-1907 APC shows this in blueprints). They act something in the middle between hard-faced caps with thick hard layers under the cap face and soft mild-steel caps, being of higher intrinsic hardness and strength than soft caps, but still kept ductile. Certain Skoda, Krupp, and Russian projectiles have these caps, though part of the Krupp Tough cap results may be due to the extra-high-strength high-temperature solder used to hold the caps on (the only company to use this solder on anything but hard caps). I use a reduced-effectiveness hard cap from my previous versions of FACEHARD for all of these tough caps (AP Cap Selection = 3 in my program logic, with -1 meaning a Hood, 0 meaning no Hood or cap, 1 meaning a soft mild steel cap, and 2 meaning a fully-functional hard cap (post-WWI for the most part)). In most cases, this changes nothing, as I had already assigned them AP Cap Selection 3 caps, reducing the cap effectiveness of these projectiles, but a couple of shells are now listed with tough caps which had hard caps before. I am thus lumping into this generic "tough" category the last 1911 WWI-era Krupp APC shells, 1908 and after WWI-era Skoda APC/SAPC shells, Russian 1907 and after WWI APC/CPC shells, French WWI-era SAPC shells from 1909 with hard caps with very thin hard face layers (see 17, above), and Krupp c.1934 WWII-era 38cm SAPC with thin AP caps (for use against light cruiser-thickness face-hardened armor). I do not have enough data to sort them out at this time.
Increased the number of lines on the screen to 43 from 25. Allows more room for printouts, if any new ones are needed in the future.
This article is copyrighted 2012 by Nathan Okun and is reproduced on NavWeaps.com with permission.
- 5 March 2012