rexford

Dispersion figures taken from official German ballistic tables for the listed guns, and the 75L48 data applies to APCBC fired at 750 m/s.

Stationary target at known range.

The T34 driver hatches that bulge upwards above the glacis throughout their entire area appear to be 75mm thick, while the earlier driver hatch with a rectangular border and slight indentation from the edges is probably on the order of 40mm or so.

"Whats apparent is that as the scope is scissored out, the eyepieces also move out! Just like when you use a pair of binocs for the first time, you have to angle the hinge element so that the eyepieces line up to your eye seperation. There is a scale on the tubular bar that shows the seperation setting also. Its 55, 60, 65, 70, 75 with ticks between those numbers. This way, you could fold them up and return to the same setting as before. The implication is that they probably could be used as rangefinders at the different angles. People with large heads and widely seperated eyes would have an increased stereo effect. They would be better then most at finding range." That scale is explained in the catalog article which shows the scissor scope in periscope and stereoscopic modes, and the scale has NOTHING to do with modifying the range dial readings for arm position. There is no implication that the scale refers to range finding, based on my reading of the text. The scale has to do with the eye distance changes related to the two modes, and that dial allowed the scissor scope to be used in two modes instead of one. It states this in the catalog article. Read the article, it explains why there are two dials (one is for interpupillory distance, or something like that). As I mentioned in an earlier post on this thread, a range dial calibrated for 90cm arm difference will not work when the scissor scope is folded upwards for several reasons: A. the distance between lenses is no longer 90cm and the range figure on the dial is keyed to a function of angle and lens separation distance B. the changes in lens angle no longer refer to horizontal changes, making any attempt to relate angle changes to range a nightmare of an exercise.

FIRST ROUND ACCURACY OF GERMAN GUNS The following hit percentages are taken from German ballistic tables (capped AP) and refer to the case where the range is known to a stationary 2m high by 2.5m wide target, and the random dispersion is doubled to model battlefield conditions: RANGE..50L60...75L48...75L70...88L56...88L71 500m........100.....100.....100......100.....100 1000m.... ...95......66......97.......93......85 1500m........68......33......72.......74......61 2000m................16......49.......50......43 2500m.................8......29.......31......34 3000m.................5......18.......19......25 Relative dispersion in the above table follows some unexpected patterns, with 75L70 being less scatter prone than 88L56 out to 1000m, but after that range 88L56 is slightly more accurate in terms of repeatability. With regard to 88L71, the scatter pattern is more diffuse than 75L70 and 88L56 at ranges out to 2000m but then attains a superior performance at 2500m which increases with range. German data for their use of the captured Russian 76.2mm L51.5 gun showed dispersion patterns similar to the 75L48 when APCBC was fired., with similar hit percentages with a known range. The 50L60 and 88L71 data applies to both the tank gun and towed weapons, while the 88L56 figures are limited to the tank gun. Review of data for the 88mm L56 Flak suggests that that weapon may have possessed greater dispersion than the tank gun. The marked inferiority of 75L48 scatter to the other guns is unexpected., since the 75L43 was used to knock-out T34’s at ranges above 1000m. It would be interesting to see if 75mm L46 Pak 40 dispersion had the same general values as the 75L48, which came along at a later date. It has been written that the introduction of the 50mm Pak in Africa greatly extended the range of direct fire combat for tank and anti-tank gun units, and the relatively close scatter pattern and excellent gun sight quality for that weapon would be superior to the 2 pdr anti-tank gun in both respects. Using the German figures for doubled random dispersion and assuming an average range estimate error of 25% with a bell shaped error distribution (typical results for average crew, based on British and American firing trials), the following first round hit percentages were computed against a stationary 2m high by 2.5m wide target: FIRST ROUND HIT % RANGE...50L60...75L48...75L70...88L56...88L71 500m...........81......75.....88.......79.....94 800m...........36......34.....51.......39.....61 1100m..........17......15.....28.......21.....34 1400m...........9.......7.....16.......12.....19 Muzzle velocities are 835 m/s for 50L60, 750 m/s for 75L48, 935 m/s for 75L70, 780 m/s for 88L56 and 1000 m/s for 88L71. All rounds APCBC except APC for 50L60 A 2m x 2.5m target size was used by the Germans as a reasonable model for the front view of a typical target tank, which simplified the calculations. Those dimensions simplify the complex variations in target width with height (T34 turret front is narrower than hull and has sloping sides, T34 hull width varies with height, etc.), and probably assume that ground rolls and folds blocked out some of the lower tank area. The above stated estimates for first round hit percentage probably represent the high side of what would be expected from average troops in battle, since “nervous and/or fatigue” origin errors were not considered during the calculations. Under the stress of combat, people can forget intermediate steps and see things on the gun sight that are not there. Discussions on the AFV News forum site have brought out the possibility that unquantifiable human errors may account for a good share of the reported misses at “sure thing” ranges. Regarding second shot corrections after misses, the Germans advised that bracketing should be used at ranges above 1200m using 200m increments below 2000m and 400m above that distance. At or below 1200m, "fire for effect" corrections to the initial shot would be made using various methods that would result in a more accurate change in shot placement than adding or subtracting 200m. An American gunnery manual for the Sherman indicates that bracketing is to be used at ranges over 1000 yards due to gun sight limitations which restrict the crew ability to make fine adjustments to the gun setting. [ August 14, 2004, 07:45 AM: Message edited by: rexford ]

Using the doubled dispersion data which is supposed to model the random scatter of rounds on the battlefield, the Tiger 88mm is acually a little better than the Panther 75mm at many ranges: HIT % When Range is Known Tiger 88mm APCBC 100% at 500m 93% at 1000m 74% at 1500m 50% at 2000m 31% at 2500m 19% at 3000m Panther 75mm APCBC 100% at 500m 97% at 1000m 72% at 1500m 49% at 2000m 29% at 2500m 18% at 3000m I have seen data for the 88mm Flak in the past which suggested that the scatter was much greater than the Tiger 88, based on lower hit percentages for the case where the range is known and the dispersion is doubled. Could someone post accuracy data for the 88mm Flak firing AP type ammo, with a reference? Thanks. P.S. made some changes to number typo's. [ August 14, 2004, 07:38 AM: Message edited by: rexford ]

The Germans noted that "fire for effect" would be used out to 1200m, and bracketing would be used beyond that range. Daniel speculated that the reason was near 100% first round accuracy below 1200m, but a reading of U.S. gunnery manuals suggests that gun sight limitations on second shot corrections were the cause. The American field manual for 37mm and 75mm gunnery indicates that out to 1000 yards the gun sight enabled the gunner the correct the aim onto the target using methods such as "burst on target" and other methods. With burst on target, the target is moved on the gun sight to the perceived location of the first shot, which should improve the second shot %. Beyond 1000 yards the gun sight did not allow the methods to be used very well, so bracketing was substituted. When the German manuals or instructions indicate that fire for effect is to be used to 1200m, and bracketing beyond that range, it appears that they are talking about the second shot correction methods. I believe that the Fibels do include a some gunnery correction methods with range limitations similar to the Americans (but with a longer range).

As noted in an earlier post, when the 75L70 range estimate is adjusted for one-half the perceived target height, the average round lands below the mid-point. The 75L48 lands close to the mid-point for the same range estimate and adjustment. So the mean trajectory of the 75L70 will scatter off the target before the 75L48 does as the range is decreased below the actual, since the 75L70 starts off closer to the bottom edge. And the 75L48 will have a higher hit % than the 75L70 when some of the estimates are less than actual. [ August 13, 2004, 01:52 PM: Message edited by: rexford ]

Jeff Duquette shared some pictures of a T34 M42 from Aberdeen with measurements of the glacis. Using his glacis measurement and scaling off a photo which showed the glacis and driver hatch thickness, the hatch was about 75mm thick. We have some other info that suggests that the T34 M42 and M43 glacis was 75mm thick, including German penetration drawings for 88L56 which do not allow penetrations of the driver hatch.

Using trajectory estimates and random scatter, the 75L48 and 75L70 APCBC rounds would have at least a 75% chance of hitting a 2.4m high target at 800m if the initial distance estimates were within the following ranges (triangles against bottom of target): 2.4m High Target at 800m 75L48: 680m to 830m 75L70: 695m to 975m The gun would actually set for a range 150m greater than the above initial estimates to obtain the final adjusted gun setting. Shots taken with initial distance estimates within the above ranges would all have at least a 75% chance of hitting the target. The above estimates assume that the crew does everything by the book and does not make any silly errors, which may have been much rarer than the other extreme. ================================ Earlier analysis of the Panther trajectory showed that adding "1/2 the perceived height in mils times 100m" resulted in a mean trajectory shot placement below the mid-point, whereas using the full viewed height placed the average shot above the mid-point. Using the lower aim point, 1/2 mils times 100m, makes sense since the target is usually wider below the mid-point and the number of shots up near the relatively narrow turret will be reduced.

Measurements of the T34 driver hatch on M42 and M43 versions resulted in a 75mm thickness, which would make that area impervious to just about anything the Germans could hit it with except the 128mm and 88L71 at close range. But hits around the edges of the hatch would suffer a resistance loss due to edge effects. Readings also indicate that some German Pak gunners were instructed to aim at the top of the hatch where the pins were, in the hope that the latches could be broken and the hatch would fall off. This suggests that effective hits against the turret front and mantlet were more difficult to obtain. A German instruction for 7.62cm Pak gunners against the vulnerable locations of enemy tanks estimates a 100m penetration range against the T34A (Model 41) glacis hull front at 30 degrees side angle, and 500m against the driver plate at 30 degrees offset. The greater penetration range against the driver hatch suggests about 11% less thickness than the glacis, or a 40mm thick driver hatch. Since the Germans did not appear to modify their calculations for cast armor, the above would suggest that early T34 had considerably less armor on the driver hatch. It is also worth noting that the German calculations did not take into account the high hardness nature of T34 armor, which would lower the resistance of a 45mm plate by about 26% when it was hit by 75mm or 76.2mm rounds. And the Germans did not consider the variable thickness of T34 armor, which varied from 42mm to 55mm during measurements of actual plates and could be responsible to the variable penetration ranges that are reported for the 75L43 and 75L46 guns.

If a 2.4m T34 is at 800m range and a StuG IIIG estimates the range as 800m, the following range would be set on the gun: 800m initial range estimate target appears to be 3 mils high so add 3 x 1/2 x 100m or 150m to the range estimate, for 950m. Place triangle on target bottom but raise gun for 950m distance. Trajectory of shot places round 1.28m above target bottom at 800m range, close to middle of target (Elvira's navel). Tiger spots 2.4m high T34 at 1000m and estimates range as 1000m. T34 appears to be 2.4 mils high, so adjustment to range estimate equals 2.4 x 1/2 x 100m or 120m. Set gun for 1120m shot with triangles at bottom of target. Trajectory places round 1.17m above target bottom at 1000m range, right around the target mid-point. Panther spots 2.4m high T34 at 600m, estimates range as 600m. T34 appears to be 4 mils high, so range adjustment is 4 x 1/2 x 100m or 200m. When gun is elevated for 800m shot at bottom of target round lands 0.8m above target bottom, a little low. If Panther crew range estimate of 600m was adjusted by perceived height of T34 times 100m, result would be 600m + 4 x 100m or 1000m. 75L70 aimed at 1000m range and bottom of target places round 1.6m above target bottom at 600m. I have read where the Germans aimed for the turret ring on KV and T34 tanks for disabling hits with light guns (even 37mm peashooters on 38t's could do it!), and 37mm Pak gunners were given pamphlets showing the vulnerable locations on the T34 front aspect which could be aimed at. So it appears that the triangles were not always placed on the target bottom with the hope that a hit would occur anywhere on the target.

In an earlier post I asked about an explanation of how a stereoscopic rangefinder worked, and no answer was forthcoming. The following site lays out the basics of coincidence and stereoscopic rangefinding, which suggest that the TF 14 would not find ranges in the periscope position: http://www.eugeneleeslover.com Go to the Gene's U.S. Navy link and then volume 2 of naval ordnance and gunnery. Chapter 16 is what you want to look at, radar and optics. Rangefinders work by measuring the angle needed to bring the view into a desired position, and the rotating range scale that one uses to adjust the view is tied into the distance between the lenses when they are horizontal. See the bottom of the paragraph to the immediate left of figure 16F8 for an explanation of stereoscopic rangefinder adjustment and what it is actually doing: "The operator adjusts the line of sight until the reticle image appears to lie at the same distance as the target image. The rangefinder has then measured the angle, and its scale indicates the range." So, if the range finder is near vertical the distance between lenses is not consistent with the turnable range scale, and the range finder will NOT correctly read or determine range. The determination of the range to the target is a trigonometric function of base length and angle, and if the base length is not the same as the horizontal arms condition the range scale is not going to work. Since the stereoscopic rangefinder works by measuring the horizontal angle needed to bring the view into compliance, if the arms are close to vertical they no longer are measuring horizontal angles when the range is adjusted. Another point to consider. Scaling off pictures of the TF 14 rangefinder/periscope, the distance between the center of the two lenses appears to be 24.8cm, which seems further than the distance one obtains by scaling off the picture of the SF 14 protruding above the JagdTiger superstructure top. With a base length of less than 25cm, the TF 14 would NOT provide the precise range measurements needed for a high percentage of first round hits at long range with the 128mm gun, or even the 88L71, even if it did work in periscope position. Now we have two sites that suggest that the TF 14 will not measure ranges in the near vertical position. Based on what I've seen, my view is that a near vertical TF 14 or SF 14 is not going to measure any ranges. Until someone brings up something to the contrary. p.s. I posted this just prior to seeing Mr. Tittles post with the same web site as noted above. We both came upon the same stuff at the same time. [ August 11, 2004, 01:44 PM: Message edited by: rexford ]

I read through the above web site and there is one interesting comparison of the SF 14 site at two different positions, with arms near horizontal and the other with arms close to vertical. Ready. When the arms are close to vertical it is referred to as the "Periscopic mode". When the arms are almost horizontal it is referred to as the "stereoscopic mode". This suggests to me that horizontal might be for range measurements via stereoscopic use, and vertical is limited to use as a periscope. ALL of the pictures that are captioned as "stereoscopic mode" show the arms close to horizontal, and none of the pics with the arms close to vertical indicate stereoscopic mode. The initial write-up for the web site piece states that the SF 14 could be adjusted either for periscopic or stereoscopic use. This seems to raise a few questions regarding the use of the SF 14 when the arms were raised up through the top of a vehicle or up and over the shield on a Nashorn.

Uh yeah, it was discussed there. maybe you should go back to that thread and catch up. </font>

Does anyone have a picture of a scissors scope that is good enough to estimate the distance between the objective lenses? Maybe by comparison with a soldiers head size or some other odd dimension that can be estimated reasonably well (I have a big fat head so won't use mine as a standard). ) I calculated three hit percentages with the T.F. 14 (90cm, 45cm and 30cm separation) to allow a curve to be drawn which would allow hit % estimates for lens separation inbetween the three points or slightly outside the outer points.

If a target is at exactly 1000m and the triangles are used, assuming 3m height or 3m width, the height comparison will result in an overestimate of the range (very few real combat targets will show up as 3m in height) and using the width will result in an underestimate of the range (some tanks exceed 3m width. If the target hull is angled by as little as 10 degrees to the firer, the effective frontal view width of a tank with a frontal width of 3m and side length of 6m increases to 4.0m, 33% wider than the actual frontal width. So a 10 degree target hull angle to a panzer results in a 25% underestimate of the effective range. If a T34 is expected to look 3 mils wide at 1000m on a frontal view, then a 4 mil wide view of the enemy tank front width suggests a range of 750m (1000m x 3 mils/4 mils). Target looks 25% closer than actual. The above explains why triangles do not result in a very high hit percentage at all ranges to 1200m.

We have a coincidence rangefinder at work which I keep in my overhead cabinet. Every once in a while I take it out and use it, and the results are not that bad. We also have laser rangefinder nocs which are easier to use, effective range out to 1200 yards. They are able to remove the irregular effects of rain drops and the like from the results, and tell you how good the quality of the target reflection is.

The Panther Fibel math seems to suggest that one half the perceived height times 100m is added to the initial range estimate to obtain the adjusted aim at the target mid-section. Tiger Fibel definitely shows the one-half the perceived height times 100m as the range estimate adjustment. Page 22 in the Panther Fibel, an 8 mil high target at 500m results in an adjusted range setting of 500m + 400m = 900m. Zielhohe is the desired height of the round, which is halfway up the target height. The Panther Fibel drawings on page 22 show the bottom of the T34 at 500m and top at 1300m, aim setting equals estimated range (500m) plus one half the perceived height (1/2 x 800m, for 400m). For 900m aim. [ August 09, 2004, 08:07 AM: Message edited by: rexford ]

We looked into the gun elevation issue and it does not matter. We ran ALOT of trajectory programs to explore the idea, since dug-in ATG would really benefit, and the trajectory is the same along the line of fire regardless of elevation. The Panther Fibel shows drawings where the trajectory is measured with regard to a line from the gun to the aim point, so gun height is not a consideration.

There are some statements that may not be on the mark. First round hits at ranges less than 1200m are not certain, especially with 75L48 and 50L60 using AP type ammo (APCBC or APC). Please explain the basis for the above conclusion. Secondly, the Panther Fibel indicates that a 200m addition/subtraction is used for the second bracketting shot at targets over 1200m. The Panther Fibel and Tiger Fibel clearly show that one estimates the range to the target, adds half the perceived height in mils time 100m to the initial range estimate, and uses that adjusted range with aim at target bottom for the FIRST shot.

While much has been said about the high accuracy range measurements possible with the T.F. 14 scissor scope, no real numbers have been put forth regarding the expected first round hit probabilities. So I sat down and cranked through the numbing numbers using trajectory analysis and German data for random scatter. The following stats apply to the 75L48 and 75L70 guns firing APCBC rounds at a 2m high by 2.5m wide target, and look at the expected first round hit percentages at 2000m and 3000m under the following conditions: 1. scissor scope distance between lenses is either 90cm, 45cm and 30cm 2. random ammo scatter equals one or two times the German test data 3. practical errors posted by Daniel are used as average values from bell-shaped normal distribution curve: 48m at 2000m and 108m at 3000m for 90cm objective distance 4. range finding function assumed to work with non-horizontal arms, which is possible but has not been proven beyond a doubt (at least to this Doubting Thomas who probably will be proved wrong in short orfer) Practical error at lens separation distances other than 90cm assumed to be proportional to 90cm divided by the lens distance. The first round hit probabilities are predicted to be: 90cm Lens Distance =================== 75L48 APCBC Single Scatter: 31% at 2000m, 6% at 3000m Double Scatter: 13% at 2000m, 3% at 3000m 75L70 APCBC Single Scatter: 58% at 2000m, 16% at 3000m Double Scatter: 40% at 2000m, 10% at 3000m 45cm Lens Distance =================== 75L48 APCBC Single Scatter: 19% at 2000m, 3% at 3000m Double Scatter: 9% at 2000m, 2% at 3000m 75L70 APCBC Single Scatter: 35% at 2000m, 8% at 3000m Double Scatter: 27% at 2000m, 5% at 3000m 30cm Lens Distance =================== 75L48 APCBC Single Scatter: 13% at 2000m, 2% at 3000m Double Scatter: 7% at 2000m, 1% at 3000m 75L70 APCBC Single Scatter: 24% at 2000m, 5% at 3000m Double Scatter: 21% at 2000m, 4% at 3000m Notes: German publications suggest that twice the test scatter be used in calculations to model battlefield conditions. It would be good at this point to find definitive information and a reference showing the ability of the scissor scope to measure range when the lens arms were not horizontal, which appears to be the case in most, if not all pictures depicting use of the T.F. 14 scissors. The above hit estimates assume that the crew works like in a highly trained and precise manner , where the commander correctly and clearly tells the gunner which range setting to use and the gunner accurately sets the range and carefully aims the finely aligned gun. Discussions on the AFV News site suggested that “errors” unrelated to range estimation or measurement might account for a good percentage of the misses, which would lower the theoretical hit probabilities presented earlier in this post. Note to Mr. Tittles: regarding the previous discussion on another thread regarding whether rounds bouncing off the pavement would detonate, the Sherman 75mm used ricochet fire to bounce HE off the ground and have the HE detonate above troops in trenches. The HE fuze had to be set to delay to avoid near instantaneous bursting. Since 88mm APCBC bursters will be set off by a 7mm plate, it isn’t that obvious that rounds bouncing off the street wouldn’t detonate.

With regard to the three man range estimation procedure, the Tiger Fibel mentions it and has quite a bit of write-up. Perhaps it was limited to Tiger crews under ideal circumstances? As was noted on the AFV News forum, when crews reach the front they sometimes experiment and find that what is in the manuals is not practicable, or there are better ways to do things. Having more than one set of eyes certainly makes for a better estimate. [ August 07, 2004, 09:18 AM: Message edited by: rexford ]

Daniel previously brought up the issue of how gunners could estimate an 800m or 1200m range for a tank target that was 3m high when it would look 3 mils high on the gunsight (and when compared to the 4 mil high triangle). Typical target heights follow, and few are exactly 3m high: IS-2, 2.7m hi x 3.1m wide KV-Is, 2.8m hi x 3.3m wide T34 M43, 2.7m x 3.0m T34/85, 2.7m x 3.0m M4A3 Sherman, 2.9m x 2.7m The above heights include any narrow protruding objects like periscopes, which might not be very visible at 1000m range with 2.5x magnification. None of the above potential panzer targets is 3m high, and if one limits the height to solid and wide areas the real observed height might be somewhat less. There is also the possibility of small ground curls and folds blocking out some of the tracks or lower front hull. Assuming a 3m target height comes with built-in errors, and would seem to result in high range estimates by the gunner. If one assumes a 3m width, and the vehicle hull is angled to the firer, the observed width will be greater than actual due to the angle and the estimated range will be low. Regarding the assumption that Panthers used the scissors scopes on more than a rare occasion, the 1945 report to Gen. Eisenhower does not seem to mention German tank use of the scissors unless I missed the discussion. It would seem that German use of such a highly accurate rangefinder in Panther tanks would have been a major point to bring out. A 3m x 3m target at 1000m with 2.5x magnification would appear to be 0.09" (0.23cm) to an observer (same size as a 0.09' object height held twelve inches, about 0.3m, in front of the eyes). Under the pressure of combat and in the rush to get in a killing hit, placing the gun sight cross hairs on the middle of an observed object that small could pose some problems for non-German guns during WW II. With regard to Mr. Tittles' comment about the German practice of adding to the range estimate (Daniel should revise his earlier post on this issue as his explanation is not consistent with the Tiger and Panther Fibels), the gun sight would be set at the final adjusted estimated range and the effective cross-hairs would be placed at the bottom of the target. If the estimated range was used as initially obtained or measured without adjustment, the hit percentage for a perfect range setting would be 50% since the mean trajectory would place the average shot on the bottom edge of the target. Half the shots would scatter off the mean point of impact. The Germans countered the above problem by setting the gunsight on the target bottom but added to the range estimation so the shot would land about halfway up the target height. The procedure was to add one half the perceived height of the target in mils times 100m to the initial range estimate (Daniel in his earlier post multiplied 100m by the full perceived height, which places the shot at the top of the target). When the target is at 500m, a 2m high target would appear to be 4 mils high, so half of 4 times 100m adds 200m to the initial range estimate. Final adjusted range estimate is 700m, which is what the Tiger Fibel shows (Elvira is 2m high in the Tiger Fibel example). Lorrin

Sir, I have written some papers on the relative size of targets and objects when they are various ranges based on theory and experiments. Also did studies on visual range estimation curve shape and probabilities, based in part on estimating range to cars on street using eyes alone (no comparison to what it should look like). Your figures in an earlier post appear to be incorrect, although the recent calculations look right. If 3m high target at 1000m is associated with 3 mils, the same height at 800m would look like 3.75 mils (you said 4.5 mils in an earlier post). At 1200m, the 3m height would appear to be 2.5 mils (you said 2.25 mils). The math for 3m height at 1000m and 800m (measured 12" in front of eyeball) would result in: 3m x 12"/800m = 36"/800, or 0.45" 3m x 12"/1000m = 36"/1000, or 0.36" A 25% difference in observed height (3 mils vs 3.75 mils is 25% difference), whereas you had 3m mils versus 4.5 mils (50% difference). You appear to have it right now but earlier post looks like something went haywire. Maybe you should go back and correct your earlier post on this thread. Lorrin