B-17G with a 20 mm cannon onboard!

[Tony Williams]

Hi Tony,

>German sources obviously differ about the percentage of shots fired which hit the targets: I have seen two figures quoted, of 2% and 5%.

Well, obviously the above figures are quite accurate and specific to the respective weapon and range, while "2%" or "5%" look much like arbitary assumptions.

I have to say that the precision of the figures you quote makes me suspicious. How could they know what hit rates were achieved at different ranges in the chaos of combat? The only possible sources for this information - the aircrew involved - would have been far too busy to make a careful note of this even if they knew, particularly since estimation of the ranges in air-to-air combat was notoriously inaccurate.

>The second point is this: just because fighter guns were harmonised for a particular range doesn't mean that this was the expected kill range.

It certainly means that it was the expected effective range. And effective range is the range at which a kill or serious damage will result from hits.

I would say that it represents the maximum range at which hits might be expected to be scored. As we know, a B-17 required many 20mm hits to be brought down, so it could take a long time to acccumulate those hits at long range.

 

[Kurfürst]

I have to say that the precision of the figures you quote makes me suspicious. How could they know what hit rates were achieved at different ranges in the chaos of combat? The only possible sources for this information - the aircrew involved - would have been far too busy to make a careful note of this even if they knew, particularly since estimation of the ranges in air-to-air combat was notoriously inaccurate.

Gun camera films. The ones I have seen from the Luftwaffe specifically mention the range at the time of opening fire, and also what the pilot actually though the range was (typically they thought they were much closer than they really were).

Range can be estimated rather easily using the wingspan of the enemy aircraft as reference from these films, and probably there was some kind of way telling how many rounds were fired - if by no other means, tracers would tell when the weapon started/finished firing, total time was known from how many frames were shot; hits could be observed visually.

 

[Tony Williams]

Gun camera films. The ones I have seen from the Luftwaffe specifically mention the range at the time of opening fire, and also what the pilot actually though the range was (typically they thought they were much closer than they really were).

Range can be estimated rather easily using the wingspan of the enemy aircraft as reference from these films, and probably there was some kind of way telling how many rounds were fired - if by no other means, tracers would tell when the weapon started/finished firing, total time was known from how many frames were shot; hits could be observed visually.

Fair point - I should have thought of that!

However, hits could not necessarily be observed, since (especially against bombers) cannon shells were fitted with delayed action fuzes intended to explode within the plane, not on the surface.

 

[HoHun]

Hi Tony,

>The only possible sources for this information - the aircrew involved - would have been far too busy to make a careful note of this even if they knew, particularly since estimation of the ranges in air-to-air combat was notoriously inaccurate.

As Kurfürst already pointed out, the report was in fact based on gun camera films from combat situations, but additionally the results from non-combat trials were taken into account, too. (The report only mentions this in passing.)

I think the air forces were so well aware of the range estimation problem you pointed out because they were routinely comparing these films with the corresponding combat reports by the pilots.

>I would say that it represents the maximum range at which hits might be expected to be scored. As we know, a B-17 required many 20mm hits to be brought down, so it could take a long time to acccumulate those hits at long range.

Hm, since the lethality of your fire depends on the concentration of hits, you would always want to have the convergence distance closely match with the typical range at which kills are achieved.

In fact, the "maximum" range (whatever this means is always beyond convergence distance because the geometry of the diverging trajectories does not lead to a sharp drop-off of lethality, but to a gradual one. (Making it more difficult to define the "maximum" effective range.)

Here is the complete hit percentage table as used in the detailed report:

Range  - 500 m - 1 km - 1.5 km
MG 151 -  9.1% - 3.0% - 0.8%
MG 213 - 10.0% - 3.8% - 1.2%
MK 108 -  8.3% - 3.3% - 0.9%
MK 103 - 10.0% - 3.8% - 1.3%
MK 412 - 10.0% - 3.8% - 1.3%
MK 214 - 10.5% - 4.2% - 1.5%

Note that the ca. 10% hit percentage at 500 m goes well with the 5% assumption from the "Bad Eilsen" overview since you'd have to allow a greater ammunition supply to make sure to have enough in a below-average attack.

The number of hits required per weapon according to the report:

Pk     - 50% - 95%
MG 151 - 18 -  25
MG 213 - 18 -  25
MK 108 -  4 -   7.3
MK 103 -  4 -   7.3
MK 412 -  1 -   2.8
MK 214 -  1 -   2.8

Note that this is slightly more than the figures often quoted in popular books. I guess this might be because the "Bad Eilsen" overview does not quote the Probability of Kill (besides using slightly different figure of 20 hits for the MG 151/20 and 5 for the MK 108) achieved with the stated number of hits. From the direct comparison, I'd say it is only 50%, making the popular "5 hits for a kill" statement a bit suspect.

The weight of the weapon including mounting and ammunition required for a 95% Probability of Kill (Pk) from 500 m:

Pk     -    kg - firing time
MG 151 - 119.7 -  17 s
MG 213 - 194.3 -  11 s
MK 108 - 140.4 -   5 s
MK 103 - 267.7 -   6 s

The report originally was provided by Karl over at alt.games.warbirds. Unfortunately, I have only two of the pages, so some details are missing

Regards,

Henning (HoHun)

 

[Tony Williams]

Thanks for posting that information, Henning, although I am still a bit sceptical of the precision of the numbers. Having had some peripheral involvement with statistics in my career, I do know that impressively precise numbers sometimes obscure lots of rough estimation behind the calculation....In particular, I doubt that the MK 214 fired more than handful of shots in combat.

Hm, since the lethality of your fire depends on the concentration of hits, you would always want to have the convergence distance closely match with the typical range at which kills are achieved.

In fact, the "maximum" range (whatever this means is always beyond convergence distance because the geometry of the diverging trajectories does not lead to a sharp drop-off of lethality, but to a gradual one. (Making it more difficult to define the "maximum" effective range.)

An interesting subject which I'm sure we could debate for hours! Having read a lot of contemporary material in the National Archives on the RAF convergence patterns in the BoB, I don't agree that the convergence pattern was usually selected to match "the typical range at which kills were achieved"; I believe that it was normally chosen to match the maximum range at which a kill became feasible.

Don't forget that there were several factors affecting the distribution of fire from fighters: the convergence pattern was only one of them. Others were the basic dispersion factors for the guns in their mountings (the .303 Browning spread its shots over a 1 metre diameter circle at 100 metres - so at 500 metres its fire would be spread over a 5 metre circle, or about 20 square metres); another factor was "aim wander" - no pilot could keep his plane pointed at exactly the same point during a burst of fire, there was always some spread due to aircraft movement. Shot dispersion and aircraft movement would, in combination, mean that the fire from aircraft would in any case be much more widely spread at long range than at short range - so it made sense to converge the gun aiming points at long range to try to compensate for that.

An interceptor could be expected to be closing on its target quite quickly when attacking. A B-17 cruised at less than 200 mph while an interceptor could be expected to be attacking at at least double that speed, maybe more. A closing speed of 200 mph is almost 100 metres per second; so if the interceptor opened fire at 500 metres range and fired a 3-second burst, he would be at only 200 metres range when he finished. And that's in a tail attack: head on, the closing speed could be over 500 mph or 250 metres per second; if he started firing at 500 metres he would scarcely have any time to shoot before having to pull up to avoid a collision. So just because convergence distances were set for a particular range, it did not mean that most kills were achieved at that range - it merely represented the starting point of the attack.

I think that the most realistic convergence patterns were those chosen by those US pilots who set pairs of .50 cal guns to converge at different distances. The outermost pair was set to long range, the next closer and the inboard pair to close range. This gave a reasonable concentration of fire at all feasible combat ranges.

 

[HoHun]

Hi Tony,

>Thanks for posting that information, Henning, although I am still a bit sceptical of the precision of the numbers.

The precision of course is only valid in the exact context. What's important is the magnitude, and that is 10% at 500 m for all of the weapons considered.

>Having read a lot of contemporary material in the National Archives on the RAF convergence patterns in the BoB, I don't agree that the convergence pattern was usually selected to match "the typical range at which kills were achieved"; I believe that it was normally chosen to match the maximum range at which a kill became feasible.

I think this is a misconception that results from the lack of definition for "maximum effective range", which in this case in implied in the word "feasible". There is nothing magic about the convergence distance that causes bullets to become ineffective once they fly past it.

>Don't forget that there were several factors affecting the distribution of fire from fighters: the convergence pattern was only one of them. Others were the basic dispersion factors for the guns in their mountings (the .303 Browning spread its shots over a 1 metre diameter circle at 100 metres - so at 500 metres its fire would be spread over a 5 metre circle, or about 20 square metres);

Well, if you look at the trajectory of the MK 108 at 500 m, the pattern centrepoint will be just 80 cm below the crosshairs, and all hits will strike within a 75 cm circle. Despite (or in actually, because of!) its low muzzle velocity, it's a precision weapon. You really shouldn't make any conclusions from 7.7 mm machine guns as mounted in RAF fighters, they're just too different.

>Shot dispersion and aircraft movement would, in combination, mean that the fire from aircraft would in any case be much more widely spread at long range than at short range - so it made sense to converge the gun aiming points at long range to try to compensate for that.

Eglin Air Foce Base test pilot Don Lopez in "Fighter Pilot's Heaven" notes that in his gun-camera recorded tracking tests, the optimum tracking result he achieved was an accuracy of 0.5 mils. At 500 m, that's an error of just 25 cm. He also mentions that a tracking accuracy of 6 mils was considered acceptable by the US, which is a 3 m error at the distance we are looking at. Considering the size of a four-engined bomber, that is nowhere near enough to make it "non-feasible" to hit it even if its far beyond convergence distance.

The adjustment of the convergence distance you are speaking of would not be chosen to give the best destructive effect at the worst possible firing range, but rather the best destructive effect at the most useful firing range.

You have to look for the integral of Pk over time in the overtaking attack you describe - that is the figure the convergence distance will be chosen to maximize. It certainly would be an operational research screw-up to put it at some extreme limit connected with hit chances appproximating zero. (However that limit might have been defined.)

Regards,

Henning (HoHun)

 

[pete_madi]

tony if you check youtube they have some good gun camera footage mainly german a/c

 

[drgondog]

Another factor not considered in the above 'bore sight' discussions is that nose mounted weapons on a 109/P-38, etc should be more reliably close to ground bore sight results than wing mounted weapons.

The inboard weapons for an Fw 190, situated in the root, should be next and all wing mounted weapons outboard of main gear are subject to the wing bending under aero loads - which in my opinion makes them the least reliable platform relative to gunsight/weapon convergence.

 

[Tony Williams]

Henning, there is a big difference between tests on the ground, or under practice conditions, and what actually happened in combat.

I refuse to believe that an MK 108, fired in combat from a plane which was constantly having its course corrected and probably bucking in the slipstream of the bomber formation it was attacking, could possibly group its fire within 75 cm at 500 m range, or anything like it.

 

[HoHun]

Hi Tony,

>Henning, there is a big difference between tests on the ground, or under practice conditions, and what actually happened in combat.

Well, the Baade report actually notes that the dispersion measured in flight tests typically is smaller than the dispersion measured in ground tests. If anything, my numbers are conservative as a result.

>I refuse to believe that an MK 108, fired in combat from a plane which was constantly having its course corrected and probably bucking in the slipstream of the bomber formation it was attacking, could possibly group its fire within 75 cm at 500 m range, or anything like it.

Re-reading the report, it must actually be 1.5 m at 500 m range - they use the diameter symbol, but in the text explain that the figures are for the 100% hit radius.

Still, have you any figures for cannon dispersion at all? The MK 108 in a nose mounting certainly should be near the low end of the range due to its low muzzle velocity, its operating principle, and due to the stiff mount. The MG 151/20 and the MK 103 due to their higher muzzle velocity have a greater dispersion, the low-velocity MG FF has even less.

Bomber slipstream ... sure, it's a negative effect, but that's why the Luftwaffe pilots were taught to avoid attacking through the slipstream. This has nothing to do with the weapon at all.

It's obvious that if the pilot would make an optimum gunnery run with 0.5 mil error, shooting with a weapon that has 1.5 mil dispersion, a bomber-sized target would be hit with 100 % probability at 500 m. The Luftwaffe expects a hit probability of around 10 % in combat, and there is no technical reason that shouldn't be possible as the main source of error is the pilot, not the cannon or the airframe.

Regards,

Henning (HoHun)

 

[HoHun]

Hi again,

>The dispersion strongly varies between different mountings, and for some reason the B-17 tail turret is the least accurate mounting (except for the even more inaccurate hand-held tail guns).

Reading the manual for the German FDL-B 131 remote control turret as used for the B turret in He 177, armed with a 13 mm MG 131, it seems that the mechanical aiming error of that turret is expected to be 10 mil judging from the calibration procedure.

Accordingly, the dispersion of the weapon would be somewhat above that 10 mil figure, so it would probably be roughly in the same order of magnitude as for the US turrets (Sperry upper turret - 11.7 mil, Bendix chin turret - 12.5 mil). Depending on the rigidity of the mount, I'd guess dispersion might even be slightly larger than these values.

(Calibration requires four bursts of seven shots each to be fired after pointing the turret at the aim point from the left, right, bottom and top respectively. The centre of each group of hits should be with 50 cm of the aim point at 100 m range.)

Regards,

Henning (HoHun)

 

[seesul]

OK, and back to the B-17 with the 20mm cannon instalation.
I got scanned the picture and it is in jpg file this time so I´ve attached it again.
The text written in the book says:
Opposite centre left: a 20mm cannon in the tail position of the 97th BG´s B-17G 42-38090, one of at least 5 such installations made by this group in the spring of 1944. The sight was a ring-and-bead on the end of the long rod attached to the gun barrel. As M/Sgt Morley Russell, who worked on this installation, commented,´The sight sure was a real Rube Goldberg type, but it worked well.´ (Morley L. Russelll)
Opposite bottom: B-17G 42-38090, on a mission in July, displays the 20 mm cannon installation. Because of the limited movement of the gun it had a poor field of fire, but its purpose was to counter the enemy fighters that fired rockets from outside the range of the normal .50 calibre armament. B-17s with these tail cannon were flown at the rear of formations, and the weapon was considered more of psychological that destructive value (Steve Birsdall).

Roger Freemann/David Osborne-The B-17 Flying Fortress story, page # 44-45

 

[seesul]

But I have a question- on the end of Hispano barrell there´s something like one more barrel, or at least a bush...does anyone know what is it for?
Or is there .50 mounted together with the M2 Hispano???

 

[seesul]

Anyone here from Michigan state that could help me with research?
Morley L. Russel, 90 years old, lives in Shelby Township, Michigan, according to white pages.
Dunno if is him, who made this 20mm cannon installation, but the name is the same. Could be nice and interesting to find him, ask him few questions and show him the pics I have...
Thanks in advance!

 

[seesul]

Here´s the actual list of the Forts with the tail mounted M2 Hispano I´ve found by now:

42-31885 (99th BG, later 2nd BG) - picture from the crash site
42-32046 (483rd BG, later 99th BG)- The B-17 Flying Fortress Story by Roger Freeman, note on page 174
42-38090 (97th BG)- The B-17 Flying Fortress Story by Roger Freeman, pic of the tail on page 44, note on page 45
42-97490 (2nd BG)- Konečná zastavka Slovensko (Final Destination Slovakia) by Peter Kassak, pic of the wreckage,page 84
42-38087 (2nd BG)- mentioned in the 2nd BG database
42-31590 (2nd BG)- mentioned by Mr. Steve Birdsall
42-97438 (2nd BG)- mentioned by Mr. Steve Birdsall

 

[B-17engineer]

Nice job with the info Roman! Looks interesting! The chine turret with six guns looks lethal

 

[seesul]

Nice job with the info Roman! Looks interesting! The chine turret with six guns looks lethal

But this chin turret must have been pretty hungry for shells...I´d like to see its feeding system...

 

[vikingBerserker]

Wow, that's pretty sweet. Never even heard of a 6 gunned turret before.

 

[seesul]

The picture on the first page of this thread shows the wreckage of B-17G, 42-31885 that crashed close to my born town.
We weren´t sure where the 20 mm cannon was but the most probably position was in tail. Now, after reading the book 'Mighty By Sacrifice' on page 92-94, where Bob Donahue, a tail gunner desribes his experince from July 9, 1944 during the raid on Ploesti, we know more. He says there that on this mission he manned a 20 mm cannon. From the database of 2nd BG was easy to find out that he flew 42-31885 that day. So 20 mm cannon instalation on this machine in tail confirmed from another source.

 

[drgondog]

One of the bomb groups in 8th AF experimented with a 20mm nose mount but took it out because the nose had no load paths to carry the stress of the recoil.