Aerodynamics of Japanese fighters (1 Viewer)

[GregP]

Here's a Seafang I drew awhile back.

 

[wells]

@ThomasP my understanding is that it was 7075 not Alu 7076 but those are almost the same material anyway. Minor quibble on my part but totally agree with everything you wrote. EDIT: I misread and you did write that 7076 is essentially the same thing as ESD whereas 7075 was introduced by Japan in 1943. Mind blowing info. Thanks!

Regarding the primary discussion topic, aerodynamics, I believe the main things that we're all talking about are the zero-lift drag coefficient and wing area of an aircraft. I only have a low-level, mechanical understanding of these principles of flight, but a quick summary for WW2 aircraft is as follows:

In general, the major aerodynamic variables for drag are the zero-lift drag coefficient and the wing area. If you have both, then you can calculate how aerodynamically efficient an air frame is.

Japanese aircraft tended to have particularly good streamlining, but then they also had big wings relative to horsepower and weight, which watered down their excellent streamlining.

But this is all speculation as I have never seen a ZLDC calculated for any Japanese aircraft, not even the Zero. Although some aviation buffs have used estimated ZLDCs which were reverse engineered from TAIC records, these are of unknown pedigree as well. This video does a really great job of reverse engineering TAIC's calculations, and some Japanese sources, to get the ZLDC for the Ki-84. But as the video pointed out, TAIC appears to have just transplanted the Ki-43's ZLDC to the Ki-84's wing area. Additionally, Nakajima burned all of their records after the war, so the final drag estimate is also just an estimate with no valid sourcing. This is definitely not the right number. And we don't actually know how the Frank's aerodynamic efficiency was.

Just a side note: I'd love to see the zero lift drag coefficient calculated for the N1K2-J and the N1K1-J but my guess is that it doesn't exist. But because it has laminar airfoils and a large propeller, it's a good indicator of Japanese aerodynamic technology.

The Japanese were early adopters of "laminar" flow airfoils. Kawanishi was the first company to fly such an aircraft outside of the US, but the J2M1 Raiden also had a semi-laminar airfoil (according to Japanese Wikipedia, which sources Maru Mechanic IIRC for this). However, except for the Suisei dive bomber and the Ki-61, all Japanese aircraft were relatively high drag radial aircraft, with extra emphasis spent on streamlining to compensate for their lower horsepower output as well as extensive hand finishing using machinists.

The Mustang (D model), as far as I'm aware, has the lowest Zero-Lift Drag Coefficient out of all aircraft produced in large quantities during the war. I think much of that is because it uses a liquid-cooled engine. The Ki-64 probably had an overall lower zero-lift drag coefficient though but that was a prototype.

That was my ki-84 video and I'm working on calculations for other Japanese fighters as well. Regarding the N1K2-J, it is quite comparable to an F4F-4 in terms of wing span, area, weight, pudginess? So, if you put a 2000 hp engine on the Wildcat, it should be close to "George". Using the Standard Aircraft Characteristics data for the F4F-4, its drag works out to 0.0198 for 260 sq ft ( A base value, corresponding to Mach 0. Compressibility correction is applied based on Mach number ), or an "equivalent flat plate area ( f )" of 0.0198 x 260 = 5.15 sq ft. This was arrived at from knowing the speed of 284 mph at sea level with 1200 hp and a 9.75 ft prop with 77.5% propeller efficiency.

The next step was to put a 3.3 m prop on it and give it 2000 hp. The speed goes up to 337 mph at sea level. Taking it up to 6000 m, with 1625 hp, the speed is 375 mph TAS. So, looking at the published figures for "George" of 369 mph, you can see it is right in the ballpark and very comparable to a Wildcat in drag terms. I suspect the fatter fuselage counters the laminar wing to a large extent.

 

[wells]

I wish there was performance data available for the M6A1-K, which was the land-based version of Aichi's M6A1 Seiran.

Would probably be similar to the difference between A6m2 and A6m2-N

 

[AerialTorpedoDude69]

That was my ki-84 video and I'm working on calculations for other Japanese fighters as well. Regarding the N1K2-J, it is quite comparable to an F4F-4 in terms of wing span, area, weight, pudginess? So, if you put a 2000 hp engine on the Wildcat, it should be close to "George". Using the Standard Aircraft Characteristics data for the F4F-4, its drag works out to 0.0198 for 260 sq ft ( A base value, corresponding to Mach 0. Compressibility correction is applied based on Mach number ), or an "equivalent flat plate area ( f )" of 0.0198 x 260 = 5.15 sq ft. This was arrived at from knowing the speed of 284 mph at sea level with 1200 hp and a 9.75 ft prop with 77.5% propeller efficiency.

The next step was to put a 3.3 m prop on it and give it 2000 hp. The speed goes up to 337 mph at sea level. Taking it up to 6000 m, with 1625 hp, the speed is 375 mph TAS. So, looking at the published figures for "George" of 369 mph, you can see it is right in the ballpark and very comparable to a Wildcat in drag terms. I suspect the fatter fuselage counters the laminar wing to a large extent.

Hey thanks for providing that information. Your video is fantastic.

Regading the George 21, The George 21 came with three primary engine variants, two of which were derated from its maximum horsepower. The same is true for the Ki.84.

The derated engine that was listed in the George 21 Prototype's Handbook mentioned that the top military speed was 644 KPH. Laurelix calculated its WEP at around 658 KPH. Japanese Wikipedia also has the same numbers, which are derived from the Shiden Kai manual. I also applied the equation for determining the aircraft's top speed in that thread and the math appears to be correctly done (if you use the only authoritative source available).

Also, the top speed in the manual aligns with some IJN pilots' recollections for the top speed of the Shiden. As far as I'm concerned, this is the truest value we're going to get given the available data.

The 369 MPH number is of unknown pedigree. It was derived from a 1946 Japanese source hastily compiled from memory. I believe this is the military rating of the George 11b (or N1K1-J-Otsu) which had streamlined its wing. Japanese records specifically state that removing the gun pods added around 7 MPH.

The report originated from a US government request for official performance values from the Japanese. Unfortunately, many of the official records had been destroyed. The only authoritative text remaining is the Shiden Kai prototype handbook. The best secondary sources after that are pilot biographies which mention and corroborate the manual.

 

[GrauGeist]

Would probably be similar to the difference between A6m2 and A6m2-N

Well, I'm sure there would be a differwnce in performance, but it would certainly be quite a difference.

The A6M2 Model 21 had speeds of 317mph plus, while the A6M2-N had speeds of roughly 270mph due to the main float and outriders.

The M6A1 had a top speed of 295mph with it's twin floats where the M6A1-K had retractable landing gear (it was land-based) and the upper folding portion of the vertical stabilizer omitted.

 

[wells]

Hey thanks for providing that information. Your video is fantastic.

Regading the George 21, The George 21 came with three primary engine variants, two of which were derated from its maximum horsepower. The same is true for the Ki.84.

The derated engine that was listed in the George 21 Prototype's Handbook mentioned that the top military speed was 644 KPH. Laurelix calculated its WEP at around 658 KPH. Japanese Wikipedia also has the same numbers, which are derived from the Shiden Kai manual. I also applied the equation for determining the aircraft's top speed in that thread and the math appears to be correctly done (if you use the only authoritative source available).

Also, the top speed in the manual aligns with some IJN pilots' recollections for the top speed of the Shiden. As far as I'm concerned, this is the truest value we're going to get given the available data.

The 369 MPH number is of unknown pedigree. It was derived from a 1946 Japanese source hastily compiled from memory. I believe this is the military rating of the George 11b (or N1K1-J-Otsu) which had streamlined its wing. Japanese records specifically state that removing the gun pods added around 7 MPH.

The report originated from a US government request for official performance values from the Japanese. Unfortunately, many of the official records had been destroyed. The only authoritative text remaining is the Shiden Kai prototype handbook. The best secondary sources after that are pilot biographies which mention and corroborate the manual.

Interesting...It makes sense that the airplane would be tested with lower grade fuel ( hence being de-rated ), since the good stuff would go to the front line combat units as a priority. Or, it just wasn't available at the time. The US did the same thing. For training, they used 91/96 grade and gave the engine limits in the handbook. As an example,

The figure of 321 knots at 5600 m is actually exactly the same as 330 knots at 6000m, as given in the handbook. The indicated speed is 240 knots in both cases. The airplane would have been tested at 6000m, since it was to meet the same specification as the J2M ( 325 knots at 6000 m ) and the Americans actually captured a document containing that figure. They thought it was low also, but maybe didn't realize it was the speed to which the airplane was to meet or exceed in the specification. Later, they found more information that gave the figure of 354 knots. This is where the figure of 407 mph comes from in the TAIC manual. Anyway, if the density altitude was 5600 m, it means the test flight took place on a cold day, where the outside air temperature was about 15C colder than "standard". Once corrected back to "standard" atmosphere, you get the figure of 330 knots at a density altitude of 6000 m.

I was always under the impression that the figure of 348 knots was just a calculated figure to which the airplane failed to meet. According to Jiro Horikoshi in "Eagles of Mitsubishi", they were using an "obsolete" method of calculating top speeds, which is why the Zero never met their expectations. He doesn't say what the issue was, but I suspect it was compressibility correction. This would have been an industry wide problem. In any case, it now makes more sense that it corresponds to the fully rated engine, versus the derated engine, especially given the climb times, which compressibility drag by itself cannot make up the difference. And, why would they publish such data in the manual? "Look at how bad our airplane performs, compared to what we thought it would do! " haha

So, I agree that 348 knots at 6000 m is the speed with rated power of 1700 HP ( 1675 hp ) @ +350 mm boost / 3000 rpm ). I will post more in the other thread. On a final note, my calcs give a Cd0 ( Mach 0 ) of 0.0174, if 330 knots at 6000 m is achieved with 88% rated power ( about 1500 HP, if power varies directly with manifold pressure and RPM ). Then, applying rated power, I get the figure of 345 knots, only 3 knots less than that published in the manual. So, that makes perfect sense.

 

[Simon Thomas]

Nice work on the video. That would have taken a while to put together.

In post #22, was your prop efficiency of 77.% found or calculated?

Power does not vary linearly with RPM, however for small changes in RPM it is likely close enough.
Power is meant to vary linearly with manifold pressure, but typically other things are changing at the same time so it never seems to work out correctly. Again small changes will likely be close enough to linear.

 

[wells]

Nice work on the video. That would have taken a while to put together.

In post #22, was your prop efficiency of 77.% found or calculated?

Purely calculated. Then, the drag is determined from the calculated thrust at some known speed/alt/power combination. It is an iterative process to determine inflow losses. The propeller sees a higher velocity than the freestream velocity. The results are quite accurate when compared to airplanes with well documented performance ( mostly British and American ). I will probably post a couple of sample calculations in the P-51 or Corsair thread at some point to illustrate. But then, the same method can be used to determine performance of any airplane, even with minimal data. All I need really, is one airspeed/altitude/power combination. For example, I've recently done calculations for the J7W1 ( not published anywhere yet ) and the results are within 1% at both sea level and 8400 m, compared to the Japanese calculations done 80 years ago. The performance of that airplane was never determined by flight test, since it only made 3 short flights of about 15 minutes each and had major trim issues, which were never sorted before the war ended. My goal is to expand on known performance figures and show things like glide ratios ( range and endurance ), airspeed for best climb rate or angle, turn performance, etc.

 

[Holtzauge]

Purely calculated. Then, the drag is determined from the calculated thrust at some known speed/alt/power combination. It is an iterative process to determine inflow losses. The propeller sees a higher velocity than the freestream velocity. The results are quite accurate when compared to airplanes with well documented performance ( mostly British and American ). I will probably post a couple of sample calculations in the P-51 or Corsair thread at some point to illustrate. But then, the same method can be used to determine performance of any airplane, even with minimal data. All I need really, is one airspeed/altitude/power combination. For example, I've recently done calculations for the J7W1 ( not published anywhere yet ) and the results are within 1% at both sea level and 8400 m, compared to the Japanese calculations done 80 years ago. The performance of that airplane was never determined by flight test, since it only made 3 short flights of about 15 minutes each and had major trim issues, which were never sorted before the war ended. My goal is to expand on known performance figures and show things like glide ratios ( range and endurance ), airspeed for best climb rate or angle, turn performance, etc.

There seems to be a big spread in the opinions about the A6M2 Zero's top speed in this forum, so I would be interested in your take on this as in what speed do you think it was capable of at different power settings and why. I recently published a paper comparing the P-40E with the A6M2 Zero which you can find here and in which I detail my take on its top speed and the reasoning around why I chose the tuning data I did. Seeing you seem to have similar interests, what is your take on the A6M2 Zero's top speed?

 

[GregP]

Beautiful. I really liked the framing of the Zero with the red dot star on the camera plane's wing. What plane is it?

AT-6

 

[AerialTorpedoDude69]

Interesting...It makes sense that the airplane would be tested with lower grade fuel ( hence being de-rated ), since the good stuff would go to the front line combat units as a priority. Or, it just wasn't available at the time. The US did the same thing. For training, they used 91/96 grade and gave the engine limits in the handbook. As an example,

The figure of 321 knots at 5600 m is actually exactly the same as 330 knots at 6000m, as given in the handbook. The indicated speed is 240 knots in both cases. The airplane would have been tested at 6000m, since it was to meet the same specification as the J2M ( 325 knots at 6000 m ) and the Americans actually captured a document containing that figure. They thought it was low also, but maybe didn't realize it was the speed to which the airplane was to meet or exceed in the specification. Later, they found more information that gave the figure of 354 knots. This is where the figure of 407 mph comes from in the TAIC manual. Anyway, if the density altitude was 5600 m, it means the test flight took place on a cold day, where the outside air temperature was about 15C colder than "standard". Once corrected back to "standard" atmosphere, you get the figure of 330 knots at a density altitude of 6000 m.

I was always under the impression that the figure of 348 knots was just a calculated figure to which the airplane failed to meet. According to Jiro Horikoshi in "Eagles of Mitsubishi", they were using an "obsolete" method of calculating top speeds, which is why the Zero never met their expectations. He doesn't say what the issue was, but I suspect it was compressibility correction. This would have been an industry wide problem. In any case, it now makes more sense that it corresponds to the fully rated engine, versus the derated engine, especially given the climb times, which compressibility drag by itself cannot make up the difference. And, why would they publish such data in the manual? "Look at how bad our airplane performs, compared to what we thought it would do! " haha

So, I agree that 348 knots at 6000 m is the speed with rated power of 1700 HP ( 1675 hp ) @ +350 mm boost / 3000 rpm ). I will post more in the other thread. On a final note, my calcs give a Cd0 ( Mach 0 ) of 0.0174, if 330 knots at 6000 m is achieved with 88% rated power ( about 1500 HP, if power varies directly with manifold pressure and RPM ). Then, applying rated power, I get the figure of 345 knots, only 3 knots less than that published in the manual. So, that makes perfect sense.

Sorry to drag this thread out, just wanted to add a bit more.

Yes, you're right. The derated Homare 21 wasn't outputting 1675 HP at 6,000M. But most sources claim it wasn't using a lower octane than 91/92.

In Laurelix 's tables, it lists the derated engine's maximum output. I believe it had been downgraded from a maximum of 3,000 RPMs with 500mm of pressure to something like 2900 RPMs with 350 mm of pressure. But that is from memory. The tables also include the WEP values.

N1K2-J Shiden-Kai Performance

There are many misconseptions and confusions regarding the real performance of this fighter. If you look at the Specification of performance of IJN Fighters table from 10th september 1945, or US TAIC for George 21 or sources like Francillon. They all state the Performance of the N1K2-J was...

ww2aircraft.net

As far as I'm aware, the manual states that there are two speeds: the derated speed and the theoretical speed. The derated military speed was at 611 KPH, although the manual lists a range of speeds. It's been a while since I read through those tables, but I believe that table 3 is displaying the derated horsepower and table 4 is displaying the theoretical horsepower of a fully rated engine (500mm/3,000RPMs).

A primary source (US Bombing Survey IIRC)` on Japanese fuel state that the Japanese military had access to 92 Octane throughout the war and had adequate stockpiles. The US's report on Japanese fuel states that a 95-octane fuel distilled from pine resin existed in large quantities. Some anecdotal reports mention that it was mixed with lower octane fuel. However, the 95-octane fuel was highly volatile causing pre-detonation due to aromatic compounds in it. So I believe some of the issue with reliability had to do with the Japanese mixing 95-octane with 87-octane fuel. But this is my interpretation of the available data. However some sources claim very little pine-root oil made its way into aircraft.

According to everything I've read, the main issue with the de-rated engines was that many were poorly assembled and had to have their RPMs and boost pressure limited because they had problems operating at their peak output. But the picture is complex and suffers from two competing narratives. One narrative is that the fuel stocks had been "watered" down so that what was listed as A87G (a volatile fuel) was more like A85G (the G means volatile).

Nakajima's chief engineer in charge of the development of the Homare published a paper on how they had to come up with a way to get the engine to run on 87-92 octane, down from 100 octane. He wrote that the main hurdle was developing a lubrication system for a high performance engine. So it might be that the pine-oil aromatics causing pre-detonation is a myth and that the engines had design problems. In fact, according to the Japanese entry on the Ki-84, 87 octane fuel was only used for training flights. Otherwise, 91/92 was required.

A chief mechanic who got non-lemon Franks' operational status up to 100% said that it was actually faulty fuel injector nozzles and once these and a few other technical issues were field repaired, the engine was actually very reliable.

 

[wells]

Sorry to drag this thread out, just wanted to add a bit more.

Yes, you're right. The derated Homare 21 wasn't outputting 1675 HP at 6,000M. But most sources claim it wasn't using a lower octane than 91/92.

In Laurelix 's tables, it lists the derated engine's maximum output. I believe it had been downgraded from a maximum of 3,000 RPMs with 500mm of pressure to something like 2900 RPMs with 350 mm of pressure. But that is from memory. The tables also include the WEP values.

N1K2-J Shiden-Kai Performance

There are many misconseptions and confusions regarding the real performance of this fighter. If you look at the Specification of performance of IJN Fighters table from 10th september 1945, or US TAIC for George 21 or sources like Francillon. They all state the Performance of the N1K2-J was...

ww2aircraft.net

As far as I'm aware, the manual states that there are two speeds: the derated speed and the theoretical speed. The derated military speed was at 611 KPH, although the manual lists a range of speeds. It's been a while since I read through those tables, but I believe that table 3 is displaying the derated horsepower and table 4 is displaying the theoretical horsepower of a fully rated engine (500mm/3,000RPMs).

A primary source (US Bombing Survey IIRC)` on Japanese fuel state that the Japanese military had access to 92 Octane throughout the war and had adequate stockpiles. The US's report on Japanese fuel states that a 95-octane fuel distilled from pine resin existed in large quantities. Some anecdotal reports mention that it was mixed with lower octane fuel. However, the 95-octane fuel was highly volatile causing pre-detonation due to aromatic compounds in it. So I believe some of the issue with reliability had to do with the Japanese mixing 95-octane with 87-octane fuel. But this is my interpretation of the available data. However some sources claim very little pine-root oil made its way into aircraft.

According to everything I've read, the main issue with the de-rated engines was that many were poorly assembled and had to have their RPMs and boost pressure limited because they had problems operating at their peak output. But the picture is complex and suffers from two competing narratives. One narrative is that the fuel stocks had been "watered" down so that what was listed as A87G (a volatile fuel) was more like A85G (the G means volatile).

Nakajima's chief engineer in charge of the development of the Homare published a paper on how they had to come up with a way to get the engine to run on 87-92 octane, down from 100 octane. He wrote that the main hurdle was developing a lubrication system for a high performance engine. So it might be that the pine-oil aromatics causing pre-detonation is a myth and that the engines had design problems. In fact, according to the Japanese entry on the Ki-84, 87 octane fuel was only used for training flights. Otherwise, 91/92 was required.

A chief mechanic who got non-lemon Franks' operational status up to 100% said that it was actually faulty fuel injector nozzles and once these and a few other technical issues were field repaired, the engine was actually very reliable.

The Homare engine in itself was very ambitious. A radial with 8 to 1 compression ratio, turning 3000 rpm. Trying to get 2000 hp from 35.8L ( The PW R-2800 is 45.9 L ) would be quite an achievement.

The full rated engine had the following settings,

Takeoff ( 1 minute ): +500mm boost / 3000 RPM
Rated ( 30 minute ): +350 mm boost / 3000 RPM

The "de-rated" engine was +450 / 2900 for takeoff and +250 / 2900 rated. So, we're comparing +250/2900 to +350/3000

 

[AerialTorpedoDude69]

The Homare engine in itself was very ambitious. A radial with 8 to 1 compression ratio, turning 3000 rpm. Trying to get 2000 hp from 35.8L ( The PW R-2800 is 45.9 L ) would be quite an achievement.

The full rated engine had the following settings,

Takeoff ( 1 minute ): +500mm boost / 3000 RPM
Rated ( 30 minute ): +350 mm boost / 3000 RPM

The "de-rated" engine was +450 / 2900 for takeoff and +250 / 2900 rated. So, we're comparing +250/2900 to +350/3000

If the engine were specially assembled to meet factory standards, the engine did meet the expected horsepower output. But many production engines weren't capable of meeting those output expectations. Japanese wikipedia mentions that they had top military speeds from around ~580 KPH to 611 KPH. And I do believe it was the result of a draft program that did not provide exceptions for critical labor. Many skilled machinists were sent to front line units (presumably to work as machinists) and replaced with high school students.

But again, a Captain Kariya managed to solve the reliability issues which he attributed to defective injector nozzles, spark plugs, and a few other subcomponents.

 

[wells]

If the engine were specially assembled to meet factory standards, the engine did meet the expected horsepower output. But many production engines weren't capable of meeting those output expectations. Japanese wikipedia mentions that they had top military speeds from around ~580 KPH to 611 KPH. And I do believe it was the result of a draft program that did not provide exceptions for critical labor. Many skilled machinists were sent to front line units (presumably to work as machinists) and replaced with high school students.

But again, a Captain Kariya managed to solve the reliability issues which he attributed to defective injector nozzles, spark plugs, and a few other subcomponents.

Yes, my understanding is that the Homare 23 used a low pressure injection system, which was apparently an improvement.

If the pilot were to use the takeoff rating for "emergencies", the speed would be higher than 348 knots. And there is also the issue of ram air, raising the critical height ( and the true airspeed again ). The airplane was tested at 6000 m, because that was the specification, but doesn't mean it wasn't faster at some higher altitude. I haven't crunched those numbers yet, will probably do so today sometime, because now I'm curious.

 

[wells]

There seems to be a big spread in the opinions about the A6M2 Zero's top speed in this forum, so I would be interested in your take on this as in what speed do you think it was capable of at different power settings and why. I recently published a paper comparing the P-40E with the A6M2 Zero which you can find here and in which I detail my take on its top speed and the reasoning around why I chose the tuning data I did. Seeing you seem to have similar interests, what is your take on the A6M2 Zero's top speed?

Well, about 20 years ago, I looked at it for a flightsim I was consulting on and came to the conclusion that I had no reason to dispute the figures published in Francillon, for example, as they matched Jiro Horikoshi's recollections exactly. I think he should know! He says 565 km/h ( 351 mph ) for model 52, which was 20 km/h faster ( 545 km/h / 338 mph ) than previous model ( A6M3 ), which was 10 km/h faster than A6M2 ( 535 km/h = 332 mph ). I recall also taking the performance published in the American "manual" for the model 52 and finding the drag coefficient was essentially the same. That was good enough for me. Every airframe or engine is going to be a little different. There could be up to 10 mph variation either way. Your data also matches and adds a source that I didn't have ( the manual for the airplane ). I don't think there's any dispute. I would need to actually do the math to see what the "emergency" performance would look like.