Greg's Airplanes: Nakajima Ki-84 Hayate "Frank" History

[taly01]

I think the problem with the Ki-84 propellor was wartime rushed production, along with loss of unskilled staff etc It must have worked OK on the well maintained prototypes. Its likely any weakness in the system affected the high speed/power pitch for the kind of reasons you mentioned. Japanese WW2 electrics do not have a good reputation.

One Japanese site I've recently found WW2航空機の性能:WarbirdPerformanceBlog (mentioned here on WW2aircraft somewhere?) says there were basically 3 versions of Homare, and when you understand that all the aircraft performance figures make more sense.

Ha-45-11 early production not good, boost limited , most prototype planes 1943 early 1944 used this engine!
Ha-45-12 good engine basically same as Ha-45-21 but boost limited, used on bombers more.
Ha-45-21 the planned design full boost, used on fighters more.

He also considers differences in stated HP critical heights is due to different planes have different cowl ram air intakes.

 

[Metrallaroja]

Apologies for double posting, but qaz M Metrallaroja I found the post where krieghund posted the engine test of the Homare. The partial clip only references that the Homare did output its rated power but unfortunately we do not have the full power output tables for it.

That is just an extract from this document: http://www.wwiiaircraftperformance.org/japan/K-84_MATSC.pdf

 

[Metrallaroja]

As previously said japanese max speed power makes very little sense power wise.

 

[Metrallaroja]

On the other hand TAIC makes a lot more sense.

 

[Metrallaroja]

No ram air intake that I know makes a 2000m pressure difference between 250kph and 600kph...
That japanese document is flawed and useless, any claim to trust that document instead of TAIC is equally flawed imo

 

[AerialTorpedoDude69]

M Metrallaroja taly01 I think I now understand why the Japanese documentation appears the way it does and why the original source documentation cited by Japanese websites was recorded by US TAIC/TAIU. The date on the document is Oct. '45, after the war ended during the exact time period during which US and Australian TAIC were interviewing Nakajima engineers and groundcrew.

In other words, these were not official Nakajima documents derived from real test data, these were responses to interview questions. That means these were highly complicated performance values that were reconstructed from memory.

We do know that the Australian TAIC documentation states that, based on their (EDIT) interviews with US(/EDIT) mechanics, Japanese propeller technology was "five years" behind the Allies. Unfortunately, this is an anecdotal report and we don't know the actual state of Japanese prop technology. However, this does suggest that the prop could have been a source of underperformance. But, again, this is based on interview data and therefore isn't as reliable as actual test data. What complicates the document is that TAIC synthesized a final report based on interview data with dozens of individuals, from mechanics to engineers. So we're not dealing with just one source of potential error. We're dealing with many.

But regarding the performance data that we see in Maru Mechanic and other Japanese sources, these strongly appear to be sourced from, or derived from, US documentation. In my opinion, this absolutely is based on the October 1945 TAIC report because it's written using the same English language typewriter with the same typeset that TAIC used to write their reports. But I'm not an expert on 1940s typewriters. Anyway, the odd performance curves could be a sign that either the blower settings varied throughout the production run of the Frank or someone misremembered some crucial bit of data.

Because they are interview questions they are bound to include errors. In fact, interview questions are regarded as being the least reliable form of data in the sciences.

This lines up with the information about Nakajima burning their records. If Nakajima burned their records, the Japanese sources are unfortunately not as valid as TAIC's benchtest of the engine. It's also notable that they do not use the full 500mm of boost that Greg spoke about and I don't see any attempt to synthetically derive the WEP values from such data.

I'm sad to say that what we believe to be the Japanese data is not really what we think it is. Nor is the TAIC data.

M Metrallaroja by the way, thank you for clarifying and annotating that document for the Homare 11. I really wish we had the data for the Homare 22, 23, and 25. My guess is that the Homare 22 was the first model to use a reduced compression ratio, whereas the 25 and 23 (due to their increased horsepower) seem like they may have had the 1:8 CR, but IDK.

 

[taly01]

The problem is alot of data floating around and looking very professional is engineers estimates. AFAIK the only real test data we have for Ki-84 is the prototype one.

In Burma around October 1944 50th Sentai converted to Ki-84 and it took months before the engines were reliable for combat, the 64th Sentai actually refused the offer to change their Ki43-III to Ki84 from what they saw 50th Sentai suffering.

the Japanese sources are unfortunately not as valid as TAIC's benchtest of the engine.

I believe the TAIC bench tested the engine before fitting it to the plane, for safety reasons mainly before flying, but did they ever actually test the HP? A test run stand is not necessarily a BHP test stand!

 

[AerialTorpedoDude69]

The problem is alot of data floating around and looking very professional is engineers estimates. AFAIK the only real test data we have for Ki-84 is the prototype one.

In Burma around October 1944 50th Sentai converted to Ki-84 and it took months before the engines were reliable for combat, the 64th Sentai actually refused the offer to change their Ki43-III to Ki84 from what they saw 50th Sentai suffering.



I believe the TAIC bench tested the engine before fitting it to the plane, for safety reasons mainly before flying, but did they ever actually test the HP? A test run stand is not necessarily a BHP test stand!

That's the relevant question. In the fragments that we have, TAIC refers back to the benchtest as indicating it did reach its stated horsepower of 1970 HP or something like that. But we don't have the full report.

The full report might be in this document: OPNAV-16-V No. T231, February 1945, technical air intelligence center report No. 31, Homare 11 and 21 engines, principal characteristics and performance. Report No. 13-c(26), USSBS Index Section 6. Greg's video refers back to the tested horsepower rating and this document is the only Naval document that specifically deals with the Homare 21 and 11.

To make this point clear, let me reemphasize this: This document isn't on the Blueprint Homare 21 but rather the very first Homare that was captured by the Allies. (At Clark Airfield.)

Regarding the Blueprint, rebuilt Homare, I don't see any document supposedly held within the USSBS archives that refers back to it. Although Greg made frequent references to this report as well. Unfortunately, I don't see where he hosts his sources and apparently he only makes those documents available to his Patreon subscribers. He definitely did not conjure that information out of thin air and I've got to assume it's out there somewhere.

According to Greg, the rebuilt-to-better-than-factory-spec engine was tested. It must have been tested at Middleton in 1946.

Regarding the information on the prototype Ki-84, isn't that information also based on interviews with Nakajima personnel? Or did the Japanese retain at least some records?

EDIT: I found a source that mentions three Franks were brought to the US. The first, known as "number 46" was a Ki-84 Ko made in the Summer of 1944. This was rebuilt. The other two were brought over post-war and were a Ko and an Otsu, late-model builds. There are no available performance evaluations, to my knowledge, on the last two.

 

[DarrenW]

In other words, Nakajima traded top speed for faster acceleration. If they upped the voltage in their electrical system they might have hit higher pitch angles, which in theory would compensate for the shorter prop blade. But if they never hit that higher voltage electrical system, then the prop never would have reached its maximum potential. That is a serious design flaw.

I have to say I think we've broken new ground here. I never thought that propeller pitch issues could have anything to do with this but now it's seems far more plausible.

Thanks to all for digging deep and presenting both new and valuable information. It will be nice to finally have an opinion on this that is not based solely on gut feeling or trust/non-trust of existing source material.

 

[AerialTorpedoDude69]

I had time for a little more research and I've got to believe two things:

1. The Frank's propeller definitely was a design flaw. The short length and low pitch angles (I'd like to one day get a copy of the diagram, but for now I assume it's correct) guarantee that its horsepower at max boost isn't converted into thrust.
2. The Frank's abundant wiring suggest that Nakajima intended on increasing its electrical system's voltage in later production versions.

So first up, let's talk a bit about US electric props and why pitch angle and prop size were important to thrust:

The only electric propeller used by the US military during WW2 was the Curtiss Electric (which is why the Middleton report compares the Frank's Pe-32 propeller to the Curtiss Electric prop). The Curtiss Electric prop and electric governor were used on a lot of aircraft, including the P-40, P-38, and some versions of the Mustang. The governor was produced under four primary model types: the 100008, the 100001, the 100005, and the 100006. But of these. only the 100008 was common during the war. Yes, if you click on that link, someone is selling a Curtiss Electric 100008 governor, for not that much money.

Getting back to the topic, you can see there's a direct relationship between prop size, engine horsepower, and pitch limit. When the Allies put a Merlin in a P-40L and increased its pitch limit, they got a maximum speed of 368 MPH. The higher pitch limit allows for the horsepower to be fully utilized at maximum output. However, the big question is how much voltage do you need (VA=W)? The Warhawk had a 24-volt system.

According to Greg, the Frank had a really low voltage system, yet extensive wiring. Greg speculated that they planned on putting more wood in the fighter, which doesn't make sense (wood is resistive and won't conduct electricity). The more likely explanation is that Nakajima planned on upgrading to a higher voltage system and therefore a better pitch angle on the governor because more volts = more push (V=MA). Because the Frank tested by TAIC is from Clark, it was not the final production series.

In all likelihood, later versions of the Frank were probably capable of better pitch angles. The US (on a bigger, badder aircraft) managed to get the pitch limit to 65-degrees inches on an electric prop. To illustrate, the 100008-3J pitch governor on the Black Widow P-61C could hit 65-degrees inches. This was the fastest variant of the Widow, although it was more or less a post-war model that never saw combat during WW2 AFAIK. My guess is that the Japanese were probably able to increase the pitch governor's angle although there's no evidence of that. The only thing we do know is that the Frank that was tested was manufactured in the Summer of 1944. That's toward the beginning of the Frank's production run. In other words, we know early production aircraft are quite different from late production. But I digress. Getting back to the pitch angle.

You can see that the US also had problems getting higher pitch angles on their electric props. The published pitch angles for electrics are similar to the angles listed by taly01 for the Pe-32 although the prop used on the Merlin-powered P-40L reaches a maximum pitch angle of 58. In comparison, I'm positive that the Hydromatic prop could hit much higher pitch angles. But more or less, it's telling that the US only used electric props on lower horsepower engines. On 2,000 HP and up, they seemed to universally use hydraulic props.

However, I can't locate the pitch limit for the Corsair's or Hellcat's propeller, but I believe it's about 65-inches. If DarrenW knows, it would be useful for comparison.

As it stands, we do have good evidence to suggest that the Frank's propeller probably gimped its performance.

 

[DarrenW]

However, I can't locate the pitch limit for the Corsair's or Hellcat's propeller, but I believe it's about 65-inches. If DarrenW knows, it would be useful for comparison.

Here's what the F6F-3/5 Hellcat pilot's manual states:

Most F4U-1Ds used the same propeller, as did some earlier F4U-1As.

 

[DarrenW]

But more or less, it's telling that the US only used electric props on lower horsepower engines.On 2,000 HP and up, they seemed to universally use hydraulic props

This seems the case until you look at the P-47 Thunderbolt which had factory-fitted CE propellers.These birds had engines with outputs well in excess of 2,000 hp.

 

[taly01]

I think its a false trail to believe the available prop angles were a problem, rather the Pe-32 propellor if a bit faulty was just slow to reach high speed pitch resulting in problems to reach top speeds, also as noted by RAF test in Singapore post-war vs Seafire and others.

Regarding the information on the prototype Ki-84, isn't that information also based on interviews with Nakajima personnel? Or did the Japanese retain at least some records?

Although most Nakajima factory data was said to be destroyed we have reliable KI-84 data from the captured Ki-84 manual, a IJAAF performance summary book of plane types, and data said to be provided by Nakajima to US TAIC.

******************************************************

"Ki-84 Pilot Manual" 3/Jan/1944

594kph @ 3370m +250mm 2900rpm
624kph @ 6550m +250mm 2900rpm

6:26 to 5000m +250mm 2900rpm

*****************************************************

"Ki-84 Performance and Drawing" from
Manuals for military aircraft created by Army Aviation Headquarters

584kph @ 2980m +350mm 3000rpm
631kph @ 6210m +350mm 3000rpm

5:54 to 5000m ?mm ?rpm

*****************************************************

Data provided by Nakajima to US TAIC

Ki84 "Hei?"

614kph @ 3700m +350mm 3000rpm
634kph @ 6550m +350mm 3000rpm

5:37 to 5000m +350mm 3000rpm

*******************************************************

I don't want to even think about Homares 21+ 23/24/25 or other developments!

 

[AerialTorpedoDude69]

This seems the case until you look at the P-47 Thunderbolt which had factory-fitted CE propellers.These birds had engines with outputs well in excess of 2,000 hp.

Good eyes, thank you!

I should have said, "with the exception of the Thunderbolt." The engineers compensate for its lower pitch angle by using an enormous propeller and as horsepower increases, so too does the size of that prop. They eventually get to 13 feet, which AFAIK is the largest prop on any fighter of the war.

But we all know the reason now: it's because in order to absorb the output of a high-horsepower engine, you need a combination of pitch angle and blade surface area.

I think its a false trail to believe the available prop angles were a problem, rather the Pe-32 propellor if a bit faulty was just slow to reach high speed pitch resulting in problems to reach top speeds, also as noted by RAF test in Singapore post-war vs Seafire and others.



Although most Nakajima factory data was said to be destroyed we have reliable KI-84 data from the captured Ki-84 manual, a IJAAF performance summary book of plane types, and data said to be provided by Nakajima to US TAIC.

******************************************************

"Ki-84 Pilot Manual" 3/Jan/1944

594kph @ 3370m +250mm 2900rpm
624kph @ 6550m +250mm 2900rpm

6:26 to 5000m +250mm 2900rpm

*****************************************************

"Ki-84 Performance and Drawing" from
Manuals for military aircraft created by Army Aviation Headquarters

584kph @ 2980m +350mm 3000rpm
631kph @ 6210m +350mm 3000rpm

5:54 to 5000m ?mm ?rpm

*****************************************************

Data provided by Nakajima to US TAIC

Ki84 "Hei?"

614kph @ 3700m +350mm 3000rpm
634kph @ 6550m +350mm 3000rpm

5:37 to 5000m +350mm 3000rpm

*******************************************************

I don't want to even think about Homares 21+ 23/24/25 or other developments!

taly01 it looks like there are three concerns here:

1. Are the captured manuals indicative of the Frank's performance throughout the war?
2. Why is prop pitch important?
3. How did hydraulic prop governors compare to electric ones?

First, do the 1944 Frank manuals accurately capture the performance of the Frank?

We don't know the details on the manuals' pedigree, although we might be able to assume they were from either China or the Philippines. Given the point at which they were captured, they seem to be for the -11 versions of the Homare and they do not include WEP values IIRC. So these aren't indicative of the performance of mid or late war Franks. It would be like describing a -4 Corsair using information on the F4U-1 of 1942. I'm not saying they're irrelevant but there are serious issues regarding sourcing, data reliability, and accuracy. Their values do line up with a lot of TAIC's data (which is partially based on those manuals) as well as synthetically generated numbers that experts like Greg have created.

Second, prop pitch limit is a crucial factor in the ability of a prop to convert torque into thrust.

The equation is mostly based on prop efficiency, surface area (or prop size), and pitch. Given the numbers that you gave me, there's no way the Frank's Pe-32 could convert 2,000 HP into thrust efficiently given its very low pitch and relatively small surface area. But I hope to one day get my hands on the Bunrindo issue that covered the Frank's Pe-32 because using that prop data we can generate the Frank's top speed as it seems we also know the zero-lift drag coefficient and horsepower.

Third, regarding hydraulic vs electric, there's a lot of points of comparison, but AFAIK, WW2 hydraulic systems were capable of hitting higher pitch angles (degrees) because hydraulic systems are capable of exerting more "force" with less power than electrical ones.

I think the reason is that as the pitch angle increases, drag on the blade increases as well, which requires more power to twist the blade. My guess is that early electrical motors didn't have the power to turn the blade and could likely adjust lower pitch angles more quickly but slowed as they reached their maximum pitch angle. But I could be wrong about faster adjustment of lower pitch angles for electric. However, I do know that while hydraulic systems weigh more, they're also capable of exerting more force. Electric on the other hand, is much faster but weaker. So for heavy loads at high pitch angles, you'd want hydraulic. At lower loads, you'd want electric.

This is relevant because at low RPMs, low pitch angles lead to faster acceleration. So an electric governor would be an ideal choice for low-speed dogfighting. So those advantages may be why Nakajima went with a low pitch angle governor.

So to recap, I think the only thing we can accept is that the early Pe-32 wasn't able to efficiently convert torque into thrust. As a corollary, that probably means the 687 KPH number isn't valid and could be substantially lower. However, we also don't know what late war Pe-32s looked like. It's entirely possible that they increased the pitch angles or prop efficiency because the engineering was constantly changing during the war. A 1944 governor was probably technically inferior to a 1945 model.

 

[qaz]

Hello, sorry for my lack of activity. A little busy and I forgot about the discussion for a while.
I finally left a comment on the Flight Model Review video now, in case the author wants to comment on this thread.

I'll post the values I have for the Ha-45-21 by the end of today, surely :V. I'm not sure they'll be very useful, but maybe someone with a bit more engine knowledge can try to understand the real performance in relation of alt, boost, and compression ratio.
-

I want to reiterate for the understanding of everyone that the engine performance issue is not only the power difference between Ha-45-12 & Ha-45-21, but also the different states of Ha-45-21: most mass-produced engines had reduced boost/rpm limits and/or reduced compression ratio to almost the same state as Ha-45-12.

I think Ha-45-21 was equipped to essentially every pre-production and mass produced Ki-84, but the US mistakenly called the types with restricted operating limits and reduced CR 'Ha-45-12' because they are almost the same. (AFAIK only the 3 initial prototypes certainly had Ha-45-12).


Sorry if I cause any confusion here, but I'm responding to several points from a few messages I missed:

Every Ki-84 has exhaust thrust. Both the early type of collective exhaust and later type of individual stacks produced thrust, but I presume the later type improved the thrust force to some extent.

To assume the TAIC numbers to be in the realm of reality, we would also have to establish that it used realistic drag, which doesn't seem to be the case. I really don't think we can treat these figures as even comparable in legitimacy to the more vague Japanese numbers.

I think it is indeed possible that Ki-84's real speed performance only jumped around 10km/h by changing from Ha-45-12 to 'fully-rated' Ha-45-21, especially if the true military power @ alt of the Ha-45-21 is ~1600, and/or the prop was inadequate for the additional power.
As mentioned there is another speed number floating around that correlates with the test document, even if it has some errors.

It would be useful to calculate a rough drag figure for the 624km/h test with Ha-45-12, and compare it with the drag figures given in the video derived from the 634km/h test. Using the same values that are estimated of course.

It should be possible using the rough drag figures to also extrapolate how much additional thrust/equivalent hp would be theoretically necessary to reach higher speeds like say, 645 or even 687km/h with said airframe.
It could also be useful for estimating what the speed would be with an arbitrary amount of extra exhaust thrust, for example.

 

[qaz]

Hi, here's the performance numbers I could quickly find from the documents I can verify.

Document 1: 「キ八十四」操縦法 (1/1944) Ki-84 Operation Manual
Army Document, Ha-45 (Ha-45-21)

Takeoff: 1850hp @ 3000rpm, +400mm

Nominal 1st Speed: 1860hp @ 2900rpm, +250mm 2300m
Nominal 2nd Speed: 1500hp @ 2900rpm, +250mm 6500m

[De-rated 21]


Document 2: 試製紫電改仮取扱説明書 (1/1944) Shiden Kai Provisional Operation Manual
Navy Document, Homare 21 (Ha-45-21)

Takeoff: 2000hp @ 3000rpm, +500mm
Nominal 1st Speed: 1900hp @ 3000rpm, +350mm 1750m
Nominal 2nd Speed: 1700hp @ 3000rpm, +350mm 6400m


Document 3: 海軍現用機性能要目一覧表 (8/1945) Navy Operational Planes Performance Catalog
Navy Document, Homare 21 (Ha-45-21)

Takeoff: 1990hp
Nominal 2nd Speed: 1625hp 6100m


Document 4: 海軍試作機性能要目一覧表 (8/1945) Navy Prototype Planes Performance Catalog
Navy Document, Homare 21 (Ha-45-21). Some various figures given.

1.
Takeoff: 1990hp
Nominal 2nd Speed: 1625hp 6100m
[Full-rated boost/rpm]

2.
Takeoff: 1820hp
Nominal 2nd Speed: 1500hp 6600m
[De-rated 21.]


Document 5: Specifications and Performance of Service Airplanes of the IJN (9/1945)
Navy Document, Homare 21 (Ha-45-21)
By the way, this is seemingly just a US rehash of the previous 2 documents with some of their own supplemental data.

Takeoff: 1990hp @ 3000rpm
Nominal 1st Speed: 1825hp @ 3000rpm 1750m
Nominal 2nd Speed: 1625hp @ 3000rpm 6100m


Document 6: TABLE OF OPERATIONAL ENGINES (9/1945)
Navy Document, Homare 21 (Ha-45-21)

Takeoff: 2,000hp @ 3,000rpm, +500mm

Nominal 1st Speed:
1790hp @ 3,000rpm, +350mm SEA LEVEL
1890hp @ 3,000rpm, +350mm 1800m
Nominal 2nd Speed
1420hp @ 3,000rpm, +350mm SEA LEVEL
1750hp @ 3,000rpm, +350mm 6450m

Compression Ratio: 7.0 [IMO this CR is erroneous, because this power is very high]


Document 7: Desc of Experimental Aircraft and Experimental Engines Under Development by the Japanese Army and the Imperial Japanese Navy (8/1946)
Army Document, Ha-45 (Ha-45-21)

Takeoff: 1,760hp @ 3,000rpm, +400mm

Nominal 1st Speed: 1860hp @ 3,000rpm, +250mm 2000m
Nominal 2nd Speed: 1780hp @ 3,000rpm, +250mm 6200m

[I think this document is mixing 'de-rated' conditions with 'full-rated' power numbers, except for takeoff power…]

----

Sorry I don't have much useful analysis to offer right now, but as we can see again, totally de-rated Ha-45-21 has 2nd speed nominal power around 1500hp (basically equivalent to Ha-45-12).

On the other hand, fully rated boost/rpm Ha-45-21 has 2nd speed nominal power ranging anywhere from lower 1600s to mid 1700s. Though CR change could partially explain this gap, personally my guess is that the upper extreme was the 'design objective', and I'm unsure if it was ever fully achieved even in the most ideal state...

 

[taly01]

It would be useful to calculate a rough drag figure for the 624km/h test with Ha-45-12, and compare it with the drag figures given in the video derived from the 634km/h test. Using the same values that are estimated of course.

This is probably a good idea to find drag of Ki84 from the test results we have, then to get an idea of HP needed for 427mph? TAIC. The Ki84 airframe although well streamlined and of minimal area, is not necessarily a good high speed design Mach drag etc.

Mach drag like propeller blade theory, is beyond my knowledge