How capable was the Ki-44?

[ThomasP]

F4U wing t/c ratio confusion arises from the way it is sometimes expressed. The actual numbers were 18% at the root, 15% at the inner/outer wing-fold joint, 9% at the tip. For some reason the 15% value is often the number used for the root in various publications.

The 18% & 9% values are from the F4U-1 and F4U-4 Detail Specification sheets. I do not know if this might be of use, but the same document also lists the average of the t/c ratio as 16% when divided by the 'frontal wing area'.

These same documents list the wing airfoil section as NACA 23018 tapering to 23009.

 

[GregP]

F4U used NACA 230 series.
18% t-t-c in the root, 9% on the tip.
web page

Yes, it used a 23018 at the root. That tapered to a 23015 at the gull wing break and then to a 23009 at the tip. Most of the airfoil with good free-stream airflow over it was 15% thick. Again, I doubt if Vought would allow the degradation of the CL Max to get as low as 1.38 overall, but I also don't have a wing tunnel and a Corsair to confirm same.

The F4U Corsair did not have a reputation as a bad turner. In fact, it was a very good dogfighter and pretty decent in roll, too. Sort of surprising for the hose-nose, huh? Overall, it was a stellar performer with respect to any real-world competition. I doubt a CL Max of 1.38 would result in that overriding impression but, again, I cannot positively verify that.

The book "Fighter Aircraft Performance of WW2 A Comparative Study" lists the CL Max for the F4U-1 as 1.691 on page 238. The author is Erik Pilawskii. It also lists the predicted turn radius at 4,000m as better than the Hellcat at a Reynolds Number of 10 million. Of course, that is predicted by an author. In real life, the F4U-1 was a good performer and rolled better than the Hellcat, but I've never seen comparative turn radii for the F4U and the F6F side-by-side anywhere elese other than in this book, making it a reference set of 1 book.

Not the best to argue with, for sure.

So, you could be correct, Tomo. I'll just say the subject is interesting to me and let it go at that.

Cheers.

 

[Shortround6]

What is often left out of these "CL Max of X.XX" lists or comparisons is what the angle of attack is, or more importantly, the drag at the angle of attack that gives the CL Max number.
The CL Max may tell you what the plane can sustain for a fraction of a second. I doubt anybody was flying full circles at CL Max even with a robot pilot.

Very few planes could fly at CL much over 1.2 to 1.4 for very long without flaps. Granted this is a generic chart. Maybe I am wrong. But it seems if you are generating high CL you are generating buckets of drag. Can your engine/propeller generate enough thrust to counter act the drag?

 

[GregP]

What is often left out of these "CL Max of X.XX" lists or comparisons is what the angle of attack is, or more importantly, the drag at the angle of attack that gives the CL Max number.
The CL Max may tell you what the plane can sustain for a fraction of a second. I doubt anybody was flying full circles at CL Max even with a robot pilot.

View attachment 851400
Very few planes could fly at CL much over 1.2 to 1.4 for very long without flaps. Granted this is a generic chart. Maybe I am wrong. But it seems if you are generating high CL you are generating buckets of drag. Can your engine/propeller generate enough thrust to counter act the drag?

WWII fighters didn't enjoy a great amounts of excess of power when it came to sustained turning. I believe most, possible none, could not sustain 6-g for a complete 360° circle, and many could not do so for a 180° turn. More important, they could not do it and maintain altitude at the same time, so most dogfights were descending affairs that stopped descending when the ground approached, as it often did over the Russian steppes ... especially since the Soviets were rarely willing to venture above their best power height and engaged in destroying enemy ground troops, which forced the Germans to come down and fight.

The entire point being you are quite correct Shortround6, best CL Max is usually associated with very high drag.

In today's very powerful jets, the CL Max is at what we call "corner speed" on the flight envelope. It is ALSO associated with high drag, but that speed is where the modern jets many times HAVE the power to sustain that turn.

But, you know that stuff as well as anyone. Good chart above.

 

[Irregualr23]

Yes, it used a 23018 at the root. That tapered to a 23015 at the gull wing break and then to a 23009 at the tip. Most of the airfoil with good free-stream airflow over it was 15% thick. Again, I doubt if Vought would allow the degradation of the CL Max to get as low as 1.38 overall, but I also don't have a wing tunnel and a Corsair to confirm same.

The F4U Corsair did not have a reputation as a bad turner. In fact, it was a very good dogfighter and pretty decent in roll, too. Sort of surprising for the hose-nose, huh? Overall, it was a stellar performer with respect to any real-world competition. I doubt a CL Max of 1.38 would result in that overriding impression but, again, I cannot positively verify that.

The book "Fighter Aircraft Performance of WW2 A Comparative Study" lists the CL Max for the F4U-1 as 1.691 on page 238. The author is Erik Pilawskii. It also lists the predicted turn radius at 4,000m as better than the Hellcat at a Reynolds Number of 10 million. Of course, that is predicted by an author. In real life, the F4U-1 was a good performer and rolled better than the Hellcat, but I've never seen comparative turn radii for the F4U and the F6F side-by-side anywhere elese other than in this book, making it a reference set of 1 book.

Not the best to argue with, for sure.

So, you could be correct, Tomo. I'll just say the subject is interesting to me and let it go at that.

Cheers.

Hi Mr. Greg

The airfoil is a 2-Dimensional stuff, which shouldn't be a representation for an airwing which is 3-dimensional, and all of these has to be account for conditions like Reynold's number ( Mean Aerial Chord Length) and airspeed. Usually the airwing and aircraft has significantly lower CLmax, L/D than the 2D airfoil series that been utilized on the aircraft.

Here's the Langley Full-scale Wind Tunnel tests on F4U-1 and F6F-3:
Airplane 6( F4U-1), at 60mph, no propeller, no flap:

Airplane 5(F6F-3), same condition:

The CLmax for F4U-1 in the test at service condition is around 1.17@16deg, for F6F-3 is 1.29@18deg. Note here the wing-tunnel test held at 60mph, which reduces the airplanes' CLmax because the airspeed was not at the optimal range(usually Mach 0.2-0.4). Also there's no propeller slipstream which further reduces the CLmax by around 0.1.

My figure of 1.38 for F4U-1 and 1.5 for F6F-3/5 were calculated from the trails held by NAS Pax. on production aircrafts of the two. But if you assume that 60mph airspeed reduces CLmax by 0.1 comparing optimal airspeed, and a loss of 0.1 due to no idling propeller slipstream, adding these two terms we have 1.37 for the F4U-1 and 1.49 for the F6F-3, which also proves the estimated power-off CLmax figure.

A comparison between CLmax of airplane 5 (F6F-3) and airplane 2 (P-63) with/without idling propeller, with flaps extended. An idling propeller gives around 0.1 increase in CLmax, and more with increased power setting.

The report attributes the higher CLmax of the F6F-3 to be that the Hellcat's wing is aerodynamically 'cleaner'. Consider that the F4U-1 has 2deg incidence angle, and F6F-3 has 3deg incidence angle, and the Corsair still stalls 2deg earlier than the Hellcat, which means the Corsair's wing stalls 3 degrees earlier than the Hellcat's wing. There may be some component that causes the early separation on the Corsair. I believe the wing-duct at its wing root was a reason, considering how the wing-duct on the P-63 affects its CLmax:

There's also a 1:2.75 scale F4U-1 wind-tunnel test of the CLmax, with propeller installed:

Though not a representation for a real F4U-1, but it shows the trend of its CLmax with/without flaps and power. You've mentioned that the CLmax to be 1.69 which could be correct in a power-on condition, where the strong slipstream will increase lift, also the vector of power has a portion lies on z-axis, and would be counted as 'Lift'.

Sincerely,
Irregular23

 

[Holtzauge]

My figure of 1.38 for F4U-1 and 1.5 for F6F-3/5 were calculated from the trails held by NAS Pax. on production aircrafts of the two.

A Clmax of 1,5 for the F6F-3/5 on a production aircraft sounds impressive. Since these were trials, how was the airspeed used to calculate the Clmax derived and what sort of corrections were applied?

 

[Steamed_Banana]

Hi Mr. Greg

The airfoil is a 2-Dimensional stuff, which shouldn't be a representation for an airwing which is 3-dimensional, and all of these has to be account for conditions like Reynold's number ( Mean Aerial Chord Length) and airspeed. Usually the airwing and aircraft has significantly lower CLmax, L/D than the 2D airfoil series that been utilized on the aircraft.

Here's the Langley Full-scale Wind Tunnel tests on F4U-1 and F6F-3:
Airplane 6( F4U-1), at 60mph, no propeller, no flap:
View attachment 851465
Airplane 5(F6F-3), same condition:
View attachment 851466

View attachment 851467View attachment 851468

The CLmax for F4U-1 in the test at service condition is around 1.17@16deg, for F6F-3 is 1.29@18deg. Note here the wing-tunnel test held at 60mph, which reduces the airplanes' CLmax because the airspeed was not at the optimal range(usually Mach 0.2-0.4). Also there's no propeller slipstream which further reduces the CLmax by around 0.1.

My figure of 1.38 for F4U-1 and 1.5 for F6F-3/5 were calculated from the trails held by NAS Pax. on production aircrafts of the two. But if you assume that 60mph airspeed reduces CLmax by 0.1 comparing optimal airspeed, and a loss of 0.1 due to no idling propeller slipstream, adding these two terms we have 1.37 for the F4U-1 and 1.49 for the F6F-3, which also proves the estimated power-off CLmax figure.

A comparison between CLmax of airplane 5 (F6F-3) and airplane 2 (P-63) with/without idling propeller, with flaps extended. An idling propeller gives around 0.1 increase in CLmax, and more with increased power setting.
View attachment 851470

The report attributes the higher CLmax of the F6F-3 to be that the Hellcat's wing is aerodynamically 'cleaner'. Consider that the F4U-1 has 2deg incidence angle, and F6F-3 has 3deg incidence angle, and the Corsair still stalls 2deg earlier than the Hellcat, which means the Corsair's wing stalls 3 degrees earlier than the Hellcat's wing. There may be some component that causes the early separation on the Corsair. I believe the wing-duct at its wing root was a reason, considering how the wing-duct on the P-63 affects its CLmax:
View attachment 851473

There's also a 1:2.75 scale F4U-1 wind-tunnel test of the CLmax, with propeller installed:
View attachment 851469
Though not a representation for a real F4U-1, but it shows the trend of its CLmax with/without flaps and power. You've mentioned that the CLmax to be 1.69 which could be correct in a power-on condition, where the strong slipstream will increase lift, also the vector of power has a portion lies on z-axis, and would be counted as 'Lift'.

Sincerely,
Irregular23

Wow that's interesting!

So forgive me, I'm still trying to get my head around CL Max. Can we say that CL Max is a constant varying on angle of attack? And does this equate to degrees of bank?

In other words, is CL Max in terms of lift efficiency more or less a constant or only really a measure of how far you can effectively push the turn angle?

 

[GregP]

Hi Mr. Greg

The airfoil is a 2-Dimensional stuff, which shouldn't be a representation for an airwing which is 3-dimensional, and all of these has to be account for conditions like Reynold's number ( Mean Aerial Chord Length) and airspeed. Usually the airwing and aircraft has significantly lower CLmax, L/D than the 2D airfoil series that been utilized on the aircraft.

Here's the Langley Full-scale Wind Tunnel tests on F4U-1 and F6F-3:
Airplane 6( F4U-1), at 60mph, no propeller, no flap:
View attachment 851465
Airplane 5(F6F-3), same condition:
View attachment 851466

View attachment 851467View attachment 851468

The CLmax for F4U-1 in the test at service condition is around 1.17@16deg, for F6F-3 is 1.29@18deg. Note here the wing-tunnel test held at 60mph, which reduces the airplanes' CLmax because the airspeed was not at the optimal range(usually Mach 0.2-0.4). Also there's no propeller slipstream which further reduces the CLmax by around 0.1.

My figure of 1.38 for F4U-1 and 1.5 for F6F-3/5 were calculated from the trails held by NAS Pax. on production aircrafts of the two. But if you assume that 60mph airspeed reduces CLmax by 0.1 comparing optimal airspeed, and a loss of 0.1 due to no idling propeller slipstream, adding these two terms we have 1.37 for the F4U-1 and 1.49 for the F6F-3, which also proves the estimated power-off CLmax figure.

A comparison between CLmax of airplane 5 (F6F-3) and airplane 2 (P-63) with/without idling propeller, with flaps extended. An idling propeller gives around 0.1 increase in CLmax, and more with increased power setting.
View attachment 851470

The report attributes the higher CLmax of the F6F-3 to be that the Hellcat's wing is aerodynamically 'cleaner'. Consider that the F4U-1 has 2deg incidence angle, and F6F-3 has 3deg incidence angle, and the Corsair still stalls 2deg earlier than the Hellcat, which means the Corsair's wing stalls 3 degrees earlier than the Hellcat's wing. There may be some component that causes the early separation on the Corsair. I believe the wing-duct at its wing root was a reason, considering how the wing-duct on the P-63 affects its CLmax:
View attachment 851473

There's also a 1:2.75 scale F4U-1 wind-tunnel test of the CLmax, with propeller installed:
View attachment 851469
Though not a representation for a real F4U-1, but it shows the trend of its CLmax with/without flaps and power. You've mentioned that the CLmax to be 1.69 which could be correct in a power-on condition, where the strong slipstream will increase lift, also the vector of power has a portion lies on z-axis, and would be counted as 'Lift'.

Sincerely,
Irregular23

Nice charts. Look at your first graph. It is operationally meaningless because it is a chart at 60 mph, where the aircraft would likely be completely stalled (assuming a WW2 fighter, anyway, not a Cessna 150) and falling from the sly, but let's look at it anyway.

At 16° angle if attack, CL Max is 1.26 or so. The Service wing;s CL Max is 1.17 or so. That is a change of 7.1%.

For the second chart (F6F), the change is only 3.2%.

Not bad for either one, though any chart at 60 mph is not really an operational consideration.

The last two charts are the best.

Thanks for the charts!

Cheers.

 

[Irregualr23]

A Clmax of 1,5 for the F6F-3/5 on a production aircraft sounds impressive. Since these were trials, how was the airspeed used to calculate the Clmax derived and what sort of corrections were applied?

I think NAS Patuxent River did very well on calibrating airspeed when testing the Hellcat and the Corsair. If you calculate CLmax according to the stall speed based on IAS listed on the manual, you would get over-stated CLmax since that IAS included pitot installation error (the one particularly affecting F6F showing lower IAS), and the stall may be counted when the wing starts to drop instead of the plane starts to lose altitude (1G stall).

Given values F6F-5 BUNO.58310
mass_kg = 5633*1 # kg
g = 9.81 # m/s^2
rho = 1.225 # kg/m^3
V_kph = 157.78 # km/h
V_f_kph = 136 # km/h
S = 31 # m^2
# Convert velocity to m/s
V = V_kph * (1000 / 3600) # m/s
V_f = V_f_kph * (1000 / 3600) # m/s
# Calculate weight (W = mass * g)
W = mass_kg * g # N
# Calculate lift coefficient (C_L)
C_L0 = (2 * W) / (rho * V**2 * S)
C_L1 = (2 * W) / (rho * V_f**2 * S)

CLmax of F6F-5 Buno 58310 (starboard pylon) power off clean: 1.515, power off landing: 2.039

Given values F6F-3 BUNO.25892
mass_kg = 5154*1 # kg
g = 9.81 # m/s^2
rho = 1.225 # kg/m^3
V_kph = 150.53 # km/h
V_f_kph = 80*1.61 # km/h
CLmax of F6F-3 Buno 25892 power off clean: 1.523, power off landing: 2.080

Hellcats in reduced condition:
F6F-3 BUNO.40164 with 2x20mm cannons: CLmax power off clean: 1.478, power off landing: 1.962 (cannon barrels reduce aerodynamic performance)
F6F-5 BUNO.72731(Full pylons, Rocket racks, airplane not very good condition as 5mph slower than Buno.78467 in same loading configuration):
power off clean: 1.443, power off landing: 1.941 ( reduced aerodynamic performance due to pylons, and reduced airframe&finish)

As for the Corsair:
F3A-1 Buno.04691
mass_kg = 5415*1 # kg
g = 9.81 # m/s^2
rho = 1.225 # kg/m^3
V_kph = 104.5*1.61 # km/h
V_f_kph = 86.5*1.61 # km/h
S = 29.17 # m^2
CLmax of F3A-1 Buno 04691 power off clean: 1.361 power off landing: 1.986

F4U-1 Buno.02155
mass_kg = 5077*1 # kg
g = 9.81 # m/s^2
rho = 1.225 # kg/m^3
V_kph = 101.5*1.61 # km/h
V_f_kph = 83.5*1.61 # km/h
S = 29.17 # m^2
CLmax of F4U-1 Buno.02155 power off clean: 1.353 power off landing: 1.999

F4U-4 Buno.80765 CLmax power off clean 1.471 power off landing 1.948 (The power-off clean here was tested when plane starts to roll so could be over-stated)

NACA values for F4U-4 still shows 1.37 for power off clean (they called Glide condition), though I think NACA does not doing as accurate as NAS Pax. on calibrating airspeed and measuring these values during test flights:

It is still possible that F4U-4 had imporved aerodynamic performance of the wing and thus giving it higher CLmax than the F4U-1, and possbily a reason it was listed to turn tighter than the F4U-1D even though the wing-loading is heavier:

Note the tests still involve a lot of uncertainties and errors, an 1-2mph error might incur some significant difference in calculated CLmax, thus I am listing various cases to compare. All stall speeds were calibrated to sea-level.

 

[Irregualr23]

Nice charts. Look at your first graph. It is operationally meaningless because it is a chart at 60 mph, where the aircraft would likely be completely stalled (assuming a WW2 fighter, anyway, not a Cessna 150) and falling from the sly, but let's look at it anyway.

At 16° angle if attack, CL Max is 1.26 or so. The Service wing;s CL Max is 1.17 or so. That is a change of 7.1%.

For the second chart (F6F), the change is only 3.2%.

Not bad for either one, though any chart at 60 mph is not really an operational consideration.

The last two charts are the best.

Thanks for the charts!

Cheers.

Which is exactly why to add 0.2 on both terms, 0.1 for not been at normal operating airspeed range, and 0.1 for missing propeller stream, and F4U-1 CLmax been = 1.17 + 0.2 = 1.37, F6F-3 CLmax been = 1.29 + 0.2 = 1.49.

Since the flow velocity term is included in the denominator of the CL formula so it doesn't really matter if the lift was less than aircraft's weight (stall), and should be well correlated with the service aircraft.

 

[Holtzauge]

I think NAS Patuxent River did very well on calibrating airspeed when testing the Hellcat and the Corsair. If you calculate CLmax according to the stall speed based on IAS listed on the manual, you would get over-stated CLmax since that IAS included pitot installation error (the one particularly affecting F6F showing lower IAS), and the stall may be counted when the wing starts to drop instead of the plane starts to lose altitude (1G stall).

Absolutely: If you have instrument errors when trying to determine the stall speed then your Clmax calculation will be off. And this is why I asked how they (NAS Pax) calibrated. So can you explain why you think "NAS Patuxent River did very well on calibrating airspeed when testing the Hellcat and the Corsair"? How did they do it?

NACA values for F4U-4 still shows 1.37 for power off clean (they called Glide condition), though I think NACA does not doing as accurate as NAS Pax. on calibrating airspeed and measuring these values during test flights:

So why do you think NACA was not as accurate as NAS Pax.? How did they (NACA) calibrate to determine the stall speed?

 

[GregP]

Which is exactly why to add 0.2 on both terms, 0.1 for not been at normal operating airspeed range, and 0.1 for missing propeller stream, and F4U-1 CLmax been = 1.17 + 0.2 = 1.37, F6F-3 CLmax been = 1.29 + 0.2 = 1.49.

Since the flow velocity term is included in the denominator of the CL formula so it doesn't really matter if the lift was less than aircraft's weight (stall), and should be well correlated with the service aircraft.

The easier way would be to run the wind tunel up to operating speed. And, it would be more accurate, too.

Fudge factors like your 0.1 increments, aren't my favorite thing to use for conclusions about real world aircraft when there are facilities to use that can do it at the actual conditions.

 

[Irregualr23]

Absolutely: If you have instrument errors when trying to determine the stall speed then your Clmax calculation will be off. And this is why I asked how they (NAS Pax) calibrated. So can you explain why you think "NAS Patuxent River did very well on calibrating airspeed when testing the Hellcat and the Corsair"? How did they do it?



So why do you think NACA was not as accurate as NAS Pax.? How did they (NACA) calibrate to determine the stall speed?

I've reviewed the stalling test held by NAS Pax. and NACA and indeed I found no mentioning on the non-linear pitot error calibration at near stall speed. NAS Pax. only provides calibration curve above 120MPH so it is not-known the calibration curve near the stall speed, especially that was a non-linear region. However, one can tell that they did calibrated the airspeed at near the stall speed, the stall speed figure from the Hellcat test is clearly higher than the stalling IAS listed in the manual, and most of the calculation using the tested data fall in a relatively good region as I have posted previously. This is not the case for most of the AAF figures on the stall speed(p-47,p-40 and p-51), indicating that the installation error was a major factor, the exception would be the P-38 which has a clear test report on stalling speed and airspeed calibration near the stall point.

When I was going through your previous discussion on CLmax for the P-40 and P-47, I found this British report been quoted and it contains a clear measurement for the CLmax of F6F-3, with specific calibration curve:

The report gives a power-off clean CLmax 1.64 and a landing CLmax 2.06 for the Hellcat Mk.I, though from the chart the specific power-off clean CLmax point was actually around 1.52. This may be resulted by different definition for "power-off", whether the throttle is fully closed or to remain for a few inches of manifold pressure, which would affect the propeller slipstream effect. This test report had cross validated the CLmax figure I've calculated from the NAS Pax. tests and also proves their calibration during the test.

I have biased toward the NACA test reports due to frequent typo involved in the report, but now I found indeed I have no other evidence to tell that NACA made the wrong calibration.
So I've also went through the stall tests for F6F-3 and F4U-4 again and found them to reflect the conclusion:

1.5 seconds of steady-state(no angular movements on all axes) CLmax for the F6F-3 power-off clean, around 1.5:

2 seconds of steady-state(steady angular movements on roll and yaw with some angular velocity on pitching because of the turning condition) CLmax for the F4U-4 power-off clean, around 1.4:

Though the tests for the F4U-4 still gives relatively lower CLmax for conditions with flaps deployed. I dont know if NACA did it correctly.

It seems that the higher taper ratio of around 0.7 for the F4U and thicker t/c were also reasons for the lower CLmax of the F4U in comparison with the F6F.

It looks like the F6F had achieved what a relatively clean NACA23000 no-twist tapered wing should have achieved at higher Reynold's number. The question remained came from the discussions that you've participated years ago about the CLmax for P-47 and P-40.

1. It looks like the P-38 and P-40 also features an exceptionally high CLmax around 1.5, the case on the P-38 can be explained for its huge AR=8, but what for the P-40 then? Sure it uses an un-twisted wing, but with NACA 2200 airfoil and shorter MAC it shouldn't achieve anything that high, I would expect the number to be around 1.4. The P-40 related tests about its stall speed remained questionable, as the instrumental error near the stall speed came from a linear extrapolation.

2. The CLmax for the P-47 looks not promising, calculated from its stall speed figure only featured a CLmax around 1.29 to 1.33, which is quite low considering its MAC and relatively high Reynolds number, then how came the Republic choose to use the S-3 airfoil?

3. It looks like many WW2 fighter aircrafts were not designed to achieve an ideal CLmax figure, for example the Japanese would rather gave specific requirements on low wing-loading instead of higher CLmax, despite the latter seems not that hard to achieve. What are facts that to be concerned about it? Or should we consider the methodology for designing ideal CLmax for a typical wing was not matured due to the lack of high-speed wind tunnel and CFD tools during that era?

Happy to disccuss,
Irregualr23

 

[Steamed_Banana]

Wow, very informative. I think I'll have to read this about four times before I fully understand it. What is AR again?

 

[Irregualr23]

Wow, very informative. I think I'll have to read this about four times before I fully understand it. What is AR again?

AR means Aspect Ratio, which is a basic factor in wing planform design.

 

[Irregualr23]

Absolutely: If you have instrument errors when trying to determine the stall speed then your Clmax calculation will be off. And this is why I asked how they (NAS Pax) calibrated. So can you explain why you think "NAS Patuxent River did very well on calibrating airspeed when testing the Hellcat and the Corsair"? How did they do it?



So why do you think NACA was not as accurate as NAS Pax.? How did they (NACA) calibrate to determine the stall speed?

Another thing to be considered is that Re effect has diminishing return at high Re numbers, and high diminishing return for planes with smaller MAC to work at optimal speed. So taking CLmax at low stall speed may not be accurate and will heavily biased toward planes sensitive to Re effect, and it can be a possible reason for contradictory reports on lift performance of P-47, P-51 and FW-190.

 

[GregP]

All this with the Corsair and the Hellcat seems pretty funny if you remember Corky Meyer.

One of the Grumman test pilots was Corky Meyer. He did a series of articles on the Hellcat and Corsair some time back, like maybe 25 years ago, and he said that Grumman was given a Corsair to test for some period of time and Vought was given a Hellcat for the same period. He did comparative test with them, side-by-side, with another test pilot. Both airplanes had the same engine and propeller at the time. Males me think the planes were an F6F-3 and an F4U-1a, but that's my supposition.

Unsurprisingly, they both flew at the same speed at the same power setting and rpm. If flown side by side and the throttle was opened up simultaneously, they accelerated almost side by side and one or the other would slowly pull away slightly. It wasn't always the same aircraft that pulled away into the lead.

What Corky, and every other pilot who was along for the test, noticed was that the Corsair always indicated about 15 knots faster than the Hellcat, even when flying side by side in formation.

I have never had occasion to ask a modern warbird pilot about that, but it seems like a good thing to ask about given the discussion.

No real point here, just an interesting thought from a comprehensive article about wartime test flying some years past.

Cheers.

 

[Irregualr23]

All this with the Corsair and the Hellcat seems pretty funny if you remember Corky Meyer.

One of the Grumman test pilots was Corky Meyer. He did a series of articles on the Hellcat and Corsair some time back, like maybe 25 years ago, and he said that Grumman was given a Corsair to test for some period of time and Vought was given a Hellcat for the same period. He did comparative test with them, side-by-side, with another test pilot. Both airplanes had the same engine and propeller at the time. Males me think the planes were an F6F-3 and an F4U-1a, but that's my supposition.

Unsurprisingly, they both flew at the same speed at the same power setting and rpm. If flown side by side and the throttle was opened up simultaneously, they accelerated almost side by side and one or the other would slowly pull away slightly. It wasn't always the same aircraft that pulled away into the lead.

What Corky, and every other pilot who was along for the test, noticed was that the Corsair always indicated about 15 knots faster than the Hellcat, even when flying side by side in formation.

I have never had occasion to ask a modern warbird pilot about that, but it seems like a good thing to ask about given the discussion.

No real point here, just an interesting thought from a comprehensive article about wartime test flying some years past.

Cheers.

Standard F6F-3 pitot orifice does indicate airspeed some 10 knots lower at most of the speed range, but at those range, the instrument error is linear. At stall speed, the error would be non-linear, that's where you see AAF test memorandum recorded a stall speed 60mph IAS for the Hellcat, and some claims the plane stalls at 20 knots, and possibiliy 0 knots( due to sideslip and flap setting).

Let's turn to the Hellcat then cuz clearly I am also a fan of Hellcat.

Regarding to the speed difference between the F6F and F4U. The BuAer documents and conference notes listed the speed of F6F-3 to be 392mph TAS and early F4U-1 to be 398mph, only 6 mph difference. From the NAS Pax. test, a standard production F6F-3 usually has an airspeed-power curve 1800BHP : 310mph at sea level, while the F4U-1a to be 1800bhp : 325mph. Around 15mph faster for the Corsair in the same power output. The Grumman may have the F6F in their test to be in higher than standard condition, and the F4U-1 to be standard or below standard conditon. The F4U-1 had bad quality control before FG-1D, the Corsair also frequently got design upgrade, and there is also widely seen carburettor problem causing a derated power even at the same RPM and MAP setting. ( The reason you often see that de-rates the military power from 2000bhp to 1920bhp and even 1850bhp on production F6F, F4U and P61 that use R-2800-B engine). So it was possible for Meyer to find his Hellcat to be the same as the Corsair in speed, while other tests and reports often says the Corsair is somewhat faster.

Also I have researched the Neutral stage problem for the F6F which has been discussed for decades, and I've reached out the NASM archive and they shared me a report on the Neutral stage test that solved the mystery :

Starting Buno.78467 the F6F got rammed air for its Neutral blower, the main stage air inlets were moved inside intercooler inlets, and there is an intercooler blocking valve that can block the air from entering the intercooler and directing stream into the main-stage when the engine is running at the Neutral stage to give the engine maximum ram-air effect to generate 2345bhp at sea-level. So starting early 1945, the last 2000-ish Hellcats got ram-air effect for its N stage, thoguh most of them may not get the blocking valve system since that may increase the risk of a malfunctioning valve blocking the air when the engine is running at Aux blower and needs intercooler, so 2215-2250bhp part-ram can be plausible for late F6F-5 to achieve at sea-level.

The speed test showed 2165bhp/332mph, 2215bhp/335mph and 2345bhp/341.5mph @SL the plane carried all pylons and HVAR launchers during the test, and for a clean F6F-5 the speed should be 7-10mph faster, and a 350mph top speed at sea level is achievable for a Hellcat. (Speed plots from 1944 ACP and Grumman report 2422C showed it was achievable even without rammed air, possibly not representing the burnishing condition of a production aircraft).

Also, even without this modification, the earliest F6F-3 can also get some slight "rammed-air effect", as been stated in the AAF memorandum report the F6F-3 can have a half inch increase in MAP by closing all cowl flaps and inter-cooler flaps at Neutral stage.

Hopefully this would help.

Sincerely,
Irregular23

 

[Holtzauge]

I Irregualr23 : Very interesting stuff and thanks for sharing all the information in this thread. The reason I asked you if you had any info on what type of calibration was done both in the case of NAS Pax and NACA data, is that it has such a large bearing on the results.

In fact, there is a NACA report about the Spitfire in which the NACA engineers who wrote it came to the conclusion that the Spitfire Mk V they tested had a Clmax of only about 1.2, and even theorized that the reason for this low value was because it had such benign stall characteristics. In the report, the NACA engineers say that calibration was done by flying in formation with an aircraft with "known calibration" properties and that this was the correction used to derive the Clmax=1.2.

Now when this came to the attention of the RAE in Britain, one of their leading aeronautical engineers, M B Morgan, wrote a note which was sent to NACA where he was very civil but pointed out that the British had already done similar trials, but using a trailing pitot system, and based on these derived a Clmax of 1.36, and implied that maybe flying in formation was not the best way to get accurate results?

I draw two lessons from this anecdote: The first is that even the sun has its spots, and sometimes even peer reviewed documents by reputable institutions can contain erroneous information. The second is that I am nowadays quite vary of any Clmax figures unless it is totally clear how they were derived, and so far I'm only inclined to trust trailing pitot results, or those from full-scale wind tunnel tests, like in NACA's, or in case of the Bf 109, the Chalais Meudon trials.

However, even so, regarding wind tunnel tests, I fully agree with what you wrote earlier about Re effects since these tunnels even though whole aircraft fit in, usually do not have a tunnel wind speed that corresponds IRL and will thus be somewhat on the low side. In noticed you mentioned adding about 0.1 to results like this to account for the Re effects, and to do a ballpark adjustment, I think that's a fair assessment.

In addition, I agree about the propeller slipstream effects: And while we both know that the power on Clmax is dramatically higher (IIRC then RAE estimated about 1.89 for the Spitfire), your point about the effects even when idling (I think you said about 0.1 there as well?) is also an interesting observation adding further to the problem to pin down an exact Clmax, since do we mean a Clamx as one with no prop or a Clmax when idling?

Further comes the difficulty of definitions: Exactly when did they read off the stall speed? When the aircraft started to drop? And how exact was defined? Or did they read it off when the aircraft started to get disturbances in roll etc.? Some reports detail that, but some just say "Clmax at stall" leaving the reader hanging. Sometimes that can be seen in the reports like you posted, but more often than not it's not totally clear.

And another important point: Is the value determined from one single test or have multiple measurements been done? For the RAE trailing pitot tests this is not always clear, but for the German Chalais Meudon tests of the Bf 109 the value they determined there was an average of about 10 runs IIRC.

Then about calibration charts: As you point out what you find in flight manuals (PEC's) usually stops far short of the stall speed, which I think is for the simple reason that these are for navigational purposes only, and when it comes to stall, the only thing the pilot needs to know is what he reads off his airspeed indictor. However, extrapolating from the PEC curve, this on the Spitfire looks to be about 15-20 mph lower than if it was calibrated. In addition, and as I see you have already pointed out, calibration in this regime is not linear, to which I think we could add maybe afflicted by large fluctuations, so I would be very vary of any Clmax values derived from calibration of an aircraft's airspeed IAS as for the Hellcat, especially if we have no idea about how the calibration was done.

 

[GregP]

If the link works, here's Corky's article.

PressReader.com - Digital Newspaper & Magazine Subscriptions