XP-39 Wind Tunnel Tuft Tests

kool kitty89 · Apr 20, 2016

View: https://www.youtube.com/watch?v=vKxe8jc9_rI

I'm surprised I've never stumbled on this video before, it's amazingly informative. The few photos of the XP-39 floating around and limited information (often misinformation) regarding its layout and turbocharger (and radiator) installation don't do it justice at all. (in fact, none seem to show the starboard side of the aircraft at all)

It might just be me, but the entire cooling system installation (oil cooler, intercooler, and to at least some extent coolant radiator) look shockingly sloppy and patched together ... almost taped-on. (except the coolant radiator which just seems awkwardly placed, or at least its exit duct mounted mid-wing towards the trailing edge ... below, and possible ahead of the intercooler scoop and turbo air intake) No boundary layer splitter/sleeves/cowls either, but that's not really surprising compared to the rest.

I had an inkling that the overall installation on the XP-39 was so bad as to be totally non-indicative of a practical turbocharger installation (even a reasonably acceptable draggy one with external intercooler -like the P-38J), but I didn't think it was this bad.

Additionally, note the comment about the wing stalling all at once (or almost simultaneously), though I'm not sure if this is relevant to the production P-39. (stall warning is supposed to be good, it's just the spin characteristics that are poor, at least when loaded tail-heavy or out of ammunition -the NACA-recommended 13 inch rear-fuselage extension probably didn't help, let alone the load of equipment crammed behind the engine)

Also interesting that the XP-39 has a nicer clear-view canopy, both taller and lacking the thick 'anti-roll bulkhead' behind the pilot. (an odd addition, I don't think any other WWII aircraft implemented roll-bars, and if they did they were omited in the bubble canopy redesign. (granted, this is a feature I noticed long ago with the more common pictures of the XP-39, though I only realized the full function of that bulkhead recently)

See:
http://img525.imageshack.us/img525/2236/bellairacobrai1939airen.jpg

(also note the small 60-round 20 mm ammunition drum, and the holes in the forward nose bulkhead for the omitted pair of 30 cal nose guns present on the YP-39 and P-39C, though there looks like enough space for another pair of .50 cal guns, but perhaps not ammunition boxes)

Edit:
on this issue (cut from earlier in my post)
This actually makes the quoted speed of the XP-39 with turbocharger rather more impressive with all that drag holding it back, admittedly with a lot less weight as well. (also makes me slightly more dubious over the 'excessive drag' created by attempted saddle-mounted turbo installations and other experimental fits tried ... unless those were all only on paper, so less attempted and more anticipated)

I see Shortround addressed this bit of misinformation as well a few years back:

Shortround6 said:
Tomo, this has been gone over in other threads, but basically, The XP-39 did not meet the performance numbers promised.

There is no real evidence that it ever came close to the numbers given for in many books/web sites. And there is evidence that it couldn't have done what is claimed during the time period it is claimed it flew those numbers.
1. There was a potential problem with drive-shaft vibration that called for a redesigned heavier drive shaft to be fitted, This was not done until after the NACA wind tunnel tests and until fitted the engine was restricted to 2600rpm. No where near full power.
2. XP-39 was plagued with both oil and coolant overheating problems which called for redesign of both the oil cooler and radiator ducts.
3. General Arnold was making arrangements less than a month after it's flight to have the XP-39 put into the full size wind tunnel at Langley.
4. Contract weight was 5,550lbs, when weighed at Wright field during "expedited" acceptance trials it weighed 6,104lbs. about 10% over weight.

Many people claim that army generals and the NACA "ruined" the P-39. They never go into any details except to say they removed the turboPlease note the above was before the NACA got their hands on the P-39.

In their report they (the NACA) claim the XP-39, as they received it was good for 340mph at 20,000ft and just under 280mph at sea level.
There were severe problems with the airflow for the radiator, oil cooler and intercooler.
For the last, US practice of the time was that the intercooler should remove 1/2 of the heat added by the turbo-supercharger to the intake air before it entered the engine carburetor. In the XP-39 the NACA estimated the intercooler (based on airflows) was removing only 25% in high speed flight and about 12% during climb. This significantly affected power output. The inter cooler was small in size to fit the airframe and had a high pressure drop across it. this meant high drag. The XP-39 needed a much larger intercooler to perform properly and this larger intercooler would not fit in the airframe.
Bell did at least two mock-ups of turbo/intercooler units in 1941 to be 'added' to the P-39. the extra drag of these units caused 40-45mph speed loss at the lower altitudes.

340 mph at 20,000 ft I would buy with that 'streamlining' on the XP-39.

Though in that same thread:
XP-39: pros cons

And on the XP-39 the oil cooler and radiator were in the wing root area ahead of the wheel, but to have good airflow you also need an exhaust duct behind the the radiators/coolers.

Shortround got this bit wrong, given the oil cooler is clearly in a bulky external box on the XP-39 (an understandable mistake given the confusing information floating around, and that video being the only source showing the oil cooler at all)

The XP-39 used the wing center-section for the turbocharger, with the radiator in the port wing center section (outlet in trailing edge). The XP-39A (and all later models) appear to have moved the radiator to the center section (port side) and the oil cooler(s) alongside it (starboard side) with the intakes for each in the corresponding wing leading edge.

On a side note, scanning the remainder of that thread's discussion: to increase fuel capacity, the best option seems to be deleting the wing armament and adding fuel cells to the outer wing compartments (that would otherwise hold the guns and magazines) and possibly in the wing leading edge. (the portion ahead of the gun bay is already clear to accommodate the blast tubes, but inboard of that may not be, and much further outboard gets too tight for useful fuel compartments very fast, so a pair of small leading-edge reserve tanks seems mostly likely) Beyond that you'd have to resort to tip-tanks, which would've been neat (and likely helped roll rate and possibly even stall/spin -due to wing-fence/winglet-like behavior) but the tip-tank arrangement seems to be too unusual and counter-intuitive for the time to be an obvious solution. (or sheer coincidence that no one stumbled on that prior to the XP-80A)

Oh, and retaining the .30 cal guns in the nose should have improved CoG a bit too. (.50s would've been better on all accounts if they could fit them ... even if some cheek bulges were needed ... maybe they could add ballast to the spinner, but useful 'ballast' would obviously be preferred)

kool kitty89 · Apr 20, 2016

Thinking on the 'wing stalls almost simultaneously' aspect, I realized wind tunnel tests of this sort simulate glide conditions, thus omitting the impact of prop-wash. The accelerated air over the wing center section would thus tend to make the tips stall much earlier than the inboard portion of the wing (especially at low speeds where prop wash has the most impact on overall airflow -transonic prop-tip shockwave situations aside).

Again, this might not have bearing on the stall/spin characteristics of the production P-39, but it's at least interesting to consider. (I still think partial-span leading edge slots are a good compromise for complexity/weight/drag in improving aileron/tip stall characteristics -like the Hudson, N9M, or Me 163 used- without the difficulties of slats -same goes for compromising on non-functional slats like the Whirlwind had)

Wiki actually notes that particular type of partial-span slot implementation:
Leading edge slot - Wikipedia, the free encyclopedia (though the list of examples is a bit off, I don't think the Beaufort -or any Blenheim derived airframes- used leading edge slots operationally).

The hudson seems to avoid complexity further by cutting slots between wing ribs rather than making a continuous slot.
http://www.cybermodeler.com/hobby/builds/ca/images/ca_0448_26.jpg
https://s-media-cache-ak0.pinimg.com/736x/2f/9d/be/2f9dbe761332155a40167e493e00b822.jpg

Shortround6 · Apr 20, 2016

Quite a few aircraft used "letter box" slots but they caused more drag than the retracting slats when the slats were retracted. It was a balancing act to get the aileron response at low speed desired without causing too much drag. Slats/slots were almost de rigueur for a few years just before WW II. It depended a bit on who had the patents. The first 50 Halifaxes had slats. The XB-24 had slots. I am not sure how long the B-24 series kept them. The XB-35 flying wing had them. Grumman Avengers had them. Curtiss Helldivers had slats.

Shortround6 · Apr 21, 2016

I have looked at a P-39 book by Birch Matthews which has several photos of the XP-39 showing the right side. the oil cooler is in a duct on the fuselage about the same place as that "box" in the video but is much smaller and more of a half tube (round instead of square).
Given the XP-39s rather alarming tendency to overheat even in ground running and taxiing they were doing some rather mad scrambling to solve the cooling problem before the plane was shipped to Langley. This cooling duct could very well be a hasty improvisation. there is mention in the text of several different ducts being tried in order to solve the oil cooling problem.

GregP · May 2, 2016

The Allison uses the engine oil for a significant amount of cooling. The Merlin, on the other hand, doesn't. As a result, most Allison-powered aircraft will have two oil coolers for an Allison and one radiator ... and a Merlin-powered aircraft will sometimes have two radiators, but only usually has one oil cooler.

The front intake of the P-40, take an N model for instance has two oil coolers up high in the intake and one radiator in the lower center. The P-38 is the same way today but, back when they actually used the turbochargers, the late-model intake had two oil coolers and the center was the intake for the intercooler. Today there are two P-38s that occasionally fly WITH the turbos, but the rest plug up the outboard turbo intakes and use the lower center opening as the carburetor air intake.

Edit: You are right, it's 2 coolant radiators and an oil cooler in the P-40N. Very similar shape, but coolant. And thanks for making me look closer! It also happens to share some parts with the P-38 and P-39 lie starters and cuno filters. Many Allisons shared the same carburetors. If I am not mistaken, many shared magnetos, too. It was probably GFE - Government Furnished Equipment. End of Edit.

Many Merlin planes, such as the Spitfire, split the coolant radiator into two small under-wing units.

Coolant is the Merlin's lifeblood. It doesn't run long without it. If you lose oil cooling, you can throttle back and get home if it isn't too far away. With the Allison, you need the radiator AND the oil coolers. if you loose oil cooling, you'll get a less damaged range than the Merlin and will have to shut it down or burn. The upside is if you loose the coolant radiator, it will fly a LOT longer than Merlin with the same fault at reduced power because much cooling in the Allison comes from the oil.

I don't believe either one is inherently superior to the other ... just different solutions to cooling a big V-12.

tomo pauk · May 3, 2016

Greg - a good overwiev, thanks.
BTW, Spitfire, at least service machines, never had two coolant radiators.

wuzak · May 5, 2016

tomo pauk said:
Greg - a good overwiev, thanks.
BTW, Spitfire, at least service machines, never had two coolant radiators.

The two stage Merlins had two radiators:
[qoute]Because the intercooler required a radiator, the radiator under the starboard wing was halved in size and the intercooler radiator housed alongside. Under the port wing a new radiator fairing housed a square oil cooler alongside the other half-radiator unit. When the engine was running at low speed, one radiator section provided enough coolant; a thermostatic switch turned off the starboard radiator section until more power was called for and extra engine cooling was required.[/quote

Supermarine Spitfire (late Merlin-powered variants) - Wikipedia, the free encyclopedia

[qoute]The Mk. IX radiators had enlarged frontal area as compared to the earlier Spitfire marks. Both radiators were identical rather than being mirror images of each other, divided into two sections - the starboard being an oil cooler, port side being occupied by the intercooler.[/quote]

Supermarine Spitfire Mk. IX in Detail -The Powerplant

Certainly single stage Merlin Spitfires and the Mk XII had only one coolant radiator.

In any case, P-38s had two radiators per engine.

http://i.imgur.com/W9W2rs4.jpg

tomo pauk · May 5, 2016

Indeed, both you and Greg are right re. Spitfire radiators. Aplologies.

kool kitty89 · May 5, 2016

Shortround6 said:
Quite a few aircraft used "letter box" slots but they caused more drag than the retracting slats when the slats were retracted. It was a balancing act to get the aileron response at low speed desired without causing too much drag. Slats/slots were almost de rigueur for a few years just before WW II. It depended a bit on who had the patents. The first 50 Halifaxes had slats. The XB-24 had slots. I am not sure how long the B-24 series kept them. The XB-35 flying wing had them. Grumman Avengers had them. Curtiss Helldivers had slats.

Retracting slats seemed to be tricky to engineer in a reliable, lightweight, compact, structurally sound manner, so (patents aside) had a lot more difficulties in implementing than fixed slots. Messerschmitt got it right pre-war and continued to do so during the war on multiple aircraft, but that was rare.

Fixed slots seem like a reasonable solution for aircraft where wing-tip stalling behavior is dangerously unacceptable either at low speed (take-off and landing) or general maneuvering (high-speed stalls and spins). I believe this was the primary reason the Me 163 employed them given the dangerous of stall characteristics on tailless aircraft. (I'm not positive, but I believe some of the YP-49's dangerous, structurally compromising high-speed stall behavior was related to similar issues; the YP-49 had letterbox style slots but they appear to have had flaps/fairings that closed during normal flight and were not aerodynamically operated like slats)

Given the high-speed nature of the Me 163, I imagine fixed slots were used as the added drag they presented was acceptable and worth the weight and simplicity gains, especially for a wooden wing. (plus only partial-span in spite of lack of prop wash, so likely more useful for improving stall characteristics than actually improving stall speeds)

Wing-tip slots do allow for other drag-cutting measures like using lower drag (lower lift) airfoil profiles, eliminating span-wise wash-out (twist) normally used for improving stall characteristics and possibly even eliminating dihedral (though this has some other stability trade-offs). In the P-39's case, and its use of NACA 0015 airfoil at the root (relatively low-lift, low drag symmetrical airfoil), using that 00xx profile for the entire span might have been a reasonable option for reducing drag in conjunction with adding the slots.

In the case of something like the F4U, it seems more like the critical need for carrier landing characteristics would have merited any added drag wing-tip letterbox slots would add. (particularly as it would still have been lower drag than the F6F, and available sooner)

GregP said:
The Allison uses the engine oil for a significant amount of cooling. The Merlin, on the other hand, doesn't. As a result, most Allison-powered aircraft will have two oil coolers for an Allison and one radiator ... and a Merlin-powered aircraft will sometimes have two radiators, but only usually has one oil cooler.

The front intake of the P-40, take an N model for instance has two oil coolers up high in the intake and one radiator in the lower center. The P-38 is the same way today but, back when they actually used the turbochargers, the late-model intake had two oil coolers and the center was the intake for the intercooler. Today there are two P-38s that occasionally fly WITH the turbos, but the rest plug up the outboard turbo intakes and use the lower center opening as the carburetor air intake.

You've mentioned this before, but as I recall in the last time it came up, someone pointed to the outboard radiators on the P-40 being glycol coolers and the central one being the oil cooler. (I forget who provided the reference, but it confirmed the labels on various cut-away examples of the P-40)

There might be some misinformation going on elsewhere, though, and the merlin powered P-40s may have swapped the radiator and oil cooler positions, among other things, leading to some confusion. That or the P-40N uses a different arrangement than earlier models.

http://img43.imageshack.us/img43/2537/curtissp40ekittyhawk1we.jpg

That aside, the large oil cooling capacity does at least mesh with the similarly sized oil cooler (port) and prestone/glycol radiator (starboard) intakes in the wing roots

tomo pauk · May 5, 2016

Serial-produced P-40, whether with V1710 or with V-1650, were with 2 coolant radiators:
Allison V-1710 Installation in P-40

kool kitty89 · May 5, 2016

Do note the size of the oil tank vs coolant tank depicted there, though. The large oil capacity might be significant.

Shortround6 · May 5, 2016

Oil was a consumable. Coolant was not. Engines were rated not only by fuel used per HP/hr but by oil used per HP/hr. Long flights required more oil than short ones, see larger oil tanks fitted to Spitfires for ferry flights for example.
Coolant was supposed to stay in the cooling system and not be used up at the rate of a gallon an hour or more.

wuzak · May 5, 2016

tomo pauk said:
Serial-produced P-40, whether with V1710 or with V-1650, were with 2 coolant radiators:
Allison V-1710 Installation in P-40

Thanks for that picture Tomo.

I have sometimes wondered if, given their size, the radiators could have been mounted on the leading edge of the wing - perhaps near the root or at the front of the landing gear fairing. Then they would only need a small duct for the oil radiator and the nose could have been sleeker.

wuzak · May 5, 2016

The P-40F appears to have had only a single radiator.

Looks a little bigger than the paired Allison radiators

http://fsfiles.org/flightsimshotsv2/images/2015/04/18/aFlyingHeritageCollection113.jpg

tomo pauk · May 6, 2016

Indeed, the P-40 have had only one radiator, rather big (from Service manual for the P-40F):

wuzak said:
...
I have sometimes wondered if, given their size, the radiators could have been mounted on the leading edge of the wing - perhaps near the root or at the front of the landing gear fairing. Then they would only need a small duct for the oil radiator and the nose could have been sleeker.

I like the proposal. Or, similar like it was done on the P-40Q, though it seems like the oil coolers went at the wings, between the wheel struts and guns. Now where is my pic of such modified 'plain' P-40, all together with clipped wings...

kool kitty89 · May 8, 2016

tomo pauk said:
I like the proposal. Or, similar like it was done on the P-40Q, though it seems like the oil coolers went at the wings, between the wheel struts and guns. Now where is my pic of such modified 'plain' P-40, all together with clipped wings...

Take a look at the XP-40K

http://img.wp.scn.ru/camms/ar/375/pics/3_103_b1.jpg

Military Aviation Archives - P-40 Warhawk Historical Development Photo Set
http://www.milavnarc.com/photos/wor...cal_development_photo_set/large/22 XP-40K.jpg

Radiators and oil coolers installed in a flat conformal arrangement between the landing gear strut bulges and look like they might be placed to avoid reducing the center section fuel tank capacity.

I haven't seen any information on performance, but it looks rather nice at least. (extending the intake ducts to a unified scoop below the nose might have made for better ram effect and reduced drag relative to mass flow through the cooling matrix -and better cooling on the ground and at low speeds due to prop ram effect)

Come to think of it, the P-39 could probably have employed fuel tankage in the center section if (somewhat draggier) semi-embedded lower-fuselage radiators had been used instead, perhaps somewhat spitfire or 109F/G-like but set below the wing roots, inboard of the gear and outboard of the belly rack (rather than in the wings). That would also free up the leading edge wing root space for fuel tanks (a rather prime location given the thick ring roots and existing open structure to accommodate the oil and radiator intake ducts). Large, center-section tanks should also hold more fuel relative to their weight than slim, flat wing fuel cells (much better volume to surface area ratio -so much less of that thick, heavy composite rubber self-sealing tank material for a given fuel capacity).

Delete the wing guns and outer most fuel cells (and maybe clip the wing tips) and you'll increase roll-rate quite a bit too. (and again, 2 .30. 2 .50, and 1 20mm cannon in the nose seems like a reasonable armament, provided a P-38 style cocking mechanism is included to compensate for reliability issues of the American Hispano's firing pin ... or just use the Navy's solution of slathering all the 20 mm rounds with so much wax that they seat into the chamber properly) Fitting a .50 cal within each wing would be nice, and maybe possible if fuel cells were deleted from further inboard of the wing than the .30 cal location, but that's just a guess. (it was really strange the P-63 had external guns AND stupidly limited fuel capacity, with the wings mostly empty in spite of being fairly large and thick -they SHOULD have been able to carry more fuel than the Mustang AND at least 1 .50 cal inside the outer wing portions ... as it was they made poorer long-range fighters than the P-39Q)

Actually, making the radiators semi-external could also allow shifting them forward enough to partially address the CoG issues in the P-39. (using fuel tanks to do this would be a bad idea as they shift weight when empty vs full ... though I suppose having reserve leading edge tanks shifting CoG forward would be OK given they'd always be full in combat)

The P-39 really seems like it should have been able to accept a 2-stage V-1710 without lengthening the rear fuselage as well. It had the physical space there already, but it was being occupied by tons of equipment including some relatively heavy things (like the entire radio array rather near the tail -probably not helping CoG at all). I wonder if the radio equipment could have been relocated behind the cockpit bulkhead ahead of the engine. (delete the rear canopy glazing for a more turtle/razorback configuration and rely on an external rather than internal rear view mirror -the super thick bulckhead framing behind the pilot made for a huge blindspot anyway, and an external rear view mirror wouldn't really be worse -particularly if the modeling in Il-2 is anything to go by ... that rear bulkhead really kills the visibility -switching from the cab style cockpit to a sliding malcolm hood would IMPROVE visibility substantially even if the entire rear glazing was replaced with solid metal)

wuzak · May 8, 2016

No, not enough space for a two stage Allison in the P-39.

Shortround6 · May 8, 2016

I am always amazed at the idea that the P-39 could fit a two stage supercharger when the Bell engineers lengthened the fuselage by about two feet to make the XP-39E and the P-63. The XP-39E was actually ordered as the P-76 to the tune of 400 aircraft but contract was later cancelled or transfered to the P-63.
If the Bell engineers figured they needed bigger fuselage ( and relocated wing) to make it work I don't know why people think it would have been so easy to stuff the needed components in the existing P-39.

Zipper730 · Jan 16, 2017

kool kitty89 said:
I'm surprised I've never stumbled on this video before, it's amazingly informative. The few photos of the XP-39 floating around and limited information (often misinformation) regarding its layout and turbocharger (and radiator) installation don't do it justice at all. (in fact, none seem to show the starboard side of the aircraft at all)

Yeah, all I've seen was the left

It might just be me, but the entire cooling system installation (oil cooler, intercooler, and to at least some extent coolant radiator) look shockingly sloppy and patched together ... almost taped-on.

Looks remarkably boxy, though. It's interesting that their earlier configuration was rounder...

Photo by David Lednicer

BTW: The radiator for the engine was on the left side, the inter-cooler was the boxy structure, and the intercooler was the smaller installation below right?

No boundary layer splitter/sleeves/cowls either, but that's not really surprising compared to the rest.

The P-38 and P-40 didn't have any form of boundary-layer splitter from what it appears. They did incorporate doors/gills to regulate the airflow (The P-39 did not).

NACA-recommended 13 inch rear-fuselage extension probably didn't help

What was the XP-39's length compared to later models?

Also interesting that the XP-39 has a nicer clear-view canopy, both taller and lacking the thick 'anti-roll bulkhead' behind the pilot.

The lowering of the canopy reduced drag if I recall...

Though in that same thread:
XP-39: pros cons

I'll look through this thread next...

wuzak said:
No, not enough space for a two stage Allison in the P-39.

If NACA wanted to add 15 inches to the design, and the P-63 was two feet longer, why didn't they just add the two feet? That would have fixed most of the problems (the P-63 did okay even without a turbo)

wuzak · Jan 16, 2017

Zipper730 said:
The P-38 and P-40 didn't have any form of boundary-layer splitter from what it appears. They did incorporate doors/gills to regulate the airflow (The P-39 did not).

The P-40 did not need a boundary layer splitter - there was little or no boundary layer at that point, as the radiator intake was right behind the spinner.