Meredith Effect and the P-51

[Holtzauge]

Some more input on the poor Me109G radiator design relative the Fw190D: The high radiator drag for the Me109G of around 14% as compare to the 8.8% figure for the Fw190D9 is according to Hoerner explained by Messerschmitt themselves: Me reports 109/10/L/1941 and 109/15/L1/1942 states that the loss of momentum of cooling air to the rear was due to excessive leakage in the radiator duct which increased the overall total drag of the cooling installation.

 

[Kurfürst]

Hmm, Hoerners study is actually a theoretical estimate of the drag % of the 109G, and is appearantly not based on actual measurements... in any case, Hoerner for example gives a drag coefficient of 0.036 for the Bf 109G, whereas the actual Messerschmitt polars IV/139/48 give 0.023, so Hoerners post-war study has some very obvious inaccuracies in his estimate.

As for the FW 190D and Bf 109F/G/K cooling installating, the major difference between the two is that the FW 190D employed the rather ingenious Junkers style annualar radiator in front of the engine (also seen on the Ju 88), probably the best solution when it comes to minimizing drag, as no other radiator housings were needed sticking out of the fuselage, wings etc., as was common on most fighter designs.

 

[Colin1]

PS - Colin - you may faithfully comply with Herr Ho Hun demand that you give him radiator area.

2.1 Sq ft for P-51A/Allison
2.7 sq ft for first pre P-51B Merlin modification by RAF, re-designed to 2.9 Sq ft for subsequent the one first tested by RAF. This radiator design remained for all B/C/D/K. I am not sure about the H.

Thanks dr
Dammit, I know for a fact I've got this data somewhere!

 

[Kurfürst]

2.1 Sq ft for P-51A/Allison
2.7 sq ft for first pre P-51B Merlin modification by RAF, re-designed to 2.9 Sq ft for subsequent the one first tested by RAF. This radiator designe remained for all B/C/D/K. I am not sure about the H.

Interesting,

For comparison, the frontal areas for the coolant radiators used on the Bf 109G/K were:

G-1 through G-14: 33.6 dm2, or 3,62 sq. feet. (DB 605A or AM powered)
G-10, G-14/AS, K-4, : 42 dm2 or 4.52 sq. feet (DB 605 AS or D powered high alt variants)

Oil cooler measured 6.5 (605A, AM) and 8.5 square decimeters (605AS, D), respectively.
Kurfrst - Leistungen Me 109 G-14/U4 mit DB 605 AM u. ASM.

(these appear as total size to me (there were two coolant radiators), but the context is not 100% clear.)



The figures for the Spitfire IX are the following.

Assymetric sized coolant radiators were used, figures without oil cooler and intercooler radiators (also housed underwing), only for coolant radiators:

Frontal areas:

Radiator 1: 0.36 x 0.28 m = 0,1008 m2 = 1,085 sq. ft.
Radiator 1: 0.38 x 0.28 m = 0,1064 m2 = 1,145 sq. ft.
Total: 2,23 sq. ft. coolant radiator frontal area.

Of course the matter gets more complicated with the depth and coolant pipe density, material of the radiator being factored in..

 

[Colin1]

So all these aircraft suffer a cooling related loss not a gain. Now if the P-51 where to produce thrust instead of drag then that seems like a spectacular design feat. I think it more likely that the good radiator design lowered the drag to a smaller percentage of the total drag which is good enough I think

Hi Holtzauge
you hit the nail on the head - that's precisely what the P-51's cooling arrangment did; roughly, propeller thrust was around 1,000lbs, gross radiator drag was around 400lbs, momentum recovery thrust from the cooling arrangement was around 350lbs - the net cooling drag was around 3% of the thrust of the propeller.

It doesn't overcome cooling drag, but it does a very impressive job of negating it to a large degree; as the article at the beginning of this thread points out, for the Spitfire IX to fly at 400mph required an extra 200hp over the Mustang at the same speed.

I thought alot of the Spitfire's potential was lost in cooling drag and so it would seem did the Air Ministry; the Martin-Baker MB5 employed the Mustang's cooling arrangement giving it a sparkling performance but by that time of course, the world had turned their attention to jets.

 

[Holtzauge]

Hmm, Hoerners study is actually a theoretical estimate of the drag % of the 109G, and is appearantly not based on actual measurements... in any case, Hoerner for example gives a drag coefficient of 0.036 for the Bf 109G, whereas the actual Messerschmitt polars IV/139/48 give 0.023, so Hoerners post-war study has some very obvious inaccuracies in his estimate.

As for the FW 190D and Bf 109F/G/K cooling installating, the major difference between the two is that the FW 190D employed the rather ingenious Junkers style annualar radiator in front of the engine (also seen on the Ju 88), probably the best solution when it comes to minimizing drag, as no other radiator housings were needed sticking out of the fuselage, wings etc., as was common on most fighter designs.

1) Hoerners was a professional aerodynamicist in Germany during the war and got a job doing the same in the US after the war so I guess they thought him competent enough. Frankly, I rather trust NASA judgement on this than yours.

2) His books are classic and well known in the aerospace business

3) I agree that his TOTAL estimate for the Me109 is to high. If you look closer at his analysis though, the high figure he arrives at is due to a too low estimate at Steig&Kamppf lesitung where he assumes 610 instead of 622 Km/h. Even with this corrected, I see no reason why not to trust his very detailed analysis of the drag components which spans 5 complete pages in the book.

4) You however, are comparing apples and pears: Hoerners estimate includes induced drag while yours is just Cdo. Even If you want to argue for a lower Cdo then the cooling drag percentage becomes even HIGHER than the 14.3%

Until you come up some figures yourself, I will continue to trust Hoerner's estimate on the percentage of cooling drag of the total (14%), especially since the relatively high cooling drag was a problem acknowledged even by Messerschmitt themselves in the reports that Hoerner refers to.

 

[HoHun]

Hi Clay,

>You do seem to be fond of bashing anything American though and claiming everything else was better designed.

I did not claim the Me 109 radiator to be better. I'm simply pointing out that there is not enough data presented here to make a judgement.

As I'm in the habit of backing up my statements with data to support them, I certainly reject the suggestion I'm "bashing" anything.

If what you perceive as "bashing" conflicts with your perception of reality, go find some data that supports your perception of reality, and we can have an intelligent discussion. I always appreciate new data.

If however you make big claims based on no data at all, as some people like to do, an intelligent discussion obviously is out of the question.

Regards,

Henning (HoHun)

 

[HoHun]

Hi Juha,

>Also the accident report of Black 6 probably would give some clues, IIRC the AVM took off with radiator flap control in manual setting, I cannot recall the radiator flap setting but anyway the cause of the crash was insufficient engine cooling, IIRC.

In the case of the Black 6 accident, the accident report pointed out that the radiator control valve was not in any valid position, but halfway between two settings (probably "manual" and "automatic"), locking the radiator flaps into position. The engine overheated, blowing some steam through an overpressure valve positioned so that the steam would become visible to the pilot so that he'd be alarmed to the overheating condition.

The Black 6 pilot saw the steam, but mis-interpreted it in some way, thinking the engine had run out of oil or something like that. Thus he made an immediate emergency landing (downwind, I believe), stopping the engine so that it would not be destroyed from running dry of oil, which had the side-effect of reducing drag in the fast downwind landing he was attempting.

He landed on the main runway, overshot, had enough speed to lift off again and overfly the public road behind the runway end, and landed again on a field on the other side of the road. Unfortunately, the field he landed on was being plowed, and when he ran into the part that had already been plowed, the wheels were caught in the furrows and he flipped over.

If he had simply put the radiator control valve in the "automatic" (or "fully open") position, he could probably have continued his airshow routine without problems.

Regards,

Henning (HoHun)

 

[Kurfürst]

I see no evidence that the Messerschmitt 109F-K's cooling drag was 'relatively high'. What was the cooling drag % on comparable fighters for example?

For example RAE gives the following figures for 'drag as the % of total profile drag - Powerplant' (no seperate cooling system referenced) :

Hurricane I: 11.8 %
Hurricane II: 16%
Spit Vc: 18.2 %
Spit IX: 19.1 %
Typhoon: 27.1 %
Tempest V: 24.3 %
Mustang X: 22.9 %
FW 190A : 21.5 %

In comparison, Hoerner estimates the Me 109G's 'engine and radiator installation' being 23.3% of the drag. That appears to be a pretty average figure, quite close to the Mustang X's.

It was very much higher than the FW 190Ds, already explained why (Junkers style radiator, which fundamental difference in installation you appear to completely ignore), but it is extremely doubtful that the installation was much different in this aspect than other more commonly used, conventional cooling systems (ie. on Spitfire, Mustang, Yakovlev etc. fighters).

The percentage of the total drag is of course is also dependent on the design. Assuming two entirely identical cooling system designs, the percentage will be higher on the aircraft with the smaller wing area, as in the aircraft with a larger wing/fuselage area, other components will contribute more heavily to drag, even if in absolute terms the drag of the radiator installations are identical.

If you would increase the wing size on the 109 by 25%, and leave everything else alone, the % of drag of the engine/coolant system will very likely be a smaller percentage of the total drag. Anyway, what matters is the actual drag generated by the cooling system installation, relative drag to total drag is a pretty meaningless figure, unless you are comparing different models of the same aircraft.

For example the Spitfire IX, with 19.1 % of powerplant-related drag to the total drag, had a Cd0 of 0.0229 with 242 sq. feet wing area, ie. 5.58 sq. feet total drag.
19.1% of that, or 1,058 sq. feet, was powerplant-related drag.

The Mustang X, with 22.9 % of powerplant-related drag to the total drag, had a Cd0 of 0.0227 with 233 sq. feet wing area, ie. 5.28 sq. feet total drag.
22.9 % of that, or 1,211 sq. feet, was powerplant-related drag (RAE figures). (Which I find odd, but that is what RAE gives).

For example the Bf 109G, assuming Hoerner's value of 23.3 % of powerplant-related drag to the total drag, a Cd0 of 0.0230 with 174 sq. feet wing area, ie. 4.02 sq. feet total drag. 23.3 % of that, or 0,93 sq. feet, was powerplant-related drag.

"...especially since the relatively high cooling drag was a problem acknowledged even by Messerschmitt themselves in the reports that Hoerner refers to. ..."

Hoerner does not refer to any Messerschmitt reports in my copy... so I have some doubts about this 'acknowledgement'. In particular the testing of Bf 109F in ealry 1942 as a G testbed considered the cooling system entirely satisfactory.

Moreover Hoerner does not state in Chaper XIV that the Messerschmitt 109 radiator had leakage, but that a radiator generally similiar to the one used on the Me 109 (subtype unspecied) may have had such leakage, because that radiator installation had high coefficients of drag.

Of further interest to the subject of the Mustang's ducted radiator, is the testing of a protounding belly cooper scoop on Me 109 V31 in October 1941. So - little surprise here - the solution was not unknown in Germany or elsewhere, but for some reason they opted against it.

Hell even the Hurricane used such!

 

[HoHun]

Hi Kurfürst,

>For comparison, the frontal areas for the coolant radiators used on the Bf 109G/K were:

As the "Meredith effect" is driven by the transferred engine heat, the relevant parameter for analysing different cooling systems is not the frontal area, but the total surface used for heat transfer.

Regards,

Henning (HoHun)

 

[Kurfürst]

Hi Kurfürst,

>For comparison, the frontal areas for the coolant radiators used on the Bf 109G/K were:

As the "Meredith effect" is driven by the transferred engine heat, the relevant parameter for analysing different cooling systems is not the frontal area, but the total surface used for heat transfer.

Regards,

Henning (HoHun)

Hi,

Kurfrst - Bf 109G/trop Middle East trials: Dimesnions, Weights and Performance

COOLANT RADIATORS:

Two light metal radiators, type SKF/Behr. Al.F.750 B.

The coolant used is a mixture of water and glycol.

The cooling surface is 125 sq. ft., capacity 1.4 gallons
and the weight is 131 lb. The test pressure is 28.4 lb./sq.in.

 

[Juha]

Hello Kurfürst and Happy Valentine Day!

on tail wheel, at least in Prien's JG 53 Vol 2 the caption on p. 480 "…Note the retracted tailwheels of "Black 1" and "2" rarely seen on Bf 109G-2s."

Also Fernández-Sommerau in his Recognition Manual p. 55 writes "a retracted tailwheel was planned for the G series, but this was soon abandoned." Now I have no firm opinion how reliable Marco's book is but that is anyway what he wrote.

Quote:" Well thats how the radiators work on the 109. They close if the temperature is much below 85 degrees, and open if the temperature is over 85 degrees."

They had many positions as the graph from Mike's site shows. BTW Thanks for the graph, haven't seen it before.

Quote:" Of course is it does, its how it is supposed to work in the first place, the external temperature was rather cold as you notice, plus it is difficult to miss the very high climb rate of tested MT 215 in the first 2500 meters, above which it is, however very similiar to other G-2 curves at 1.3ata.

The only reasonable explanation I can think of is that the drag was much less with the radiator flaps so closed in the initial climb."

You can of course think what you want, but as has been seen for ex 100octane fuel and FC thread what you think doesn't always match with reality. In Finnish if he had meant that radiator flaps opened first time at 2500m he would have written that flaps opened first time at 2500m, not that they were FULLY OPEN first time. Of course cool air and higher speed kept the radiator flaps less open than would have been case if Kokko had flown at the speed recommended in handbook. And as noted in the report because there was not exact info on the weight of the plane and how weight was distributed the figures were not fully comparable to German figures.

Hello Holtzauge and Happy Valentine Day!
thanks for the info!

Juha

 

[Holtzauge]

Hi Holtzauge
you hit the nail on the head - that's precisely what the P-51's cooling arrangment did; roughly, propeller thrust was around 1,000lbs, gross radiator drag was around 400lbs, momentum recovery thrust from the cooling arrangement was around 350lbs - the net cooling drag was around 3% of the thrust of the propeller.

It doesn't overcome cooling drag, but it does a very impressive job of negating it to a large degree; as the article at the beginning of this thread points out, for the Spitfire IX to fly at 400mph required an extra 200hp over the Mustang at the same speed.

I thought alot of the Spitfire's potential was lost in cooling drag and so it would seem did the Air Ministry; the Martin-Baker MB5 employed the Mustang's cooling arrangement giving it a sparkling performance but by that time of course, the world had turned their attention to jets.

I think the good low drag characteristics of the Mustang is also in part due to the wing profile. Surface imperfections will certainly cause transition to turbulent flow in a wedge shaped area behind the imperfection but going spanwise a wing kept in reasonable conditions by a good crew chief would at least in some areas have laminar flow longer than more conventional profiles. Every little bit counts and a little more laminar flow than that the NACA 230 series that many other design had or the 2R1 on the Me109 would add up to a slight advantage. Some people erroneously compare profile drag looking at the figures with NACA standard roughness which is more like sandpaper. Service condition may not be polished, but it certainly is not sandpaper

About the Spitfire radiators I certainly agree. The thing that strikes me as odd is that AFAIK the British never introduced a system to remove the boundary layer from being ingested into the radiator. Letting the boundary layer into the expansion after the inlet usually leads to a boundary layer separtion and excessive drag so it surprises me that they never introduced something similar as on the Me109F which was certainly a step in the right direction and must have led to a significant reduction of cooling drag as compared to the Me109E.

Seeing they beefed up the Spifire wing, improved roll rate etc it certainly is strange that they did not do something about the radiators. Or maybe they did? I don't have Morgan's excellent book on the Spitfire history but maybe someone can check on possible radiator fixes or how the design team of later models reasoned on this point?

 

[Juha]

On Meredith Effect and Bf 109
DB's Nallinger mentioned among his complains on Mtt that "The oil back pressure and the water back pressure have been too high from the outset and had to be approved under protest - made together with officials in the face of delivery quotas..."

To me, a complety layman, that indicates at least some utilisation of Meredith Effect or at least radiators produced more trust than DB would has liked.

Hello HoHun
thanks for the Black 6 accident info. I have seen a article on it in an old AM but clearly recalled it wrongly. I had not time to dig the magazine from my archives and in fact I have misplaced some of my AMs so I'm not too eager to go through them.

Juha

Later ADDITION
Quote:"If he had simply put the radiator control valve in the "automatic" (or "fully open") position, he could probably have continued his airshow routine without problems."

Yes that part I remembered, that was why I mentioned that he was/is an AVM = Air Vice Marshall.

 

[Holtzauge]

For example RAE gives the following figures for 'drag as the % of total profile drag - Powerplant' (no seperate cooling system referenced) :

Hurricane I: 11.8 %
Hurricane II: 16%
Spit Vc: 18.2 %
Spit IX: 19.1 %
Typhoon: 27.1 %
Tempest V: 24.3 %
Mustang X: 22.9 %
FW 190A : 21.5 %

Please quote the RAE report or post the data. Just quoting "RAE" is a bit vague no?

Also profile drag if just a part of the total drag so no surprises if the PERCENTAGES above are higher than 14% as Hoerner claims as the cooling drag percentage of the TOTAL drag not just the profile drag.

A big warning bell for these numbers: The fw109A is air cooled and comes up with 21.5% comparing with Mustang at 22.9%. So now air cooling provides lower drag than liquid cooling?

Comparing apples and pears again are we?

 

[Kurfürst]

Hello Kurfürst and Happy Valentine Day!

on tail wheel, at least in Prien's JG 53 Vol 2 the caption on p. 480 "…Note the retracted tailwheels of "Black 1" and "2" rarely seen on Bf 109G-2s."

Well two out of three Bf 109Gs on that picture have the semi-retractable tailwheel. Unfortunately, pictures of in-flight Bf 109G-2s appear to be rare, but all Bf 109Fs (which had the same tailwheel) show them as retractable.

In addition, both Bf 109G-2s captured by the Russians in the end of 1942 (WNr 14 513 for example), they only noted that the G-4 did not have a retractable tailwheel anymore.

It appears there are plenty of examples of Bf 109Gs with retractable tailwheel, which was the rule, and not the exception. The tailwheel was fixed down with the introduction of the enlarged tailwheel in early 1943.

Also Fernández-Sommerau in his Recognition Manual p. 55 writes "a retracted tailwheel was planned for the G series, but this was soon abandoned." Now I have no firm opinion how reliable Marco's book is but that is anyway what he wrote.

He appears to be writing of a fully retractable tailwheel, not the semi-retractable one that was present already on the Bf 109F, and plenty of pictures show it retracted in flight.

Fully covered wheel wells were also planned for the 109G, but these things did not realize until the 109K.

The only reasonable explanation I can think of is that the drag was much less with the radiator flaps so closed in the initial climb."

You can of course think what you want, but as has been seen for ex 100octane fuel and FC thread what you think doesn't always match with reality.

I understand your frustration with that thread, given that you were unable to support your 'reality' with any evidence. But please, vent it off somewhere else.

In Finnish if he had meant that radiator flaps opened first time at 2500m he would have written that flaps opened first time at 2500m, not that they were FULLY OPEN first time. Of course cool air and higher speed kept the radiator flaps less open than would have been case if Kokko had flown at the speed recommended in handbook. And as noted in the report because there was not exact info on the weight of the plane and how weight was distributed the figures were not fully comparable to German figures.

Still, the fact is that the Bf 109G-2 MT 215 was showing much higher climb rates, up to 24.7 m/sec than in any other German or Soviet test of G-2s, but very similiar above 2500 m.

One explanation is that the G-2 was capable of climbing at 24+ m/sec with its radiators being open even at the 30-min rating, the other is that the drag of the aircraft was less in the initial stage of the climb, for which a likely explanation, given Kokko's notice of the radiators opening ('At 2500m coolant radiator flaps open fully for the first time. After that varying between open and closed position.') at around 2500m. We know they would open if the coolant temperature begun exceeding 85 degrees Celsius or thereabouts.

In any case, there is no indication that the coolant capacity was inadequate in either of these three tests, quite the contrary, given that the radiators did not need to open fully in the Finnish trials during the first 1.5-2 minutes required to climb to 2500 meter alttiude to maintain the 85 degrees or so degrees Celsius of coolant temperature, there appear to be plenty of reserve (unsurprisingly, given the sheer size of the radiators).

 

[Glider]

Hi Juha,

>Also the accident report of Black 6 probably would give some clues, IIRC the AVM took off with radiator flap control in manual setting, I cannot recall the radiator flap setting but anyway the cause of the crash was insufficient engine cooling, IIRC.

In the case of the Black 6 accident, the accident report pointed out that the radiator control valve was not in any valid position, but halfway between two settings (probably "manual" and "automatic"), locking the radiator flaps into position. The engine overheated, blowing some steam through an overpressure valve positioned so that the steam would become visible to the pilot so that he'd be alarmed to the overheating condition.

The Black 6 pilot saw the steam, but mis-interpreted it in some way, thinking the engine had run out of oil or something like that. Thus he made an immediate emergency landing (downwind, I believe), stopping the engine so that it would not be destroyed from running dry of oil, which had the side-effect of reducing drag in the fast downwind landing he was attempting.

He landed on the main runway, overshot, had enough speed to lift off again and overfly the public road behind the runway end, and landed again on a field on the other side of the road. Unfortunately, the field he landed on was being plowed, and when he ran into the part that had already been plowed, the wheels were caught in the furrows and he flipped over.

If he had simply put the radiator control valve in the "automatic" (or "fully open") position, he could probably have continued his airshow routine without problems.

Regards,

Henning (HoHun)

Re the accident I was there when the crash happened. I cannot add anything to the above as I have not read any accident report. All I can say is that this was the second flight of the day and the first one had been delayed due to a technical problem, what I do not know, but it was delayed. The second flight was the last flight of the day and most people believed that it was only done as this was the last day that the plane was deemed airworthy, its certificate running out that night and the authorities were refusing to renew it under any circumstances.
The emergency landing wasn't downwind, of that I am certain as all fights that day had been in the same direction, but it was fast and late the initial touchdown being if I recall correctly, about 1/3rd of the way down the runway. At the end of the runway the land banks up quite quickly and the pilot did well to gain enough altitude to make it over the road.
I am pretty certain that the engine wasn't cut but there was reduced power as if he was trying to lose speed before touchdown and there was little wind that day to help him.

Thats all I can say without guessing

 

[drgondog]

I think the very good radiator design was one of the major reasons for the the P-51's very good speed performance relative other contemporary designs. However, I seriously doubt that the system actually produced thrust. What it probably did though was to produce considerably less drag than the competition.

I have long felt the same way Holtzauge. When I came across Gene Lednicer's VSAERO study via Crumpp - there was a supplement discussion relating to designe study of both the trailing edge of the fillet as well as the radiator cowling design for the racer Strega... which support this thesis

Some percentage numbers of drag from "Kuhlung" from "Widerstandsdaten von Flugzeugen" dated december 1944:

Fw190A8: 15%
Fw190D9: 8.8%
Ta152H1: 12.2%

This most likely also includes drag from oil cooling but gives the general idea.

From Hoerner Fluid dynamic drag book page 14-6:

Me109G: Radiator drag 0.8 sq ft of 5.6 sq ft total=14.3%

So all these aircraft suffer a cooling related loss not a gain. Now if the P-51 where to produce thrust instead of drag then that seems like a spectacular design feat. I think it more likely that the good radiator design lowered the drag to a smaller percentage of the total drag which is good enough I think.



Hoerner also gives a hint of why the P-51 is better on page 9-3 in the same book:

Pressure loss in a radiator system is according to Hoerner roughly proportional to internal speed w**1.8

Heat transfer in a radiator system is according to Hoerner roughly proportional to internal speed w**0.8

So dividing the pressure loss and heat transfer we deduce that the lower the speed over the radiator core the better. I would expect this formula to be valid for reasonable variations of radiator core speeds since Hoerner uses it for design comparisosns so it should serve it's purpose for what we are looking at here.

Now I do not have the actual figures but by just eyeballing the proportions of radiator inlet area to core area on the P-51 and Me109K I would say that the speed over the core should be lower for the P-51 than in the Me 109K seeing that the speed over the core will be proportional to inlet area divided by the core area.

Good speculation on your part - This is what I have.
The P-51B modifications from P-51A/Allison Original duct resulted in the following for M=.570/V=430mph:
Type Flap Opening Mass Flow Rate
oil coolant oil duct coolant duct
P-51A 0.6 1.3 0.107 0.650

P-51B 0.6 1.3 0.093 0.412
3.1 5.9 0.244 0.796
8.0 14.5 0.299 1.220

Reference pg 79 Mustang by Gruenhagen

There was a very slight increase of CD (.0006 delta) for the new modified duct but a complete removal of 'rumble' and a dramatic reduction of mass flow rate over both the radiator and the oil cooler. The total Duct opening went from 197sq in to 177.3 sq in (138.7 Radiator plus 38.6 oil cooler (split internally))frontal area


From this we can deduce that the P-51 has a more efficient radiator design than the Me109K due to the lower speed over the radiator core.

I have the complete design drawing/spec package for all variants but have been unsuccessful at obtaining either the aero or structural analysis details.

 

[Holtzauge]

The P-51B modifications from P-51A/Allison Original duct resulted in the following for M=.570/V=430mph:
Type Flap Opening Mass Flow Rate
oil coolant oil duct coolant duct
P-51A 0.6 1.3 0.107 0.650

P-51B 0.6 1.3 0.093 0.412
3.1 5.9 0.244 0.796
8.0 14.5 0.299 1.220

Reference pg 79 Mustang by Gruenhagen

There was a very slight increase of CD (.0006 delta) for the new modified duct but a complete removal of 'rumble' and a dramatic reduction of mass flow rate over both the radiator and the oil cooler. The total Duct opening went from 197sq in to 177.3 sq in (138.7 Radiator plus 38.6 oil cooler (split internally))frontal area

I remember reading about the P-51 duct problems and the "rumbling" sound somewhere else as well but do not remember the source right now. Interesting that they both solved the rumbling AND managed to lower the mass flow at the same time. Since the mass flow would be governed by the cooling need this would indicate that the necessary heat transfer could be done with less air flow which would directly translate into lower drag. Also interesting to see that the Mustang with the higher powered Merlin managed to make do with less mass flow than the Allison powered version.

Another thing that that could go some way to explain why the P-51 had lower radiator drag than the Me109 is the very rapid expansion after the inlet in the Me109 compared to the more gradual compression in the P-51 duct. It's a well known fact in aerodynamics that the steeper the adverse pressure gardient (compression) is the higher the risk of a boundary layer separation right? Given the short distance from the inlet to the radiator, the Me109 engineers have no choise but the rapid expansion we see in the drawing provided by Kurfurst above. For the P-51 OTOH there is a more gradual compression which would lessen the risk of a separation in the duct.

Finally a thought about radiator placement: If the wing position was a good way to go one has to wonder why Willy did not plan to place the radiators in the wing for the Me109 replacements under consideration: The Me309 placement was in the fuselage and Me209 had an annular radiator like on the Fw190D

 

[Kurfürst]

Another thing that that could go some way to explain why the P-51 had lower radiator drag than the Me109

Do you have any figures for the respective radiator drags or its a guess/speculation?

Finally a thought about radiator placement: If the wing position was a good way to go one has to wonder why Willy did not plan to place the radiators in the wing for the Me109 replacements under consideration: The Me309 placement was in the fuselage and Me209 had an annular radiator like on the Fw190D

Take for example the Me 209 and Me 309. Both designs opted for wide undercarriage, and a simple glace on their drawings reveal that there is simply no place in the wings to place the radiators there.

Me 309:

Taking a look at the Me 209 II,

again we see the wide undercarriage, and no space for installing of radiators or ducting in the wings. Given that the proto was merely a modified, stock Bf 109G-5, there would be no space in/under the fuselage either (the fuel tank is there, and taking the radiator even further is probably a not a good idea CoG-wise).

In addition the team was forced to use the Jumo 213 instead of the DB 603, and the Jumo already used the annular radiator in existing designs (Ju 88 etc). Why change that, esp. when there is nowhere else to put that radiator anyway? The reason for the choice of an annular radiator is rather obvious... besides if I would want to go down on the road of good old Holtzeuge-style rhetorics, I would ask, why was the Me 209 II and Me 309 eventually loose out against the old Me 109 design...? (for a myriad of other reasons than radiator arrangement, of course).