About Propellers Other Products Articles Bookshop Gallery Links Contact Supercool

Propeller Dynamics

Essential reading for model aircraft contest fliers. This is the only book on the market explaining propeller theory in non-mathematical terms. A rattling good read, I know, I wrote it.


Ducted Cooling for Model Airplane engines 


See the new article More on cooling airplane engines  by reader Graham White

Advances in wing design, especially of F3D pylon race models, have reached the point where further performance gains may not be related to the surface configuration of the model. If there is one area where gains may be expected, it is in engine cooling.

On full-size aircraft, quite large gains in performance have been achieved by reworking the cooling system. I have in mind Roy LoPresti's work on the Mooney light aircraft. From my memory, I recall he turned a 160 MPH airplane to a 201 MPH airplane, largely by reducing cooling drag. This seems like an improbably large figure, so let us look more closely at another example, this time not from memory. 

In the magazine 'Aeroplane' for May 1999, there is an article by the late Lee Atwood, vice-president of North American Aviation in 1940, entitled 'We can build you a better airplane than the P40'.  The aircraft alluded to is no less than the P51 Mustang , so we can expect some excitement!

Lee indicates that the propeller thrust at full power was about 1000lbs. However, the drag of the cooling radiator was of the order 400lbs!  This is shocking, nearly half the available thrust was required just for cooling the engine. For you farmers out there, this is not unlike using horses to plough and harvest your wheat fields: for every 1000 bags of wheat, the horses eat 400! And that is a fact!

By careful design of the radiator and its ducted cooling system, it was possible to use the heat released by the radiator to generate 350lb of thrust, thereby reducing the net drag of the cooling system to just 50lbs. This was a rather special achievement, possible due to the work F.W.Meredith, in 1935, at the Royal Aircraft Establishment at Farnborough. This reduction in cooling drag was mainly responsible for making the  Mustang some 30 MPH faster then the Mk IX Spitfire, despite the higher critical Mach number of the Spitfire wing.

There are a variety of opinions current on whether it is possible to obtain thrust from the heat  residual of I/C engines. That is, the Meredith effect cannot be applied , as the theory is simply wrong. Regrettably, I am too ignorant of thermodynamics to contribute to this debate: however, I believe there may be a book coming out on the cooling theory for I/C engines so watch out for it. In the meantime, if you are an expert and would like to contribute your view to my website, I will happily place your comments here in this article.

Now here is food for thought. Perhaps our speeds are too slow to use the Meredith effect on our racing models, but just how bad is our cooling drag? Are our cowling designs giving rise to more drag than, say, the wing itself? Highly possible, based on those figures for the Spitfire and Mustang.

Recently I had the need to test my bar-stock double-exponential tuned pipe on an OPS 40 mounted on my stationary test stand. In characteristic Supercool fashion, things went wrong very quickly. With the motor completely uncowled, air rushing past it in all directions, I figured that cooling was not an issue. How wrong can you be!

From cold, the motor would burst up on pipe to 27000 RPM, then sag and sag down to 25700, and become impossible to tune. Opening the needle did nothing. Getting pipe data was impossible, so in set the Blues as my alter ego started jabbing me with a pin and making me feel despondent. Was my pipe at fault? I had heard of pipes burning holes in pistons: was my 3 months of development and coding all wasted?

Looking closely at the engine, I noticed that there was castor oil burnt onto the back of the cylinder, offset slightly to starboard. This seemed to indicate excess heating, or at least lack of cooling, at that point. Talking to the C/L speed guru Grant, he spun me a wonderful story. When dyno tuning his K&B 40 racing motor, he was able to determine the cooling pattern from coloration of the liner and piston. Cooling was fine in the front of the engine and in the vicinity of the transfer ports, but not at the rear of the cylinder in the region of the exhaust port. So air was doing the cooling in front, and the fuel also contributing to the cooling.

Cooling with the fuel mixture? This does not sound so good to me. After all, the Spitfire had an intercooler between the supercharger and the inlet manifolds, for the express purpose of cooling the fuel mixture. Just the opposite of what we are doing.

The MB40 pylon motor is well known for its heat stability, especially on the starting line. Previously I thought this may have been due to some wonderful metallurgy or the elimination of the cylinder liner, the latter giving better heat transfer to the cooling fins. But when I looked more closely, I noticed that this motor has rear transfer ports which can cool the rear of the cylinder, unlike the IR, Nelson or RPM.

To my way of seeing things, the MB40 has never performed up to my expectations. The front induction Nelson and RPM have been able to keep pace, despite inferior design specifications. Could it be that the MB40 has gained its temperature stability from using the fuel to induce more uniform cooling around the whole cylinder? If so, great for handling, lousy for power. Hot fuel bad, cold fuel good.

Now back to my sad, obsolete OPS. If I was going to test this darn pipe, I better cool the rear of the cylinder so that, at least, the motor would hold a setting. But how to do this? I didn't want to build a cowling for use on my test stand, what a pain in the butt that would be. Everything I do is conditioned by access to my CNC mill, and its enormous power in sculpting shapes. Could I machine up a metal cowl from bar stock, something that wouldn't flap around in the breeze like composite cowlings are wont to do?

Then it hit me. My gloplug was held in place by a cllamp-ring, which also carried the head fins. What if I was to machine up a new cllamp ring, which retained the head cooling fins, but included a shroud that dropped down over the barrel cooling fins and forced the air round the back of the motor! Piece of cake! Have it done Friday!

Well, it was Friday a month later I had it done, and very racy it looked, too.

OPS duct 1.jpg (16660 bytes) OPS duct 2.jpg (15343 bytes) OPS duct 3.jpg (17105 bytes) OPS duct 4.jpg (17223 bytes) OPS duct 5.jpg (17210 bytes)

Click to enlarge

There are some principles involved in designing a cooling duct such as this. I don't really understand them, but I can regurgitate very well. We Australians are great regurgitators, it assists with our beer drinking.  It goes like this. Blowing cold air at high speed over an engine is a lousy way to cool it. Never mind that that is standard practise, it is still lousy. The heat transfer is poor, and the drag is very high.

The correct way to cool is to blow low speed, high pressure air over the fins. This is standard practise in full size airplanes, and if you know where to look, you will find that this is true. But where do you get low speed, high pressure air from? Well, chances are, if you look in the front cowling holes on a Cessna or Piper, you won't see the cooling fins. That is because the intake air is directed by baffles into the top of the cowling, from where it moves downwards thru the cooling fins, then out.

That chamber formed by the upper cowling is called a plenum chamber: it has a wonderful property. Being of large volume, the air that rushes in through the intake holes is permitted to expand to fill the volume of the plenum chamber. Now the magic part is that when air expands in the fashion, its speed falls and its pressure increases! Wonderful, just what we wanted, low speed air at high pressure.

So my OPS duct had to have a small opening that expanded as much as possible to form a plenum in front of the cylinder. Not much room there, but I was able to expand it about 3 fold. Now that is a lot better than nothing, but nothing like the expansion in, say, the Rare Bear cowling, but I will return to that exotic later.

A baffle was incorporated in front of the boost port to reduce excess cooling of the front of the cylinder, as this would put the liner out of round just as much as overheating the rear of the cylinder does. The duct then closed again to force the low speed high pressure air around the cooling fins, exiting at the pipe manifold. Not so good, I didn't really want to heat the pipe manifold, but then these engines are not really all that cleverly designed, so you must live with it.

Now, where were we? Oh yes, the creaky old OPS now had a cooling duct, which carried my hopes for further glory. Yes folks, just in case you have missed the plot, this is all about Honour and Advancement!

Well I fired up the OPS, and sat back to see if the revs fell. No! Eureka! Over a period of thirty seconds, the revs even climbed slowly from 27000 to 27300, and the needle setting held! This was better than  I dared dream. Pipe temperature analysis shows that the tuned point rises with the temperature, about 100 RPM per 10 degrees, and this was where we were at!

Overjoyed, I fired up again. This time, the motor burst into song and the head fell off! Yes, all 8 head-screws took off in the slipstream and hid in the grass. Supercool, you know that sorrow follows joy, fall cometh from pride! I had torqued the screws lightly, as I always do, to allow for the expansion of the alloy head against the steel screws. But now the cooling was so great the screws were simply not tight enough.

Unfortunately to this date I have not been able to fly either the ducted head or the bar-stock pipe. But not to worry, where to next?

Well about this time a generous colleague in the UK supplied me with an IR40, as being a bit more realistic in performance compared to the 15 year-old OPS. The motor was in perfect condition, beautifully made and appeared to have been run very briefly, as it had light castor oil burn marks on the rear fins, just in the same location as the OPS.

IR cylinder head.jpg (16669 bytes)
Click to enlarge

Right, I thought, now lets do this right this time. Rather than smoking up this fine engine on my test stand, I would fit it with a cooling duct before I ever ran it.

Not only that, I was finding the cost of plugs a bit more than my poverty-level income could sustain. I do enjoy food, it gives me energy: I prefer Testosterone pills, they give me lust: but it was Honour and Advancement I was after, not aching nether-regions.

So I drew up a new list of Specs for the IR shroud:

  1. Cool the cylinder fins with a deep shroud

  2. Make the intake lips round to avoid poor airflow at the intake

  3. Provide the cylinder head with its own separate duct

  4. Provide the gloplug with its own cooling duct

  5. Improve the aerodynamics of the head cooling fins to reduce drag

It is hard to imagine a worse design than the head cooling fins on our racing motors. One would like to think that the air would rush between those lovely head fins, carrying away the heat and cooling the plug so it won't burn out so easily. Well Mr Manufacturer, I suggest you stop spending our money on Testosterone and start using your brains instead.

There is no way the air rushes between those head fins. They have square ends, they are chopped into short lengths for the head bolts, the head bolts stick up into the fins and they have square exits. Aerodynamically, they are hopeless. High speed, low pressure air rushes over the top of the fins, producing high drag and poor heat transfer. Absolutely woeful. Give me those pills, I use them for brain food!

The IR head duct has the following features, which I hope you can see in the photos.

IR cooling duct.jpg (16468 bytes) IR full duct.jpg (14876 bytes) Closed duct assy.jpg (16870 bytes) Duct head fins.jpg (17215 bytes)  
Duct in position.jpg (15667 bytes) Exit view.jpg (15795 bytes) Head view.jpg (16718 bytes)  

 Click to enlarge

  1. The inlet end of the fins are rounded, and the exit section tapers to a fine edge, simulating an airfoil section.

  2. The fins are curved, so that they pass around the head bolts without interruption

  3. The fins lie in a duct, which provides for expansion to yield low speed, high- pressure air. This in turn produces high heat transfer and low aerodynamic drag by the fins.

  4. The gloplug has its own duct, with its own plenum and finning to cool the plug effectively.

Admittedly, the whole assembly has the appearance of a steam engine, but this is a developmental prototype. There is much more that can be done to increase its efficiency. For example, the slotted air- intake is quite the wrong design. LoPresti, in his work on the Mooney, showed that the air intake should be circular, with rounded lips. Not to kiss you with, you Testosterone addled turkey's!

Now on the subject of lips, I promised to return to Rare Bear. There is a lesson to be learnt from this mighty Bearcat, just from standing on the ground and looking in the cowling. Remember,  this is a 530 MPH 3000+ horsepower monster. The heat generated in one flight would warm my house all winter.  So just what does one see of the cooling system?

Firstly, the air-intake is an annular slot around the back of the spinner. It looks to be about 4' wide, which isn't very much. The outer lip of the annulus is highly rounded, to permit smooth flow of the air into the cowling of the air-cooled radial engine. The air then enters a very large plenum chamber, formed between the front row of cylinders and the forward section of the cowling. The expansion must be very large, I would guess a factor of 50, maybe more. So you see, this engine is only going to see very slow moving, high-pressure air at the cooling fins.

The flow of this air after it leaves the fins was not apparent from the ground, but I would not be surprised if the Meredith effect described earlier was in action. One of the key requirements for the Meredith effect to work is control of the back-pressure of the exiting air. This can be provided by the variable cowling gills at the rear of the motor. There is certainly plenty of heat available for generating thrust!

There is yet more to say about the intake flow. If the high speed air were not correctly ducted, it would not expand smoothly to generate the high pressure air. It is important that the fast moving air does not become turbulent. That is also the reason that the P51 intake duct is stuck out so far into the airstream: that prevents turbulent air from the fuselage from entering the duct.

On the Bear, the smooth ducting continues from the lip right up into the plenum chamber. Not only that, but the crankshaft on the reverse side of the spinner is also heavily faired, just as though there was a reversed spinner backed up against the true spinner.

I have often wondered how aircraft like the Bearcat, FW190, Thunderbolt, La-7, Lagg-3 etc, with radial engines, were able to compete so effectively against aircraft with liquid cooled engines, such as the Mustang, Me-109, and Spitfire. These latter slender aircraft had in appearance much lower frontal area. But appearances can be deceiving!

The aircraft that brought understanding to me was the Fw190-D9. In appearance, this is a radial-engined fighter. At least, it has a radial cowl, hardly what you would expect to see on a liquid-cooled in-line engined fighter. But the D-9 is exactly that! The engine is a Junkers Jumo 213A, a 12-cylinder in-line liquid cooled unit.

The radiators are mounted in the radial duct in front of the engine, with the cowling gills immediately behind the radiator duct. This is an even better set-up than in Rare Bear. All the elements are there for a Meredith effect, or at the very least for efficient low drag cooling─..low-speed, high pressure air.

So if you have read those accounts of Kurt Tank, the Fw designer, being bounced by Mustangs and simply flying away from them, then perhaps the story is not so apocryphal as it sounds. Which brings us back to the conundrum. How can a big fat radial engined fighter compete on equal terms with a slender liquid-cooled in-line engine?

The answer lies in the cooling drag. The physical flat-plate area of the airframe comes a bad second to the drag of the cooling system.

Well that just about wraps it up on ducted cooling. All I have left to say is some more bad words to the manufacturers. You are dragging your heels, you are out of date.

Future racing engines will be built with ducted cooling systems, not the primitive stick-it-in-the-slipstream rubbish we have now. Cooling fins will be aerodynamically shaped and placed in plenum chamber ducts. Most likely, the air intakes will be annular, like Rare Bear, and the fins placed vertically on the crankcase to provide up-draft cooling which exits above the pipe, not onto the pipe. The head and plug will have their own ducts, with circular inlets. In this way, fuel will no longer be used as a working fluid for the cooling system. Power will rise, as the ducted cooling actually cools the fuel in the transfers. Engine handling will improve, as the uniform radial cooling of the cylinder maintains the perfect circular shape of the cylinder. Plug life will increase. Compression ratios can then be increased, and pipe stinger exit diameters reduced, without risk of burning holes in the piston.

And all this will be sourced back to Supercool!  Supercool, the father of the modern racing engine. Weep your hearts out Nelson, Metkemyer and Phelan! If you are real nice to me, I will even let you give me some engines and money to develop your obsolete engines for you!  Hurrah!                                                                 

 Back to Top