Revised 05/21/00

Before I started, I read everything that I could find on the internet with regards to liquid cooling as well as "Alternative Engines" compiled by Mick Myal. AE has some theoretical as well as practical information on cooling engines in aircraft. I came to the conclusion that many people have no idea about what they are doing nor are they applying physics or even common sense in most cases. We have a lot of people with relatively low hp engines and huge radiators with marginal cooling. From a practical point of view, if you look at a P51 with 1500 hp and compare the radiator size to a 150 hp aircraft, it should be obvious that you don't need a radiator that big to cool an engine with 1/10th the hp if the installation is properly designed. Yet, many people have multiple auto rads hanging out in the breeze to cool a 130 hp engine.

Liquid cooled aircraft present several challenges. The biggest one is adequate ground cooling without incurring a huge drag penalty in cruise flight. I can see several common mistakes made by many.

1. It is foolish to expect good ground cooling when the inlet air is picked up at the prop shank, close to the spinner. While this is usually a high pressure zone in flight, it is essentially dead air on the ground with many metal and wood propeller types.

2. Fixed geometry air outlets make no sense on aircraft with a huge range of speeds operated at low altitudes. All military aircraft in WW2 had variable geometry outlets to allow adequate ground cooling and still have low drag in cruise.

3. Proper ducting is a must! Hanging a huge rad out in the breeze without proper ducting does not work because the pressure differential across the rad is very small so there is little net flow through the core. The air must be slowed down from the free stream velocity considerably before reaching the core. This is done by using a diffusion duct, small at the opening for low drag and larger at the core. As the air slows down in the duct, its pressure increases and velocity decreases. This is a win-win scenario. The slow moving air spends more time in contact with the core fins and the high pressure allows us to take advantage of using a deeper core with minimal frontal area. Again, look at the P51 as a good example.

4. Don't dump exit air into a high pressure zone. Basic physics tells us that air always flows from high to low pressure. It serves to reason that for the highest possible flow, the inlets should be located in an area of clean, high pressure flow and the outlet should be placed in a low pressure area. The other side of the coin is that dumping air into a low pressure zone may cause more drag in cruise due to turbulence and airflow separation. This may be another compromise solution between ground cooling and low darg in cruise flight. An adjustable cowl flap on top of the cowling away from the windshield might prove useful for ground cooling. This could be closed for cruise, dumping air out the bottom as most aircraft do with an adjustable cowl flap there to regulate airflow and reduce drag.

5. Air at high speed has inertia, therefore it cannot quickly turn tight corners without help. Many people seem to expect a huge rad laid down on edge with a poor scoop setup to actually cool. Usually it doesn't. Air likes to be directed straight between the tubes and fins, not at some obtuse angle. Guide vanes can help considerably here in the ducting.

Automotive radiators could be used on aircraft use if they could be fitted in the tight confines of an aircraft cowling. Usually they won't fit or cannot be properly ducted. Cars operate at relatively low speeds and generally have minimal pressure differentials across their rads because of unfavorable inlet to outlet geometries so they need a large area rad. Frontal area is of little concern on a car. Cars use a thin core with large frontal area because this makes sense in the application. Aircraft are frontal area sensitive and operate at high speed hence a deeper core with smaller frontal area usually must be fitted. Since the engine only makes 10-20 hp running on the ground, the small frontal area should be adequate if the rad inlet is located somewhere in the outer 3/4 of the prop radius and has moveable outlet flaps to help airflow at low rpm but this is no means certain. Electric fans may be needed with small rad areas for ground cooling due to the low velocity air off the prop at idle and its inability to penetrate the thicker core.

Many people are going to say that there are some liquid cooled aircraft flying successfully with fixed outlets. This is true, but they MUST be incurring a huge drag penalty at higher speeds and low altitude or be marginal on the ground for cooling.

Theoretical Considerations

The basic variables with regards to successful cooling on your aircraft are:

1.Core Size. The radiator must have enough physical area to dissipate the heat rejected into the coolant for your maximum climb power and minimum climb airspeed. This is usually the worst case condition. I like to refer to core volume, height times width times depth, rather than frontal area. Ground running with low rad areas may be another problem which may be encountered even if the design cools well in the air. Once a maximum rad area is chosen, adequate cooling must come from increases in core depth to achieve proper cooling. Unfortunately a thicker core may compromise ground cooling at the expense of improved flight cooling.

2. Core Effectiveness. This refers to how efficient a core is per unit volume. Tube spacing, core depth, fin spacing and design and materials are important considerations here.

3. Mass Flow. This is the mass or weight of air running through your rad. The higher the mass flow, the more heat that can be dissipated through the core. It should be noted that your are exchanging heat between 2 different mediums here, water and air. Water has about 800 times more mass than does air. Therefore you need a lot higher airflow through your rad than water flow from the engine to transfer the heat away. Fortunately, in an aircraft we have lots of airflow available. A perfect duct with an area of 1 square foot will flow 17,600 cubic feet per minute at 200 mph or about 1346 lbs. per minute. If our coolant flow is 15 gallons per minute, we are pumping about 150lbs. of coolant per minute. This gives you some idea of the flows involved. It should be noted that air density and mass decrease with altitude. On a naturally aspirated engine, this is not a problem as hp also decreases with altitude. On a turbocharged engine, rated hp stays the same up to critical altitude but mass flow available drops off. This can present cooling problems sometimes at high altitudes and high power settings. Fortunately, these problems are offset somewhat by increased TAS and colder inlet temperatures.

4. Temperature Differential. The higher the difference between what you are trying to cool (engine coolant) and the cooling medium (ambient air), the more efficient the process. Since we cannot control the ambient temperature and the coolant temperature should be held fairly constant by the thermostat, we have little control here. On a hot day, there is less difference between coolant and air, hence less BTUs are being dissipated. Most engines prefer to run between 160 and 190 degrees F and this cannot be safely increased without lowering engine life to some degree.

What to use for a Radiator

Radiators are usually made from either aluminum or brass. Aluminum is about 1/3 the weight and has about 20% better thermal conductivity than brass. The choice is pretty clear here. For interest sake, copper and silver have about twice the conductivity that aluminum has but they are 3 to 4 times heavier. If weight or cost is not a consideration and you can find someone to build you a custom core from these materials, you could certainly reduce the core size for less drag. Copper was used on the Italian seaplane which still holds the world piston seaplane speed record at 440 mph.

Some people have employed automotive air conditioning evaporator cores for rads on aircraft and I would have to agree that these are an economical choice for higher speed aircraft. However recent flow bench testing has shown that the GM cores are very draggy compared to modern, oval tubed rad cores. While many people are flying the evap cores successfully, they are not the best solution if drag is a concern. Cheap they are.

I sourced 3 GM evaporator cores in the boneyard. One was from a 1980 Olds Cutlass. It measures 10.75 X 10.5 X 3.65 inches and the actual core is 7.75 X 10.25 X 3.65 inches. Total core volume is 288 cubic inches. Weight is 5 lbs. The second core was from a 1981 Chev Caprice/Impala. This measured 13.5 X 9.125 X 3.625 with a core of 10.5 X 9 X 3.625 inches and a volume of 342 cubic inches. Weight is 5 lbs. 9 oz.

GM Evaporator cores,'80 Cutlass left, '81 Caprice on right

The third core was from a '92 Cavalier. It measured 10 X 7 X 2.875 with a core volume of 201 cubic inches, 13 passes and a weight of 4 lbs.

Caprice core on top of Mazda test radiator

Pressure Drop

I also did pressure drop tests on the cores using my cylinder head flow bench. The thicker the core, the more resistance to airflow through it. With the cores mounted over the exhaust port of the bench flowing 160 SCFM (area was 9.6 square inches), the thin Mazda core only showed a drop of 1 inch of water across it. The Caprice core showed 5.5 inches across it. It would appear that pressure drop is fairly proportional to core depth with the Caprice core having about 5.8 times the depth of the Mazda rad. For comparison, I also flowed my 3.5 inch thick Spearco intercooler core. It showed a drop of 7.5 inches across it.

Update 07/16/03

After running the engine we discovered several things which back up the original theories above before we started:

1.The air is pretty dead coming off the prop root and there is almost no pressure at the ducts beside the spinner at ground idle, especially with the 2.2 to 1 gear ratio between prop and engine. Consequently ground cooling was a big issue.

2. On the RV6, no amount of cowl flap depressed into the prop blast made any difference in the pressure differential or airflow through the rads at ground idle.

3. 50-50 Glycol/ water is not the best for cooling. Use the highest proportion of water possible for your climatic conditions and add some Redline Water Wetter. This gives the best possible heat transfer and a reasonable boiling point with a 12-16 psi pressure cap. For winter operation, increase the proportion of glycol as required. 25% glycol is good to about -16C, 33% to -20C.

4. Install drain cocks in any low points in the cooling system to purge trapped air and make sure your header tank is mounted as high as possible in the system. Trapped air will cause many problems!!

5. Tightly duct all heat exchangers on the inlet with a smooth transitioning difuser shape. Don't mount rads at 90 degrees to the airstream unless you fit guide vanes. This causes drag and is inefficient. Provide exit areas at least equal to the inlet area.

We have ended up with 4 rads situated around the cowling and baggage bay as space and ducting permits. We can climb at 85 knots and 35 iches/ 4600 rpm on a 25C day without coolant temps exceeding 90C. This is not an elegant setup but works well now after many changes. Our new RV10 will have a dedicated belly mounted rad, duct and in-flight adjustable exit door based on the lessons learned on the RV6A.