Liquid cooling


Help Support HomeBuiltAirplanes.com:


Well-Known Member
Feb 26, 2011
San Francisco
Some race motorcycle designer tried a meredith style cooling system about 25 years ago. The bike had many interesting features. I never saw it again.


Well-Known Member
Jun 26, 2012
Calgary, Alberta, Canada
This thread has prompted me to come out of my 1 year self-exile to comment on some of the misunderstandings and conjecture here.

First, people who've never built, flown, tested and proven their own cooling system designs probably can't be considered authorities on the subject, the same as on any other subject. Theory is proven (or disproven) by experimentation/ demonstration in the real world.

Background information. I wrote the article for Kitplanes in 2015 after 2 years of study and experimentation on the subject. I built and flight tested my radiator duct design based on bench measurements I took from my flow bench and coolant test rig to quantify pressure drop and HX effectiveness (not covered in the article for brevity)

My inspiration came from the theoretical work by Meredith in 1935, Gothert and Russell Sherwood who flies a very successful and fast (multiple SARL winning), Subaru EG33 powered Glasair with a ventral radiator setup. We both ran airflow tests on our ducts and incorporated internal guide vanes to mitigate separation prior to the HE face. This was very important to their good performance.

In flight testing, I measured exit velocity at 104.4% of the freestream velocity, validating the Meredith Effect. As far as I know, this is the only published information on this subject from flight testing. All other information I found was theoretical or simple conjecture with little basis in facts including the 2014 article from an Italian university researcher who published some unbelievable numbers, apparently totally disregarding the effects of pressure drop across the core and consequent momentum loss.

I’ll start from the top of the discussion in this thread. Copper isn’t used for aircraft radiators because it’s heavy, weak, work hardens under vibration and you have poor thermal conductivity between the soldered tubes and fins. Aluminum is superior because of these aspects.

Maximum heat dissipation is obtained by having coolant divided into thin tubes, attached to fins with high surface area and turbulence. A big chunk of copper inside a duct won’t cool much of anything.

Cooling Drag = Mass Flow X Momentum Loss- in simple terms, the more airflow you use for cooling and the more you slow this down before it exits the ducting into the free stream, the higher the drag.

Mass in = mass out. If we discharge that mass at higher velocity than we ingested it at, we create thrust.

Paul Lamar was no pioneer in aircraft cooling system design since he never built, flew or quantified the results in the real world. While his book collects a lot of previous information in one place, it contains a number of incorrect conclusions unsupported by any actual instrumented flight testing. Secondly, these sources rarely considered the drag associated with cooling or how to mitigate those losses for aircraft.

His much espoused wedge diffusers do cool but at the expense of high momentum loss and therefore high drag, since air turns approximately 180 degrees during its path through the HX. So if drag is not a concern to you, wedge diffusers may do the job for you.

Consider these facts; No military or race aircraft use wedge diffusers for ducting radiators.

The fastest liquid cooled aircraft in the world, (Voodoo and Strega) use traditional or “long” streamline type diffusers with low diverging angles to minimize separation prior to the HX face. BTW, the ducts and rads used in these two aircraft are NOT stock P51 parts.

The record setting Anequim design used a CFD developed long diffuser for its oil cooler to minimize drag, not a wedge diffuser and not a "bell" diffuser.

All WW2 and race aircraft use variable geometry exit doors to control mass flow and minimize momentum loss. Why? Because you want maximum ground cooling where mass flow is low and minimum drag at high speeds where mass flow is high. Any cooling system design not employing variable geometry exit doors will be a less efficient compromise. Exit door actuation is easy. Copy my Bowden cable design or that of Russell’s with a servo motor. Both are extensions of the P51 design.

Most any tractor aircraft with cowling mounted rads will suffer high drag in exchange for more convenient component packaging. Take your pick. You usually don’t have sufficient length to fit a proper diffuser here and exit the air parallel to the free stream, so you’ll necessarily have separation and high momentum loss.

You don’t need fans for ground cooling with a properly designed layout. Russell and I have both proven that. Fans are a crutch for poor design and cause high drag in cruise due to the disc area, blocking flow. No WW2 liquid cooled aircraft used fans- other than the big one out front.

In simple terms, heat dissipation on a given HX and water flow rate and temperature is dependent on air mass flow through it. Mass flow is dependent on Delta P or the difference in pressure between the front and rear faces of the HX. Worst scenario cooling is on the ground (very low Delta P) and in the climb (relatively low Delta P and high hp (heat). GA aircraft need to have adequate ground and climb cooling on hot days. This is an important consideration for non-race aircraft.

All modern auto engines use a pressurized cooling system to raise the boiling point. We aren’t concerned with boiling at high altitude. Not sure where that idea came from?

Like all things in aircraft design, an efficient cooling system is a compromise between idle and low speed needs and those of low drag at cruise or high speeds. We have a pretty fair idea on how to do this after flying liquid cooled aircraft for many years. No need to re-invent what was well known and understood 80 years ago.

Back to Paul Lamar for a moment. He contacted me over 20 years ago after reading my web page, warning me that my Subaru powered RV would be a failure because he believed piston auto engines would never survive such “abuse”. Only a Wankel could do the job in his words. Hmmm. Here I am along with Russell Sherwood (racing in SARL no less, at almost max power and rpm) along with hundreds of other flying Subarus- working just fine years later. Today, piston automotive powerplants in Experimental aircraft outnumber Wankel powered ones at least 25 to 1.

Paul also contacted me after my Kitplanes rad article and offered some more unsolicited advice, saying I had it all wrong. In his words “The 7 degree rule is dead”, referring to the streamline diffuser used by me and all successful WW2 and Reno race liquid cooled aircraft. Hmmm again. When I asked him why, he presented no facts to support his contention and just said to read the two oft quoted texts he loved so much.

He went on to say that “every HVAC system in the world uses a K&L diffuser” Not true, but aircraft also have VERY different application and design goals from HVAC systems so that was an irrelevant comment. He also stated that “the late P51s and B36s used the K&W diffuser”. Really? The K&W text was published in 1953. Production of the P51 started 13 years before that. Production of the B36 started in 1946, so this was all nonsense.

To his everlasting credit, Paul Lamar was a champion of the Wankel engine for use in experimental aircraft. He compiled a massive site to that end with some very useful information for those inclined to follow that path. He did some other cool projects and had some innovative ideas. His vision and help resulted in the impressive time to climb record a few years back with Wankel power- which he was never to see unfortunately. I spent many hours on his site over the years. It had some interesting content from multiple sources.

Paul wasn't right on a number of other things though and fought/ banned people who went against his advice and built things their own way successfully, proving him incorrect in the end. Confirmation bias as evidenced from my interactions with him above, was sadly a big part of rotaryeng.net whereas it could have been a much more friendly and productive melting pot of ideas from many experimenters IMO.

My advice is to look to the people who’ve actually flown their ideas rather than those who just talk about them. Real world experience is often much more valuable than untested theory.
Last edited:


Well-Known Member
HBA Supporter
Mar 26, 2012
Timaru, NZ
I don't think so. Britten's earlier bikes had a lot more aero on them, but the later ones ditched it - probably the extra weight wasn't worth it. That and he seriously injured a foot that got caught in the big fairings in a slide!