What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Engine cooling

elippse

Well Known Member
Guys, I have to tell you! These airplanes we build are gorgeous creations, and give us lots of bang for the buck! Way back when, Bellanca used to crow about how it got 135 mph on 135 horsepower, and now here are -6s doing 193 mph on 150 HP; Augustus would be mightily impressed! But for all of that, with the go-fast streamlining on these planes, the cooling systems are reminiscent of something the Red Baron would have had on his Fokker! After going to all of the trouble to build this to-kill-for plane, and then totally ignoring the air flow in and out of the cowling is montstrous, to say the least. So many take the output of their oil cooler and just dump it in front of the firewall and hope it will find its way out. Basically the same can be said of the engine cooling air; just have an inlet and place some baffles so that most of the air will find its way through the fins on the cylinders, and then let it find its way out the back! Why bother putting fairings on the gear struts and pants on the wheels? Just use the same thing the venerable old J-3 did and don't even bother with a cowling. Just let the cylinders hang out in the air! If you're going to go to all of the work and time and effort to build this thing of beauty, please at least take the time to put in a low drag cooling system.
Do you know how a jet engine works? It takes in air at its inlet, heats the air up, and then takes this increased volume of air and exhausts it out the back. When you heat the cooling air with your engine and oil, you've created a jet engine, not a very efficient one, but one just the same. When you allow the cooling air to find its way out the back through struts and wires, you take a lot of energy out of it that increases your overall drag. Have you noticed those dropped, streamlined sections on the bottom of the cowling of the Malibu and Cessna Corvallis, nee Columbia 400, nee Lancair ES? That's a duct that takes the air right from the bottom of the cylinders and conducts it smoothly out into the airstream. Each 1% extra speed you get, say 2 mph, is like having 3% more power, 4.5-5 HP! So please, do some experimenting with your inlet and outlet ducting. Make youselves nice expander ducts on the inlet-side to take the air, slow it down to increase its pressure, and then guide it into the cylinders.
And also do some experimenting with the use of your exhaust with an augmenter to help propel the air out the back. I've come up with some very short augmenters that use the Coanda principle to get the air on its way, and I'd be happy to share the drawings with any of you for free. I'd like to see you guys come all the way into the 21st century with your planes. When one or more of you more-daring individuals puts one of these together, you're going to say "This thing will never work!" and then you'll be amazed when you see it, cause everyone who's ever made one had to put the output of the Shop-Vac into it to see if it would work or if they wasted their time.
 
Hey Paul - I must admit being "guilty as charged" in terms of my very stock RV cooling system.

Problem is, I don't have any good ideas on how to get the exit flow routed better. I ran a simple compressible duct flow analysis to see how much momentum flow energy might be available and it wasn't much, by my numbers. But a few of my assumptions may have been off.

Other than doing things like running a plenum; annular inlets; inlet diffusers; good sealing everywhere - what can we do other than maybe variable exit area to help? You have me considering a nozzle to help direct oil cooler exit flow, but that's all I can conjure up... thanks.
 
losing my cool

also guilty of a stock J3 system, and wishing the air found a much smoother path out of the cowl, at least I wish for Vetterman exhaust bumperizer.

Once I've totally encased my engine in a cold air pressure intake plenum, and covered the outlets and exhaust with carbon fibre ducting, how the heck do I inspect, service, or preflight my engine?
I won't be able to see or get to a thing, so when something starts leaking a little, I won't know about it...until it's too late.

sounds good, but for now, I like to see my engine, and will put up with the drag.
 
Paul, Have you considered communicating with Larry Vetterman on this matter? He has some experience with testing cooling airflow on RVs.
 
Dave Anders did this to his RV4 from 1992 to 2000. And increased his speed to 261 mph. Or econo cruse of 190 mph on 4.5 gph. Here are his notes...................

http://sacrvators.com/Aircraft Efficiency N230A.pdf

I talked to Dave Anders at Oshkosh several years ago, and I made some photos of his exhaust augmentor. I wish now I had made notes. I'll see if I can find those photos.

Paul, I'm not far enough along to take you up on your offer, but this area has been of interest to me ever since I read how much drag the cooling airflow contributed to the total drag of the typical single engine airplane. I think a lot of Dave Anders's improvements came from Professor Raspet's work at Mississippi State. However, I do remember Dave saying some of his findings ran counter to some of the published literature on the subject. For example, he found that the exit area for his highly modified RV-4 should be about 75-80% of the inlet area, rather than >100% as some had published before. But remember he was at that time running an augmentor (not to be confused with a Piper or Cessna "Augmentor" of different design and purpose).

I like what DanH is showing us. It reminds me a lot of what Dave Anders did. Dan's workmanship is impressive and his exhaust system just looks like it should work!

I have read that Dave has recently altered his exhaust system to compete in the Sound Abatement contest sponsored by NASA. So I don't know what condition his exhaust system is in now.

Interesting stuff. I believe there is lots of potential drag reduction in improvinig the RV cooling airflow; maybe you should write an article for publication with your thoughts.
 
I talked to Dave Anders at Oshkosh several years ago, and I made some photos of his exhaust augmenter....

My arrangement is not currently an augmenter, but you never can tell where things may lead.

75-80% of the inlet area, rather than >100% as some had published before.

Novice studies of theory say there is no golden number....it varies with several factors. And on the practical front, consider that the average RV cowl is a leaky sieve, ie it has "exit area" all over the place, most of it wrong.

.....exhaust system just looks like it should work!

I only picked the 4 into 1 because its shape compliments efforts to increase cooling exit velocity....the tubes are mostly parallel to airflow.
 
I've come up with some very short augmenters that use the Coanda principle to get the air on its way, and I'd be happy to share the drawings with any of you for free.

Paul,

I'm a ways off yet from this part of the build, but I'm very interested in your thoughts and your drawings. Send them to me, please!

I have already talked about this subject with a couple of local builders. I think the awareness is rising a little bit.

Please keep your thoughts coming to this forum, too! Thank you.
 
Paul,
I'd like to see some drawings, read your ideas on this, too.

Glenn Wilkinson


P.S.
Thinking of "Bluff Bodies" instead of a central, single cowling exit
 
Exit venturi

A few years ago on my BD-4 I fabricated a dual venturi system for the exit area with the idea of entraining the heated exit air in the slip stream at the venturi neck due to the low pressure zone.

it was simple: fabricated an exit ramp of about 30 deg whose projected area was about 3/4 the exit area. I fab'd a fiberglass shroud around this ramp such that the freestream would be accelerated through the venturi with the low pressure zone at the aft 1/4 of the cowl exit area. It wasn't pretty at all but my goal was looks.

Results: 40 deg cooler CHT under similar conditions (dropped from 390's to 350s) and increased speed under similar conditions from 168MPH to 179MPH.

I will admit a couple of things could account for the speed increase.

1. The exit air was accelerated in entrapped venturi freestream

2. My BD-4 cowl exit was a flat opening flush with the bottom of the fuselage exposing some firewall flat area to freestream. The exit ramps alone might simply have more smoothly diverted the air under the exposed firewall reducing the drag. Dunno for sure...

Here is a pic: If you look closely and follow the nose strut up you can see the shroud of the left venturi (its the white box)
11glyt3.jpg
 
Results: 40 deg cooler CHT under similar conditions (dropped from 390's to 350s) and increased speed under similar conditions from 168MPH to 179MPH.

QUOTE]

Incredible! That's a 6.5% increase in speed, which is like having 21% more power! That is truly awesome! It just goes to show the drag decrease that can be accomplished by more attention to the cooling system. Not everyone can expect your results, but by having a more efficient means of getting air into, around the engine, and out of the cowling, not only will the engine run cooler, that will allow the inlets and outlets to be reduced in size.
 
I have increased speed by modifying cooling air system

Calling Bob Axsom! :D

I have been able to increase the speed of my RV-6A by modifying the cooling air system for the engine and other mods. It is not simple and many efforts resulted in speed losses rather than gains. Because of the supposed large gains possible cooling drag reduction was my first effort. The results of all the mods I have made are cumulative and do not have the same baseline speed and the speed gains are not directly comparable for effect on speed.

My baseline stock speed included 1.5' wing span extension for 17 gallons of extra fuel. Using the three leg US Air Race Handicap procedure at 6,000 ft density altitude and the the speed was 170.67 kts. Later plugging the three leg averages into the NTPS spread sheet confirmed the speed - under under most wind conditions the straight average approach of the USAR handicap procedure will underestimate the speed. I varied many things in my airplane's engine cooling air flow and one combination gave me a 4 kt increase in top speed or 2.3%. The elements of this configuration did not incrementally increase the speed. I installed an aluminum panel (made up of many small pieces put together with dimpled platenuts and #8 flathead screws) from the rear of the engine to the bottom of the firewall so the the bottom end of the forward surface is tangent with the direction of flight. A 3/16" by 7" vent on the back side of this panel below the fuselage was provided as an exhaust path for ports in the plenum (the lower surface of the cowl is the top closure of the plenum) such as the oil cooler and blast tubes.

On the front of this curved panel I installed at angles determined by the staggered rear positions of cylinders 3 & 4 a vertical aluminum panel with rubber outer edge seals from each side of the cowl to the edge of the of the cooling air outlet channel in the lower cowl. A test in this configuration showed a decrease in top speed of ~2kts.

I then installed horizontal aluminum baffles with rubber outer edge seals outboard of the valve covers and standard baffling (sealed to them with high temperature RTV) just below the upper edge of the lower cowl. This isolated three chambers or zones in the cowl. Air could no longer pass freely between the zone below the engine and the zone behind the engine. The test in this configuration showed a 6 kt increase over the previous configuration and 4 kts over the baseline speed.

I have reduced the inlet area incrementally by 1/4" vertical plugs that extend past the spinner rear edge to just clear the prop at maximum pitch with a shape to distribute the air above and below the cowling and out of the cooling air inlets and flown many tests that show a direct relation between reduced inlet size and increased cylinder head temperatures but no increase in speed.

I have tried a cover plate over the top of the recess in the filter air box (FAB), a cover over the opening at the top of the web structure (inverted pyramid) in the nose gear mount structure, an aluminum baffle to close the path from the lower cowl to the opening in the front of the cowl and a lower baffle to turn the air below the engine and take it all the way to the back of the cowl above the FAB and the NLG mount web structure. All of these changes showed a reduction in speed. I made racing cover plates for the heater air and blast tube openings in the read of the plenum and was able to measure a slight increase in speed from that change (I don't have the number handy).

Larry Verterman's experiment with the fairing at the rear of the lower cowl and alternate co-location of the exhaust pipes and the cooling air outlets is impressive.

Sarah Beam has augmenter tubes on the exhaust of her Glasair III but I don't recall any claims of speed gain.

Bob Axsom
 
Last edited:
Seems to me there are only two fundamental goals. Everything else leads back to them.

Maximize heat transfer efficiency, ie raise the cooling air temperature as much as possible during its pass through the system.

Maximize exit velocity. Exit velocity below freestream is drag.

I'm also convinced cooling drag reduction requires a systems approach. A single mod alone, without complimentary work, will often net nothing. For example, consider this excerpt from Bob:

I have reduced the inlet area incrementally....and flown many tests that show a direct relation between reduced inlet size and increased cylinder head temperatures but no increase in speed.

That's been a common observation after simple inlet size reductions. Makes sense; it results in a reduction of cooling air mass. Naturally that raises CHT, which is both bad and good; high temps are not welcome but increase transfer efficiency due to better deltaT. The exit velocity effect is more subtle. If you flow less mass but maintain the same exit area, you reduce exit velocity and increase drag. Net result is no gain in speed.
 
The outlet

The outlet continues to be an area of interest to me but nothing I have done there has contributed anything to speed yet. I have created a "pinched" area there, back when I was experimenting with the "turn back" second lower cowl baffle, between the curved surface of the upper/rear curved baffle and the opposite curved surface at the rear end of the lower baffle. I thought the air would accelerate through the smooth surfaced pinched exit space but the plane slowed down. I rationalized that the the airflow attached itself to the lower surface and injected itself into the previously undisturbed air below the airplane and increased the drag. I remain convinced that this is an area that one can work with to improve speed (reference Larry Vetterman's fairing and alternate co-located cooling air and engine exhaust location) and I have a lot of ideas in my mind but none that I am ready to implement yet. Any change I make has to be removable and not detract from the airplane in any way.

Bob Axsom
 
Outlets

I have two rectangular 2.5" by 5" ducts, one on each side of the plane, that carries the combined cooling-exhaust flow from the two cylinders on each side. They come out right below the firewall on each side, and the fuselage immediately behind each of them follows the slant of the ducts and is lined with RTV-covered fiberfrax for about 10". These ducts each have a variable outlet that can be be opened to 4" by 5" for takeoff and climb, then closed to 1.5" by 5" for cruise. That's 15 sq. in. outlet area vs 12 sq. in. inlet area which also includes the exhaust flow. When I close them I pick up about 5 mph. That's 2.5% speed increase, which is equivalent to having another 7.7% or 9.6 HP! There's also an additional 2 sq. in. inlet and outlet for the oil cooler.
Keep in mind that the airflow during takeoff and climb must be greater because of the higher power, and because of the slower airspeed, the inlets and outlets must be larger. But when you get to cruise at higher altitude, you usually have cooler air which compensates for the loss of cooling from a rich mixture, and higher forward speed which increases the total flow through the inlets and out the back. So by having a cooling system which extracts the maximum amount of engine heat for a minimum amount of air flow, you will have the least drag. Because I guide the air in, through the engine, and out the back, my inlet area is quite small for the installed horspower. 12 sq. in. plus 2 sq. in for the oil cooler for 125 HP at 200 mph says that if you have 160 HP and go 190 mph, you should be able to get away with 18.9 sq. in., 9.5 sq. in. per side, or about 4.3" by 2.2" and 23 sq. in. outlet area, say 3" by 8", in cruise. Why not use these real numbers as a comparison to your own and maybe a goal to shoot for.
 
Note to all....nothing in Paul's cooling setup resembles anything found in your stock RV. Take a look here:

http://www.eaa.org/experimenter/articles/2009-09_lipps.asp

The cylinder wraps are a good example of maximizing heat transfer efficiency. Air temperature taken at the underside of a cylinder is likely quite a lot higher than with typical baffles. Paul, have you ever taken any temperature data (ambient vs cylinder underside plus CHT)?

Interesting system. Do you think the tiny inlet system will work without being exhaust-driven? What sort of climb speed and ROC are typical to maintain what CHT?
 
Wow, how foolish I feel for suggesting Paul write an article for publication. He already has. See the link in DanH's previous post. Thanks, Paul. There is no doubt in my mind that you know what you're talking about. (Here's the link to Paul's article on propeller design on EAA's web site.) I saw Tom Aberle's biplane at Reno with your propeller design. I believe his lap speeds were nearly 100 mph faster than the closest competitor in that class.
 
Last edited:
What do you think would be good for a 180?

I have two rectangular 2.5" by 5" ducts, one on each side of the plane, that carries the combined cooling-exhaust flow from the two cylinders on each side. They come out right below the firewall on each side, and the fuselage immediately behind each of them follows the slant of the ducts and is lined with RTV-covered fiberfrax for about 10". These ducts each have a variable outlet that can be be opened to 4" by 5" for takeoff and climb, then closed to 1.5" by 5" for cruise. That's 15 sq. in. outlet area vs 12 sq. in. inlet area which also includes the exhaust flow. When I close them I pick up about 5 mph. That's 2.5% speed increase, which is equivalent to having another 7.7% or 9.6 HP! There's also an additional 2 sq. in. inlet and outlet for the oil cooler.
Keep in mind that the airflow during takeoff and climb must be greater because of the higher power, and because of the slower airspeed, the inlets and outlets must be larger. But when you get to cruise at higher altitude, you usually have cooler air which compensates for the loss of cooling from a rich mixture, and higher forward speed which increases the total flow through the inlets and out the back. So by having a cooling system which extracts the maximum amount of engine heat for a minimum amount of air flow, you will have the least drag. Because I guide the air in, through the engine, and out the back, my inlet area is quite small for the installed horspower. 12 sq. in. plus 2 sq. in for the oil cooler for 125 HP at 200 mph says that if you have 160 HP and go 190 mph, you should be able to get away with 18.9 sq. in., 9.5 sq. in. per side, or about 4.3" by 2.2" and 23 sq. in. outlet area, say 3" by 8", in cruise. Why not use these real numbers as a comparison to your own and maybe a goal to shoot for.

A lot of good info in this thread. One of my good Canadian friends and I have discussed this over the past couple of years and he has implemented one or two concepts. I have only thought about it. What would you recommend for a 180 HP inlet and outlet for low altitude and high altitude racing?

Bob Axsom
 
The cylinder wraps are a good example of maximizing heat transfer efficiency. Air temperature taken at the underside of a cylinder is likely quite a lot higher than with typical baffles. Paul, have you ever taken any temperature data (ambient vs cylinder underside plus CHT)?

Interesting system. Do you think the tiny inlet system will work without being exhaust-driven? What sort of climb speed and ROC are typical to maintain what CHT?

I don't know what the air temp is at the bottom of the cylinder since my exhaust comes in only about 2" below the cylinder to mix with the cooling air. I think the combined temp at the outlets is 450F-550F at WOT. Even though my Coanda augmenters are at each cylinder, the single augmenter I designed for Kevin's "Relentless", which some of you might have seen at Oshkosh two years ago, was driven by the outputs of his dual turbos. When he first welded it up, his curiosity got the better of him and he used the Shop-Vac output to see if it would really work, and saw how a piece of paper was drawn into it!
That's why I designed these two Coanda-effect augmenter ducts which can be used with either 1.5" exhaust or 1.75" exhaust. I'd like for all of you to be able to make use of these to help extract the cooling air from your cowlings to not only help the flow but to reduce drag with smaller inlets and outlets sized to the real requirements. As far as CHTs go, I have the plug-in style on the bottom of the cylinder. I can idle for extended periods of time on California hot days with no overheating, and my CHTs run about 370F-400F in a climb on a hot day. My take-off climb is at 85 mph IAS, best angle, to pattern altitude, and then at 1350 lb it's about 1500-1550 fpm at 105 mph IAS. For a trip I usually climb at 500 fpm to 11,500' or 12,500' at 150 IAS initially then down to 130 IAS as I get to altitude. I have one cylinder with an intermittent CHT probe that will jump back and forth from 380F to 420F, but two others are usually about 360-380 and one at 340-350. I think my fiberglass intake manifold in the carbon fiber pan is not as good a fuel distributor as I'd like! I asked one of the forum members to contact me at my e-mail address and I'd send him the drawings if he would put them on the forum. I still haven't the knack with pix!
 
Paul, you don't have an email link in your VAF profile, but if you click on my name above you'll find mine. Shoot over the drawing and I'll post it here.
 
Paul-

I reviewed your article that Dan cited. Have you taken manometer readings that establish that the left and right ducts operate at the same pressure above the cylinders?

The reason for my question is to learn what affect the "corkscrewing" of the air through the propeller has on the mass air flow into the inlets, which appear to me to be symmetrically sized.

Thanks

LarryT
 
Paul-

I reviewed your article that Dan cited. Have you taken manometer readings that establish that the left and right ducts operate at the same pressure above the cylinders?

The reason for my question is to learn what affect the "corkscrewing" of the air through the propeller has on the mass air flow into the inlets, which appear to me to be symmetrically sized.

Thanks

LarryT

One of the big mis-conceptions is that the air coming off the prop disc is at a really high speed. When I'm clipping along at 200 mph, 293 fps, the axial velocity increment, vi, is about 5 fps, and the tangential velocity, ut, is about 7 fps. That says that the angle of the total velocity is about 1.3 degrees, not much of a corkscrew! Because we have such a high velocity coming off the prop static at WOT, we tend to maintain that memory and think that is what the prop is doing in flight. That's why so many ask me that if the prop is producing thrust in the root region, won't that make the fuselage drag much higher? But at very low speed there will be more of this twist effect. Keep in mind, the thrust is the mass through the disc times vi. At 100 mph, the mass flow will be about half, and vi and ut will have to be twice as much for the same thrust, so the angle will be about 5 degrees. 'Course the thrust isn't the same because the drag will be about 1/4, but that was just an illustration. Now those values of vi and ut were for a three-blade, and a two-blade will be about 50% more because of the lower mass flow with fewer blades. That's also why multi-blade props are so much more efficient at low speed because they have greater mass flow and less energy is wasted in vi and ut, giving better ROC. But properly-designed multi-blade props are not less effcient in cruise, another myth!
 
...don't know what the air temp is at the bottom of the cylinder since my exhaust comes in only about 2" below the cylinder to mix with the cooling air.

Darn.....would have been a good data point. I'd love to know if there is any significant difference in temperature rise for a full wrap like yours and the gasketed baffle I'm using.

I installed a screw-type junction block on the firewall and ran 6 general purpose wiring leads through the sealed firewall pass-through for later use. The junction block will allow easy, temporary installation of various engine compartment pressure and temperature sensors. I have two unused temperature inputs on my GRT EIS which may be used as instrumentation. Or I can hook to a logger of some kind. Or a steam gauge and a clipboard ;).

I can idle for extended periods of time on California hot days with no overheating, and my CHTs run about 370F-400F in a climb on a hot day.

Tells the tale....a powered cooling system works.
 
I installed a screw-type junction block on the firewall and ran 6 general purpose wiring leads through the sealed firewall pass-through for later use. The junction block will allow easy, temporary installation of various engine compartment pressure and temperature sensors. .

I did the very same thing! I have a terminal block on the firewall and one on the back of the instrument panel interconnected with a cable with twisted-pairs and an overall shield. So far I've used it once, but it was worth having it!
 
Posted for Paul....

Postscript: Paul tells me there are errors in the original drawing dimensions below, so be aware. I'll post new drawings when Paul sends them.
_____________________________

1.5" and 1.75 " Coanda Exhaust Nozzles

These Coanda exhaust nozzles are used to entrain and eject air surrounding the nozzle. They do this by means of the Coanda effect which creates a low-pressure region where the exhaust gases emerge from the pipe on each side and follow the curve to the rear. It is actually the air surrounding the exhaust-gas jets that pin them to the rearward curved surfaces. This is the effect noted by Henri Coanda who first defined it. The flow of air attracted by the low pressure then mixes with these gases and is accelerated to the rear. The dimensions of the two nozzles are to allow them to be easily attached to a present 1.5" or 1.75" exhaust pipe. This can readily be done by the use of a piece of similar diameter pipe 1.5" long slit lengthwise and clamped over the last 3/4" of the exhaust pipe and over the 3/4" long round portion of the transition. This transition is made from a 1.5" long piece of 0.035" wall exhaust tubing that has one of its ends formed into a rectangular section to which the 0.035 thick sheet metal, that makes up most of the nozzle, can be welded. This can be readily accomplished by first marking on the tube where the corners of the rectangle will be from the drawing dimensions, placing a piece of metal with a hole of the same size, such as an exhaust flange, over the tube in its center, and then squeezing the sides of the tubes between the marked corners with pliers to flatten the sides. The use of the flange is to preserve the round shape of the forward portion of the tube while the rear portion is being formed. The three curved sections that make up the nozzle that guides the exhaust gases out and then rearward are made from 1/4 round sections of 0.035" wall tube. Once the rear section with its four forward curved-sections and the two long formed-curve sections that make up the trailing portion are assembled, the assembly is attached to the forward portion by means of the extended plates that form the sides of the nozzle from the transition to the trailing edge. These extended sides seal off the sides of the nozzle and the inside of the airfoil-like trailing wing. These extended side-plates are shown with broken lines on the drawing, and are 1/4" to 1/2" larger than the wing section. It goes without saying that, except for a model to demostrate how it works, the real thing should be formed of exhaust-grade 0.035" wall stainless steel. All welds where exhaust flow will take place must be ground smooth to prevent flow turbulence and separation.
This nozzle may be used as the driver for an augmenter tube and will result in more efficient energizing of an engine's cooling air since the Coanda effect, with its low-pressure region, causes air flow to move toward it immediately at its output, whereas a typical augmenter, fed by a round tube, can only cause a low pressure to be formed downstream of the exhaust by the mixing interaction of the two gases. This is why an augmenter tube needs to be at least five times as long as its diameter in order to provide sufficient length for this interaction to occur. Using this nozzle, an existing augmenter will perform much better, or the augmenter tube may be made as short as is the Coanda nozzle. It also may be placed just inside a cowling outlet to maximize cooling air flow. It should be noted that there will be a great deal of heat where the exhaust emerges, both of the conductive and radiative variety. The curves that cause the gases to turn will get extremely hot and give off infrared radiation which can heat any surface in the vicinity. For this reason, if used at a cowling outlet, especially with composite construction, it is best to encase the nozzle in an aluminum tube conforming to the outlet shape. This tube will remain fairly cool since it will be cooled by the engine's cooling-air flow. Any tube such as this should be provided with a generously-curved bellmouth on the inlet edge to help the flow of air that comes in to it tangentially. The outlet area of the two jets was kept the same as the cross-sectional area of the exhaust tube so that it should have a minmal effect on exhaust back-pressure.
An air-cooled engine will typically require a total cooling air inlet area, engine and oil, that can be calculated by dividing the horsepower by the WOT TAS and multiplying this by 20 to 30. So for 160 HP at 210 mph you need 15 to 23 sq. in. total cooling inlet area. Since the cooling air will be heated about 120F, increasing its volume, the oulet area will have to be increased by the ratio of the inlet and outlet absolute temperature, or (520 +120) / 520 or about 20%-25%. This will give an outlet velocity of about the same as the forward velocity to minimize cooling drag.



 
Last edited:
What would you recommend for a 180 HP inlet and outlet for low altitude and high altitude racing?

Bob Axsom

An engine requires a certain mass flow to carry off the heat. It is pretty much proportional to power. Think of the mass flow as the inlet volume per second, a tube with the cross-sectional area of the inlet, and the length is the speed in fps, 22/15 time mph. So the faster you go, the less inlet area you need for a given power. The numbers I gave for mine are reduced to sea-level power and speed, which really doesn't matter for a normally-aspirated system, since as you go up in altitude the power and speed drop off almost proportionately. My sea-level speed is 213 mph at 2950 rpm which is about 131 HP. Using my inlet's 14 sq. in. total inlet area, engine and oil, for calculating the required area for 180 HP at 210 mph at sea-level would give 14 X 213/210 X 180/131 = 19.5 sq. in., about 4.5" X 2.25" each. Now that's with a very efficient cooling system. You might want to start off with about 50% more area about, 2.75" X 5.5" and see what your CHTs are; I like mine to be about 370F-390F for efficient engine operation; 400F to 420F would be even better! Because of the temperature rise in the cooling air, the output volume will be about 20% greater, so maybe your total outlet area should be about 20% greater, a minimum of 23.5 sq. in. to 35 sq. in., or about 3" X 8" to 4" X 9". Does this make sense? An augmenter makes up for the pressure drop across the engine to maintain the flow, whereas you normally use a larger output area to reduce the pressure to get the flow, with its resultant drag!
 
Paul: I'm a bit confused. Are the Coanda nozzles and the augmentor(s) you are talking about, the same thing? I'd like to clean up the airflow in the tight space aft of the 540 in my Rocket, but I'd rather fly it than work on it, to tell the truth!

I'd still like to get my oil temps down a bit as well, but a little less drag wouldn't hurt my feelings either. For those of us who don't want to change our cooling air inlets (aesthetics), what one or two things internal to the cowling and cowling exhaust area do you believe would have the greatest payoff from a drag reduction/cooling efficiency perspective?

Thanks!

Lee...
 
Paul: I'm a bit confused. Are the Coanda nozzles and the augmentor(s) you are talking about, the same thing? I'd like to clean up the airflow in the tight space aft of the 540 in my Rocket, but I'd rather fly it than work on it, to tell the truth!

I'd still like to get my oil temps down a bit as well, but a little less drag wouldn't hurt my feelings either. For those of us who don't want to change our cooling air inlets (aesthetics), what one or two things internal to the cowling and cowling exhaust area do you believe would have the greatest payoff from a drag reduction/cooling efficiency perspective?

Thanks!

Lee...

An augmenter is a device that, of course, augments, or helps along, but in this case it is the use of high energy exhaust gases to help overcome the pressure drop through the cylinder fins. A typical augmenter is a large diameter tube with an exhaust pipe in the center that makes use of the mixing of the exhaust flow with the air in the tube to expel it out the back and create a low pressure region downstream of the pipe to draw in more air. Typically they have a length-diameter ratio of at least 5 to be effective; the longer, the more interaction between the two gases. The Coanda nozzle uses the Coanda effect to create a low pressure right at the discharge jets where the exhaust gases curve out and around, so it can be more compact than the tube-type augmenter. These designs can be mounted on the exhaust pipes just inside of the cowling outlet.
As far as the inlets are concerned, one of the ways of reducing the inlet area is to fill in the inlets with well-radiused curves along the edges of the aperture. That way you not only decrease the inlet area but you help to guide the flow in and guide the excess flow out and around. A true experimenter places aesthetics a far second, to the extent that their plane will always be a W.I.P. i know of at least one example of this!
 
Temperature sensors

For those who want to measure the cooling air at the bottom of the cylinder might I recommend the National LM135,235,335 familyof integrated citrcuits. They put out an analog voltage scaled to 10mV/K(C) over a range of -50C to 150C, and they only require an 11k resistor from 14V and return to power them. The output is easily read on a DVM and 20C would appear as 2.93V, corresponding to 293 deg. Kelvin. For those who want to measure pressure diffentials, the Motorola/Freescale MPXV5004 DP which is available for about $20 has a range of 1.5 psi full-scale and is fully temperature compensated and has an output of 0.5V to 4.5V for the pressure range. The DP in the nomenclature indicates differential pressure and has two inputs barbs, one for static and the other for the pressure. It is powered from regulated 5V which can be obtained from a regulator such as a 78L05 which can power several of these devices.
 
The drawing needs clarification...

...especially the side view.

The Coanda drawing top view is understandable.....the side view is very confusing.

Thanks,
 
I'm with Pierre on the drawings, i don't quite understand exactly how they work. Are they like a nasa duct inlaid into the exhaust tube?
 
Possible drawing clarification

I'm with Pierre on the drawings, i don't quite understand exactly how they work. Are they like a nasa duct inlaid into the exhaust tube?

The forward, pointed portion of the left part acts as a flow divider which causes the exhaust flow to divide into an upper and lower jet. As this jet flows outward, Coanda effect causes it to follow the rearward curve back over to the trailing edge. What makes it go around that curve rather than going straight out? The Coanda effect! What this amounts to is that there is no air between the jet and the surface pushing out on the jet to counteract the push of the outside air against it, so it stays attached, even though it looks like it shouldn't. It's actually the same thing that makes the air follow the curvature around the leading edge of the wing, and why the air pressure is the lowest where the LE radius is the smallest. Here's a good illustration of Coanda effect. Hold the outside surface of the bowl of a spoon into the flow of water coming from a faucet. What you will see is that the flow follows the surface and flows off at an angle to the original flow. What you will also see is that the spoon will move into the stream, not away from it. That's counter-intuitive, right? You would think that the pressure of the water against the bowl would push it away from the stream. What causes that? Suction? No! There is no force called suction, only pressure differential! It is the higher-pressure air on the other side of the bowl pushing it into the water stream since there is a lower pressure outside of the curved stream. (BTW, that is what constitutes 2/3 of the total lift of an airfoil. The other third comes from the bottom of the airfoil) That is the low pressure that the exhaust jet will create to cause the cooling air to move toward it! The reason there is a low pressure is that molecules of air moving toward the surface by Brownian motion are not replaced by molecules moving in the opposite direction, so there is a net flow in one direction. I have a friend who's brother had a loose latch holding the hood closed on his car. My friend cautioned his brother not to drive the car until he fixed the latch, as the hood would open up. His brother poo-pooed that suggestion as he knew that the air would hold it shut by pushing down on the curved hood. Bad error in judgement! Once he got up to about 60 mph that hood popped open from the higher pressure air inside the engine compartment pushing out the hood out into the lower pressure on the curved surface. The brother almost crashed because the hood wrapped around his windshield making it impossible for him to see ahead!
 
How is the pointy-looking left part....

....attached to the right side..because it appears that they are two separate pieces?

Thanks,
 
Side plates?

Pierre,
I guess the side plates are here to keep the pointy thing in place and hold it together with the right part of the drawing.
My 0.02 ?
 
Pierre,
I guess the side plates are here to keep the pointy thing in place and hold it together with the right part of the drawing.
My 0.02 ?

That's it! 'Sorry I wasn't more descriptive in my write-up! The side plates extending beyond the outline of the left-hand portion, as shown by the dotted lines, also keep the air from coming in at the sides to destroy the low pressure, sort of like tip plates on a wing!
 
It is a bit hard to grasp. Try these:



Exhaust gas is red, cooling air is blue. The nozzle would be inside a rectangular cooling air exit duct. Entrainment at Point A greatly increases velocity at exit Point B.




Got it right Paul?
 
Last edited:
It is a bit hard to grasp. Try these:



Exhaust gas is red, cooling air is blue. The nozzle would be inside a rectangular cooling air exit duct. Entrainment at Point A greatly increases velocity at exit Point B.




Got it right Paul?
Excellent 3D and flow visualization , Dan! Thanks so much for these illustration to make it clear to others! A job well done!
 
Sound dampening?

Paul,

Did you study if your system would also have a sound dampening effect?
I've read somewhere that the noise of big jetliners has been reduced by "encapsulating" the hot and noisy flow from the jet with a cold air flow.
Dan's illustration of the different flows just reminded me that.
If that works as I think, it would be a double winner for us in Europe where noise regulations are getting harder every year.
Thanks for sharing your brilliant ideas with us, anyway. Every time I click on "New Posts", I hope to read some additional info from you or Dan who also brings a lot of facts based knowledge to this forum.
Keep your brains up, you're feeding us.
 
Paul,

Did you study if your system would also have a sound dampening effect?
I've read somewhere that the noise of big jetliners has been reduced by "encapsulating" the hot and noisy flow from the jet with a cold air flow.
.

It does have some sound dampening effect, but probably not any where near enough for the European standards. Have a look and listen to the video posted on YouTube that Pat Panzera of Contact! magazine made last year. It is under "Paul Lipps FlyBy at Hanford Ca". What bystanders tell me is that it takes out somewhat of the sharp bite of un-muffled exhausts and it is different but still noisy. See what you think. My hangar is not far from the runway at Santa Maria, Ca. and I can always tell when an RV with 160-180 HP, of which we have a few, takes off!
 
My exhausts stick out....

....quite a bit on the -10 and I suppose that all of this device should be inside the cowl. Not enough room in my airplane, I'm afraid.

Dan, thanks for the three-view...very clear,

Best,
 
Hey Paul:

Dan's drawings let me finally understand what the heck all this is about. Can you tell us what the device on Kevin's NXT looked like? That design might be easier fitted to our planes...

Thanks!
Mark
 
Hey Paul:

Dan's drawings let me finally understand what the heck all this is about. Can you tell us what the device on Kevin's NXT looked like? That design might be easier fitted to our planes...

Thanks!
Mark

That design took the output of the two turbos which came in, one on each side, through 90 deg pipes and then into a 10" long 3" tube that was open on the front and with a divider in the center to keep the exhaust from either side trying to go straight through. Then from the front opening it curved around to the rear. Keep in mind this was pretty large as this was for an TIO-540 operating at over 50-some inches of MAP. I guess it could be scaled down for your applications, but I thought that this was where these smaller ducts might work better just inside the cowling. The ducts from my cylinders are at least 15" long from the exhaust port to the firewall, so I thought that you could fit something like the ones in the drawing in that much space inside the cowling with no problem.
 
.... I suppose that all of this device should be inside the cowl. Not enough room in my airplane, I'm afraid.

The true nozzle Paul has sketched for the end of a pipe would indeed require reworked headers, but you could probably get two of them inside the current cowling exhaust ramp and split the outlet area into two small "blown" exits.

Don't want to modify too much? Here's an alternate idea. Nothing says the Coanda principle must be used in a symmetrical nozzle. Heck, the most common application is the blown flap.

I'm seeing an appropriate shape (in stainless) grafted to the fuselage belly just behind the exit, and a small hinged extension to the cowl exit ramp.

 
And use 4 straight pipes, side by side???

Just throwing out a conceptual sketch Mike. It may work with straight ends and may work better with fan nozzles on each pipe. Four, two or a single tailpipe? The "package" concept allows grafting to what you have. It should also simplify in-shop modeling/testing without fooling with the airplane.

Whatever works....it's all about exit velocity.
 
Concept

Dan; it's nice to see that some fertile minds have been stimulated by this post. What you propose will work. You might want to invert it so that the cooling air is close to the fuselage and the exhaust on the outside. It looks somewhat like the one I designed for Bruce Hammer which he powered from two pipes on each side. But his was also somewhat like yours with the exhaust side next to the fuselage and he was getting lots of heating on that stuff his plane was made out of and removed it. What do you call it? Compost or something like that, not good old aluminium!
Since my outlet ducts are over 15" long, the exhaust and cooling air get pretty-well mixed together before they exit below the firewall but tangent to the bottom of the fuselage. I coated a 5" wide by 10" long area behind the duct with high-temp RTV, covered it with fiberfrax, then covered the fiberfrax with a coat of the same red RTV. Seems to work OK! Bruce is one of those SARL guys who likes to race, especially if he can beat his brother Steve. He's a really nice guy, and would probably give you drawings of the one he made, or maybe even give you the actual duct! He flies guys out to the oil wells in a chopper, and I hope none of his friends or acquaintances was injured or killeds!
 
Back
Top