Installing a WAM-120 diesel in an RV-9A

Ray clegg & Ian Bellamy have written an article in this months "Popular flying" magazine (UK) on their experiences installing a WAM-120 diesel engine in their RV-9A.

http://www.pfa.org.uk/PF/2007/Sept/DieselVans.pdf

They have taken a more complex approach to fitting the engine than I have, as they have removed Wilksch cooling pack from below the prop and mounted the intercooler and radiator on the engine mount. It takes a but more doing, but avoids the "basking shark" aesthetics of most of the Wilksch installations.

Dave

Dave_Boxall said:

Ray clegg & Ian Bellamy have written an article in this months "Popular flying" magazine (UK) on their experiences installing a WAM-120 diesel engine in their RV-9A.

http://www.pfa.org.uk/PF/2007/Sept/DieselVans.pdf

They have taken a more complex approach to fitting the engine than I have, as they have removed Wilksch cooling pack from below the prop and mounted the intercooler and radiator on the engine mount. It takes a but more doing, but avoids the "basking shark" aesthetics of most of the Wilksch installations.

Dave

Click to expand...

Dave,
Nice looking work. I like their graphic. Their layout is the technique I want to use for my 3-rotor powered RV-10
Bill Jepson

Engineers have more work to do.

I admire these poor chaps for trying to make lemonade out of a lemon. But their mistake was taking the radiators as a given. They would have done almost as well to simply bolt the radiators to the exterior of the fuselage.

Time and time again, the single most disastrous design element of water-cooled alternative engines is the shape of the radiator itself. The engineers who design engines and then slap a square radiator on it ought to be ashamed (or quit).

An engine is but one element of a whole system (RV-9A in this example). If that engine imposes undue compromises on that system, then it is no good. To illustrate extremely: if you come up with an engine that will give you 120HP, burns only 1/10th as much of the cheapest, commonest fuel but weighs 2 tons, then that engine is unsuitable for the application.

The same goes for the radiators. Why does an engine of similar power and greater efficiency need the airframe to donate a 50% to 100% increase in frontal area (compared to a standard 9A cowling) for cooling? This fails the suitability-for-the-application smell test.

Alternative engine designers must live within the same "aerodynamic budget" as mainline engines or forego some claims to efficiency. There is plenty of room inside a standard cowl for a more imaginative arrangement of cooling surfaces that is capable of using a given amount of frontal area with an efficiency comparable to a typical installation.

A Leyland bus, indeed!

Very well said, and I agree with all of your points.

But you must remember that the liquid cooled engines have a much lower Delta "T". This means more air is required to achieve an equiv level of heat transfer.

Maybe, in fact, those engineers know what they are doing.



Franklin

cujet said:

Very well said, and I agree with all of your points.

But you must remember that the liquid cooled engines have a much lower Delta "T". This means more air is required to achieve an equiv level of heat transfer.

Maybe, in fact, those engineers know what they are doing.



Franklin

Click to expand...

This has been discussed at length here on VAF and there is no flying evidence to prove this contention. This is a very complex area with multiple variables. Definitely not as simple as Delta T.

Here was some info from the water vs. air cooled thread from a couple weeks ago which actually proves the opposite may be true:

http://www.liquidcooledairpower.com/index-flash.html

http://www.liquidcooledairpower.com/...ss_arrow.shtml

Hmmm. Maybe something to this liquid cooled stuff after all.

This rad setup is similar to my thoughts for RVs.
************************************************** ************************************************** *****

For those new to the discussion, here's what I posted about a year ago on VAF:

I know the air cooled guys love to say that water cooling is draggier but never produce any facts to support. Here are some tidbits I dug up:

Bristol Beaufighter. 1280hp Merlin 330 mph, 1590 hp Bristol Hercules 323 mph.

Tempest I. 2240hp Napier Sabre 466 mph. Tempest II 2520hp Bristol Centaurus 440 mph.

Reggiane RE 2001. 1175hp Alfa Romeo 337 mph. RE 2002. 1175hp Piaggio 329 mph.

DC-4/ Northstar. 353mph with 1760hp Merlins, 280 mph with 1450hp Pratt R2000s.

Note these are identical or close to identical airframes save changes for the different engine installations. In the case of the Merlin engined Beau and Northstar, these used the Rolls Royce "power egg" system of engine/ radiator in one, firewall forward- hardly the most efficient from a drag standpoint but making engine changes much quicker.

The Tempest in particular shows how much more efficient liquid cooling can be with 280 hp less and 26 mph faster at about the same altitude. The Tempest I had nicely done leading edge mounted rads.

Ross,

again....the air cooled had the frontal area of a radial...then a different cowl for liquid cooled....this in your examples.

On the other hand there is substantial evidence that for a given power or fuel flow, the liquid cooled rvs are slower....sometimes much slower...so there is the evidence.

There is a fundamental need for greater volumes of cooling air....can it be overcome? Probably, but there is not much wrong with the rv cowl now...so it won't be as easy as loosing a radial. No one has managed it yet. Your belly scoop approach may be the first.

Ross, you want your cake and want to eat it too!

You firstly state that the air/liquid cooling debate is much more complex than just the delta-T, but then quote a bunch of airframes to justify your position where there are many more factors at work than just a simple change of cooling system.

Likewise on the CoolJugs Arrow, I watched that video and when ho hum. It just doesn't tell you anything, since much is unknown about the state of either aircraft. It's pretty easy to make any aircraft go slowly - even unintentionally. Just as one possibility, poorly maintained baffles and seals on an air-cooled engine can dramatically increase cooling drag, reducing max speed in the process.

A

Just posting the info guys, take it for what it is worth.

We'll be waiting for the fully instrumented tests and side by sides sometimes in the future.

If anyone has any real world evidence where a properly designed liquid cooled setup (not cheek mounted rads) in an identical airframe with similar propellers has been inferior, I'd be interested in seeing it. I have not seen a nice installation yet on a liquid cooled RV where proper control and use of the cooling air has been fully implemented for low drag.

Under cowling space for a liquid cooled and intercooled engine is limited in an RV and intercooled engines will always produce higher drag than those without. It is a tall order to do a low drag setup on an RV, especially an A model with the gear structure in the way.

A friend just started flying his turbo EJ22 Glastar a couple weeks ago. This uses twin GM evap cores mounted at an oblique angle to the airflow and fed by a plenum arrangement with guide vanes. Despite massive area and volume, cooling is only adequate at OATs up to about 15C. He will be revamping the setup with a custom built rad and exit flap. I'll report on the success or failure of these changes.

There are at least 3 aircraft flying with belly rads now which are cooling really well with relatively small inlets and radiators. We don't know what total drag is like on these yet but this route seems to offer the best chances of success with lower drag and decent ground cooling.

Please, I didn't intend to reignite the air vs. liquid cooled engine debate nor intend to diss certain engineers. I admit I'm not schooled in the math, but the concepts behind heat transfer are simple. The devil is in the details.

I made a forest vs. trees type of observation. I'd love to have any excuse to by an engine that offers the advantages of a 2-stroke turbo diesel. It's just that I'm not willing to stick my leg out the side of the fuselage to do so.

My challenge to the engineers is to find a way to shed the necessary heat without penalizing the airframe. Since the rate of transfer is reduced, and since mass (or drag anyway) is limited, then I guess the transfer of heat must be given more time to work. This must be old hash somewhere.

I hadn't seen the liquid-cooled-air-power website before today. The pictures of their radiator are quite revealing. Its doubled over which probably serves to increase the amount of time that heat is transferred. I would guess that air exiting this radiator is hotter than a single plane radiator all else being equal.

Since the location of the propeller flange and the slice-o-wonderbread shape of the RV-9 firewall are pretty much a given, then the cowling is going to be cavernous compared to the size of the Wilksch engine. This may be sufficient space to design effective diffusers and nozzles (rather than mere passages) that work with appropriately shaped/sized radiators.

I like the idea of the Wilksch engine. I don't like the idea of handicapping the airframe. I sure wish I had the funds to p-o-c it myself.



Other potential heat sinks:

Hammer a cowling out of 1/2" aluminum then carefully rout miles of groves/fins on it. Bond standard flat radiator tubing on the inside. Your spinning propeller will ensure that you'll never again fear having enough cooling while being 24th in line for takeoff at Oshkosh.

Bond a length or two of radiator tubing to the skin between each wing rib then hook em up. P.S. if you do the right combination of series and parallel connections and fly full-throtle long enough, you'll be able to generate enough electricity for a wig-wag flasher.

Bond surplus Carrier/York a/c tubing to the bottom of the fuselage. Don't forget to make a U-turn at the tail.

Bond radiator tubing to the underside of the wing. A handful of parallel loops oughta to the trick.

Bond radiator tubing to the top of the wing at the appropriate depth of chord. Then bond PC heat sinks to the tubing at alternating angles and call them vortex generators.

And if, for crying out loud, you just gotta have that massive radiator up front, at least put it on a tray that can pivot up and down. Use a cowl flap like cable and knob so the pilot can control the amount of droop. Drop it down for high power/low speed regimes and tuck it up for less demanding regimes and show times. In your spare time develop an automatic temperture sensing system to control the droop with Lego servo motors.

What drives the cooling air is the differential pressure between inlet and outlet. An underbelly scoop placed at the max pressure sone under the wing, and exit behind the wing somewhere at a much lower pressure zone, just has to be way more effective than the normal approach. I also think the original Wilksch position of the coolers, more or less at the stagnation point is a good design, even though it looks strange. With the exit at the relatively lower pressure at the sides, instead of below, this should not give very much drag either (and it will look even more like a shark ). Seriously, real sharks and other fish, would not have their intake at the front and the exit at the sides if this wouldn't be effective.

F1 cars (800+ hp) have forward facing intakes after the front wheels, and exit partially down in the low pressure zone under the card just in front of the rear wheels. This takes away lots of downforce, but it enables smaller rads, so they can go faster on the straights where they don't need much downforce.

I'm not a mechanical engineer, but....

Speed is not a function of liquid verses air cooling, its a function of HP verses drag.

You only have to look as far as cars to get the answer on which works best. They're all liquid cooled, they all have low drag numbers. Before you start yelling about the realitely low average HP used on public roadways, look at F1 racers (per SvingenB). Weight-to-HP and drag are all important and they run HP settings that dwarf ours. I don't see any of those guys using air cooled engines.

Liquid cooling has some obvious advantages in terms of having a more even temperature throughout the engine, eliminating shock cooling, etc. Another advantage alluded to by ergie63 & SvingenB is that you can move the cooling to wherever makes the most sense.

Many liquid cooled RVs are sporting automotive engine conversions that are probably not putting out the HP expected - so - they are slower. Probably drag issues as well.

The RVs were designed and optimized around air cooling. Trying to plug & play a liquid cooled engine in that airframe is not likely to have optimal results.

What if the an RV cowling was aluminum - same thickness as the rest of the plane - and that tiny radiator tubing was spread all over the inside of that cowling?

Then just close the cooling holes entirely.

Anyway, there are probably 20 reasons why that is a dumb idea, but the point is to engineer a liquid cooled solution from the ground up. I don't possess the skills to do that, but I suspect some on this site do.

Wilksch cooling pack

Chaps

The standard wilksch pack is not the ideal solution for low drag cooling, but it does allow Wilksch to produce an engine with a finished and tested cooling solution that is applicable to a wide variety of airframes. It's not the only solution, and you're not restriced to using that solution.

As things go I'd have to describe the Wilksch cooling pack as "pragmatic" rather than ideal or elegant. Similarly I have to say that my choice to build an installation around the wilksch cooling pack is also pragmatic. Designing and building a new engine installation is hard work, and I didn't want to make it any harder than I had to.

Dave

Steve Brown said:

You only have to look as far as cars to get the answer on which works best. They're all liquid cooled, they all have low drag numbers. Before you start yelling about the realitely low average HP used on public roadways, look at F1 racers (per SvingenB). Weight-to-HP and drag are all important and they run HP settings that dwarf ours. I don't see any of those guys using air cooled engines.

Click to expand...

Cars and aircraft makes for a falacious comparison. They neither must produce lift, nor thrust and they are surprisingly unaerodynamic compared to aircraft. Cooling drag is a pretty insignificant part of the total drag on a car.

200hp will make a car do approximately 140mph with no lift (or downforce) to speak of. 200hp in an aircraft will see it doing 220mph (nearly 60% faster) and generating 1g of lift!

Cooling drag is much more important on an aircraft, where the drag coefficient is around a tenth of that of a typical road car, which in turn is much more aerodynamic than an F1 car.

Steve Brown said:

Liquid cooling has some obvious advantages in terms of having a more even temperature throughout the engine, eliminating shock cooling, etc.

Click to expand...

liquid cooling is invariably thermostatically regulated, whereas air cooling invariably isn't, but that's not to say it cannot nor should not be done.

Steve Brown said:

Another advantage alluded to by ergie63 & SvingenB is that you can move the cooling to wherever makes the most sense.

Click to expand...

...at a cost - usually paid for in pounds, and not pounds sterling!

Steve Brown said:

Many liquid cooled RVs are sporting automotive engine conversions that are probably not putting out the HP expected - so - they are slower. Probably drag issues as well.

The RVs were designed and optimized around air cooling. Trying to plug & play a liquid cooled engine in that airframe is not likely to have optimal results.

Click to expand...

Most piston engine cooling systems are hardly optimised be they air or liquid cooled. A comparison of fully optimised air and liquid cooled aircraft would show that the air-cooled aircraft has less cooling drag - however neither aircraft actually exists as yet...

Scarce Resource...

Dave_Boxall said:

pragmatic

Click to expand...

Coming from someone in the midst of it all makes me laugh at myself.

Really?

Andy_RR said:

..... A comparison of fully optimised air and liquid cooled aircraft would show that the air-cooled aircraft has less cooling drag - however neither aircraft actually exists as yet...

Click to expand...

Andy,

Some airplanes only get 140mph or less out of 200hp.

Anyway, your conclusion is surprising to my simple mind.

Assuming you are correct, what is it then that drives every auto manufacturer and every racing team to a liquid cooled conclusion? Aerodynamics & weight are both important. As important as an airplane? No, but engineers don't give away performance for nothing.

Serious question

Steve,

I don't speak for the race teams or the auto manufacturers, but I would say the reason race teams don't use air cooling (though Porsche did!) is due to the specific power that they require from their engines.

Auto manufacturers choose liquid cooling because it makes NVH (noise) refinement much easier. Also they have legislated drive-by noise targets which arguably killed the air-cooled boxer engine in the 911.

There's no doubt that in certain circumstances, liquid cooling is an advantage, but looking at it purely as rejecting heat to the environment, liquid cooling will dump heat to a higher total volume flowrate of air than direct air cooling will.

For a more relevant motor racing comparison, look at intercoolers. Air-air intercoolers are generally chosen because they are more efficient and lighter...

Specific hp, emissions, noise, longevity, proper cabin heat all all reasons why the air cooled engine is dead in modern cars. Many of these criteria are either not so important or non-existent in aviation.

Saying that more mass flow is required to dissipate the same heat energy on a liquid cooled engine is a simple and perhaps inaccurate way of looking at the issue. This may be true but as the total mechanism of heat transfer from the critical high flux areas of an engine goes to actually transferring that energy to the atmosphere with the lowest drag, it is far more complex.

As I've said before, because of the high mass and far superior conduction of water in close proximity to the cooled parts, combined with the alumimum radiator being many times more efficient than a steel barrel or finned aluminum head at radiating heat per unit area and volume, finally combined with a diverging/ converging duct which makes maximum use of the available mass flow of the cooling air and being able to efficiently re-accelerate the air to near free stream velocity with lower pressure loss- simple delta T does not adequately cover the whole process from a drag perspective.

Without hard data, we simply don't know. I haven't seen anything that would suggest that air cooling is superior in this respect. I invite anyone who can supply such data to share it here so we can stop speculating.

Intercoolers are of course exchanging heat between two mediums with nearly equal mass. Routing is simple and we don't have any parts likely to melt from inadequate heat transfer. These factors make air to air intercoolers the clear choice for most land or air bound vehicles which have adequate space. However when we have tight packaging like say under the cowling of a Spitfire or P51 or a boat, liquid to air intercoolers are employed. Just like with engines, both types have been proven to work well.

Without hard data, we simply don't know. I haven't seen anything that would suggest that air cooling is superior in this respect. I invite anyone who can supply such data to share it here so we can stop speculating.

Hard data: every RV with a liquid cooled, horizontally opposed engine and the same cowling as the aircooled, is slower as a function of HP/fuel flow. Since the cowls are the same, and the configurations are the same, it appears that drag is the issue.

Now, the problem is that the need for converging/diverging ducts has not been answered. It would be difficult to do so on an RV, I think. My personal thought is that perhaps the belly scoop approach offers a great opportunity.

In looking at TSweezy's install, I also wonder if it will work pretty well, the side radiator/side exit concept seems like it would allow a diverging duct into a plenum on the inboard surface of the rad, and perhaps the a converging duct on the outboard side, leading to the side exit....will wait to see how it works.

It seems clear that on an RV, the front mounted rad, dumping eventually into the lower cowl exit, it a very draggy configuration....probably designed with other goals in mind (ease of installation...selling plug and play installations).

In the end I do not think, without a specific airframe design approach, weight and drag parity will be accomplished, but the above two configs seem like they will get closer than we have been thus far.

rv6ejguy said:

...a diverging/ converging duct which makes maximum use of the available mass flow of the cooling air and being able to efficiently re-accelerate the air to near free stream velocity...

Click to expand...

I agree. What strikes me about the longitudinal cross section of diverging/converging duct systems is that they are, broadly speaking, the same as a jet engine. In sequence: 1) Build up static pressure, 2) add heat, and 3) extract work.

Since the amount of mass and heat in a cooling system are orders of magnitude less than a jet engine, we're not talking about useful thrust. What seems to be possible to me (a lay person) is that the careful management of pressure and heat can be used to offset the penalty for having to make air move.

Turbulence in cooling passages, whether the engine is directly (air) or indirectly cooled (liquid), works the air which is to say adds heat. We experience that transfer of energy as drag. But all you engineers already know that.

Thoughts on shape (answering to jconard):
The respective cross sections of the WAM (narrow) and RV-9 (wide) together offer an opportunity not possible with other combinations. There is enough room for ducting roughly in the shape of a beetle's pincers on either side of the engine. In my mind, the fellows at the start of this thread were on the right track with their initial effort. It seems to me that they didn't appear to be mindful of managing their pressure gradients evidenced by the location of the engine air inlet in front of the radiator. That configuration steals precious mass and pressure from where it is needed the most. It would have been interesting to see the effect of changing that one thing alone.

Andy_RR said:

Cars and aircraft makes for a falacious comparison. They neither must produce lift, nor thrust and they are surprisingly unaerodynamic compared to aircraft. Cooling drag is a pretty insignificant part of the total drag on a car.

200hp will make a car do approximately 140mph with no lift (or downforce) to speak of. 200hp in an aircraft will see it doing 220mph (nearly 60% faster) and generating 1g of lift!

Cooling drag is much more important on an aircraft, where the drag coefficient is around a tenth of that of a typical road car, which in turn is much more aerodynamic than an F1 car.

Click to expand...

The downforce of an F1 car is enough to enable it to drive upside down at already at 80 mph or so. F1 teams spend billions on making as much downforce as possible with as little drag as possible, cooling drag included. As I wrote earlier, the cooling drag destroys much of the downforce and creates drag. Aerodynamics is much more than drag coefficients. The main thing is that F-1 cars use 2.4 litre normally aspirated V8 engines producing up to 800 hp. If this engine was to be air cooled, it would probably max out at 300-400 hp if it would last a GP race. There simply is no way you can get the needed 250 kW of heat transported out from the core of that little engine by conduction through aluminum alone.

If the same amount of money spent on aerodynamics on F1 cars were spent on airplanes with an F1 engine installation, this airplane would be the size of a -3 and cruise around MACH 0.9 It would probably also have underbelly cooling and look very much like a mini P-51

ergie63 said:

I agree. What strikes me about the longitudinal cross section of diverging/converging duct systems is that they are, broadly speaking, the same as a jet engine. In sequence: 1) Build up static pressure, 2) add heat, and 3) extract work.

Since the amount of mass and heat in a cooling system are orders of magnitude less than a jet engine, we're not talking about useful thrust. What seems to be possible to me (a lay person) is that the careful management of pressure and heat can be used to offset the penalty for having to make air move.

Turbulence in cooling passages, whether the engine is directly (air) or indirectly cooled (liquid), works the air which is to say adds heat. We experience that transfer of energy as drag.

Click to expand...

The Meredith Effect was to have produced a net gain in speed by using the energy from the expanding cooling air exiting the radiator duct at higher velocity. Theory proved somewhat different from practice and later wind tunnel testing showed that it was not possible below about 350-400 knots to achieve unity in cooling drag at the typical 110C coolant temperatures and state of the art radiators in the late 1940s.

It has been fairly clearly shown that a well designed radiator system such as the P51 or leading edge setups like the Westland Whirlwind, Tempest I, or Hornet were quite a bit more efficient that the underwing rads employed on many aircraft and could offset some of the cooling drag associated with these setups.

Obviously most under cowling setups on an RV are not going to be as efficient as what could be done with a lot more work but the drag of a belly scoop must be figured into the equation with its extra frontal area. Clearly cheek mounted rads are the least efficient with poor wetting of the HE surface and poor pressure recovery, not to mention willy nilly air exit paths and poor velocity recovery at the cowl exit- but they are convenient and simple as far as minimal changes to the cowling and plumbing runs go.

Water cooling = higher horspower

Andy_RR said:

Steve,

I don't speak for the race teams or the auto manufacturers, but I would say the reason race teams don't use air cooling (though Porsche did!) is due to the specific power that they require from their engines.

Auto manufacturers choose liquid cooling because it makes NVH (noise) refinement much easier. Also they have legislated drive-by noise targets which arguably killed the air-cooled boxer engine in the 911.

There's no doubt that in certain circumstances, liquid cooling is an advantage, but looking at it purely as rejecting heat to the environment, liquid cooling will dump heat to a higher total volume flowrate of air than direct air cooling will.

For a more relevant motor racing comparison, look at intercoolers. Air-air intercoolers are generally chosen because they are more efficient and lighter...

Click to expand...

Andy,

I'm not a Porsche fan, but my recollection is that Porsche made the jump to water cooling to support performance and reliability unobtainable with air cooling.

In other words, the performance didn't go down due to water cooling, it went up.

Also, with your "specific power" comment I think you are conceding that liquid cooling is more efficient at getting the heat out.

To me that equals better performance, if the airframe is optimized to take advantage of it.

Emissions and noise were the main factors in Porsche going water cooled in production cars.

In racing, where the 962 reined supreme for so many years in IMSA GTP, it was simply an issue with power density and specific output once the Nissan VG30s and AAR Toyotas entered the picture. Even with later water cooled heads, the Porsches were no match for the water cooled narrow engines in the hp department or in the downforce area where engine width was a detriment.

The same thing is seen in Unlimited racing at Reno where in years past, the Dwight Thorne Merlins were making about the same hp (about 4000 ) as the best R3350s with only HALF the displacement. The P51 Dago Red still holds the fastest race lap ever run at Reno by a substantial margin at 512 mph.

The latest F1 engines are unbelievably small packages for the 800hp they produce. An 800hp air cooled engine would probably have to be triple the volume of a liquid cooled one even allowing for the radiator volume to be included.

The Sport Class Lycomings and Continentals are nearing their thermal and mechanical limits now even with spray bars and ADI. These will probably not be able to exceed much over 750-800 hp no matter what with any reliability. The liquid cooled Falconer V12s and Chevrolet V8s being raced and developed in the wings will supplant the air cooled engines in the not too distant future IMO.

Bottom line, liquid cooling works just fine and we just have to find better ways to package them into an RV airframe.

I was just uploading some photos to Flysoob and noticed some new photos of a Subaru installation on a Europa there. Details a dedicated Kuchemann/ Weber derived radiator duct with the rad mounted underneath the oil pan. The inlet is relatively small and the duct diverges to the radiator and then converges aft of it to the exit. I was thinking of a similar arrangement when I redo my 6A cooling system. I should have built a -6 rather than a -6A as the nose gear structure makes a nice duct shape more difficult. I'll have to split the exit flow.

As for the intercooler- I like the setup in the P47 which used a dedicated duct to feed an aft fuselage mounted intercooler with nice exit ramp exits. The aircraft was designed around the turbocharger and intercooler installation much like the P51's cooling system was designed into the airframe rather than onto it. The P51 integrated the liquid to air intercooler heat exchanger into the radiator duct. This was also a pretty clean installation.

A liquid cooled, turbo charged and intercooled RV presents some real challenges in efficient packaging to fit it all into a low drag setup. Ask me how I know.