Quantcast

Successful Auto Conversion Trials, Tribulations and Tips

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
820
Location
Jackson
Most of the info in the paper you linked (especially the graphs & formulas) appear to be lifted directly from Kuchemann & Weber's Aerodynamics of Propulsion, Chapter 12. (Note the re-publishing date of the copy I linked: 1953.)
https://www.amazon.com/Aerodynamics-propulsion-Dietrich-Johanna-Kuchemann/dp/B004H6Y45K

I really thought you were coming here to 'spitball' ideas; trying to determine what will work. Nothing wrong with that; I've done it in the past and will do it in the future, when I'm trying to explore new-to-me ideas. Thing is, not many ideas are truly new (as demonstrated by the paper you linked). I've floated a lot of 'new' ideas, and more experienced people were nice enough to show me where/how they'd already been proven ineffective. In the case of a 'universal cooling duct', you floated the idea, and Ross gave you some very good reasons why it wouldn't work. You persisted, floating ideas for techniques that also won't work, and I tried to explain why they won't work.

The purpose of my post(s) is to hopefully save you some wasted time and effort, trying things that have been proven over and over to *not* work. If you've already decided what you're going to do, that's fine (though unfortunate, since most of what you propose has already been tried, unsuccessfully). In that case, I'm sure that most of us here will be happy to withhold our advice. But the problem with you posting your ideas without proving them first is that someone with even less experience than you will read your uncontested posts, and wander down those same paths, believing you've tested them. Unless of course, someone with more experience answers your posts with more accurate data.

Perhaps a better approach would be to say, "I've got this idea; has anyone seen it tried before, and did it work?" and realize that the people who answer are probably trying to help, even when the answer is some form of 'nay'.

Now, as requested, I'll stop trying to help. But you should expect people (including me) to publicly correct info that's obviously incorrect.

Charlie
 
Last edited:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,758
Location
Calgary, Alberta, Canada
I see, I understood it to be a splitting of the tank to reverse the flow back through the radiator.

Did you try honycomb core or was it always tube and fin?

Also, are you saying that each engine would need a new duct geometry based on radiator size? Heat output? I didn't think that the effect was that specific for each application. So a duct system could not be designed that, say, covers a radiator from 150-200sq in? To see any benefit you need a 150sq/in duct and a 200sq/in duct?

Thanks again for your answers!
Tube and fin is more efficient than honeycomb designs which went out in the 1950s.

Geometry could remain similar for the same HX height to width ratio however inlet size would have to be altered for different speed ranges and increase as the HX area increases. One size can't fit all if you're concerned about drag.
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,758
Location
Calgary, Alberta, Canada
Well I found some info.. generally the diffuser is the money maker:

http://www.arpnjournals.org/jeas/research_papers/rp_2016/jeas_0416_3979.pdf

Therefore, the traditional compromise-dimensioning advices a value of the semi-angle opening of the duct of about 7°, that is the value usually used in the wind-tunnel diffuser.
However, with modern polished RP (Reinforced Plastic) construction and anti-friction paints it is possible to reduce this angle down to about 2°

A fully deployed streamline diffuser is long 3 times the height of the radiator divided by two (YB).

The other factor is that the duct is rarely rectilinear and cylindrical, but often has S-shapes and ovoidal sections, that further reduce efficiency in “critical” ducts. - This means make it straight into the radiator.

"The Meredith duct should be embedded in the fuselage or in the wing to avoid excessive external drag. Only the air intake is positioned outside. The optimized intake is positioned in the lower part of the aircraft at about 2/3 of the wing chord, where the pressure reaches its maximum."
I can't understand this report saying that divergence angle could be reduced from 7 to 2 degrees. You'd want to INCREASE that angle if you could, to shorten the duct for less internal drag. The small inlet area has to transition to the much larger HX face. A 2 degree divergence angle would massively increase duct length.

I kept my duct as rectangular as possible since the rad is that shape. My diffuser is about 5 times as long as the actual height of the HX and it has a guide vane at mid height as well.

Russell and me both found in actual flow testing than some of the often quoted theory on duct design didn't match our results. Our ducts are longer than theory suggests would work fine and both had guide vanes where usually they're not employed.

Anyway, we both have flying examples here that cool well with small inlets. Take that any way you wish. Practice trumps theory every time in my book...
 
Last edited:

pictsidhe

Well-Known Member
Joined
Jul 15, 2014
Messages
7,977
Location
North Carolina
2 degrees is going to be awful long. The optimum angle is actually dependant on a whole host of factors. Whetehr you want maximum efficiency or maximum pressure recovery, shape, area ratio, Re, intake turbulence. It's probably going to be easier to do some reading on what has worked well, then make several test systems bracketing what you think is the optimum geometry. Test systems only need to last the test, so cardboard, papier-mache etc could be employed. You are not going to nail it first time. With some book research, you could probably build something reasonable first time if you aren't going after every last drag count. If in doubt, going a little shallower than theoretical optimum is at worst a small penalty. Cooling load can be approximated as the hp output of the engine, from that you can pick a delta and know how much air flow that will require.

Good books, in the order you should buy them:

Compact heat exchangers: Kays and London This one is compulsory!

Diffuser design technology: Japikse and Baines. Detailed information on diffusers, the easiest part to gain unwanted losses.

Extended surface heat transfer, Kraus, Aziz, Welty Will give deeper insight into effiicient radiator or fin design. If you are just going to pick a car radiator, you can probably skip this one.

Aerodynamics of propulsion, Kuchemann and Weber mentioned above is probably worth a read too.
 
Last edited:

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
820
Location
Jackson
Ross, pictsidhe,

Just doing a quick scan of that doc, it sounds like the authors are claiming that reducing skin friction in the duct with 'paint', etc, allows using that incredibly long duct (2 degree expansion) . Of course, they ignore the complete impracticality of putting one on an actual aircraft. The statements following, where they refer to 'figure 2' (apparently meaning figure 8) and the 'green line' are trying to tie the figure 8 graph to a straight-line duct. That graph is lifted directly from K&W chapter 12 page 272 (fig 12-11), and refers not to a straight duct, but an 'oblique' duct (a wedge duct, in todays' common parlance). Figures 10 & 11 are also direct lifts from K&W pg 276, though they did add nice color to the drawings and add the claim that the duct curve is from a P-51 duct (that much might even be accurate; I've looked up a P-51's 'skirt' when the duct was off, and it really is shaped like that).

The paper seems a bit 'multiple-personality', because the 7 degree (reduced to 2 degree) reference is to straight sided ducts, but they suddenly shift to what K&W call a 'streamline' (trumpet shaped) diffuser, while throwing in that heat exchanger turning vane drawing which K&W gave us to show how to deal with losses in a wedge duct. (?)

Charlie
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,758
Location
Calgary, Alberta, Canada
Another useless paper based on theory and unproven conclusions. Lots of those out there which is why I built my own test rig to evaluate different HXs for drag and effectiveness did my own experiments with actual ducts and HXs before building my final one.

Even some of the accepted texts don't have much useful information to apply to aircraft radiators IMO where low drag is a major consideration- or should be. None I read had any useful guidelines for momentum recovery. Most were concerned with the diffuser which is only half the story.
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,758
Location
Calgary, Alberta, Canada
Last edited:

slociviccoupe

Well-Known Member
Joined
Apr 1, 2011
Messages
266
Location
Palm Bay Fl.
Could a universal car hood scoop be used as a starting point? There are a few that are actually wind tunnel tested, mainly the ones used on pro stock drag cars.
 

pfarber

Well-Known Member
HBA Supporter
Joined
Feb 21, 2019
Messages
551
Location
Pennsylvania
"However, for tropical conditions, it is completely inadequate, since the flaps open fully already at +42°, which causes a speed loss of 48 km / h. At an air temperature of +50° + 5° the coolant reaches 115°. "

Yikes that's 120F OAT. Were there any water cooled AC that could live at those temps?

Also, is the report stating that engine is underperforming at that temp, or that the flap drag is causing the reduction in speed?

"It should be noted that the speed decreasing as the temperature increases is not only due to the larger radiator opening. It is superimposed by the effect of the change in specific weight of the air and the engine performance. "

To me it reads more like "The 109 is not a good tropical AC because the radiator is to small' not that the design is bad.
 

slociviccoupe

Well-Known Member
Joined
Apr 1, 2011
Messages
266
Location
Palm Bay Fl.
Say if one were to put belly scoop, then use area behind rear seat for radiator and ducting, then run ducting down to tail. Would this work? Also thinking ducting could converge down to a round shape and flap inside would be much like a large throttle body. Tube exiting the bottom would have a nice long slash cut to it and might act like a ventury to pull air out.

Ive read all the pages from rv6ejguy's belly scoop and russel's belly scoop on the lancair. Both sucessfull installations.
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
820
Location
Jackson
Some thoughts opinions.
You *could* start with one, but I suspect that you'd get better results for less money by rolling your own. The shape inside is even more important than the outside, and none of those automotive scoops are intended to feed a heat exchanger. Also, the one you linked would have awful aero at the back end. They probably get away with it on a car hood because it sits in the highest pressure point on the car. If it were open back there in an automotive install, it would get almost as much flow into the scoop as the forward facing end.

One thing to keep in mind is that most of the documented high-performing diffusers are 'one dimensional'. The inlet is the same width (or height) as the core, and the expansion is in the other dimension. You can make it work with expansion in both dimensions, but it is a *lot* harder to design and build one that avoids internal flow separation and poor efficiency. Ask me how I know....

On fuselage buried heat exchangers: It's not terribly complicated to do on a tube/fabric airframe, but be very careful with aluminum. Punching large holes in a stressed skin fuselage can have big negative structural consequences. The P51's structure was designed to accommodate the heat exchanger duct. Might be a little easier with composites, but I'd still be very leery.

Air exit. The CAFE foundation did some testing on exits a number of years ago, and they found that a 'bluff body' was the most aerodynamically efficient way to get air out of a cooling system. Visualize a teardrop, slice it in half vertically, then cut off most of the tail. Basically a quarter of a sphere. The interior, feeding into the 'bump', needs to have smooth, large radius lips. You can read the CAFE studies on their web site. Lots of great info there.
CAFE Foundation

Hope that's of some use...

Charlie
 

slociviccoupe

Well-Known Member
Joined
Apr 1, 2011
Messages
266
Location
Palm Bay Fl.
Charlie appreciate the insight. Im planning on pro composites vision with a subaru ej25.
Taking into consideration weight and cooling.
Building my firewall forward first then the plane. Get weight sorted out and a reliable engine. I just need to do more reading on the scoop.
 

pfarber

Well-Known Member
HBA Supporter
Joined
Feb 21, 2019
Messages
551
Location
Pennsylvania
Some thoughts opinions.
You *could* start with one, but I suspect that you'd get better results for less money by rolling your own. The shape inside is even more important than the outside, and none of those automotive scoops are intended to feed a heat exchanger. Also, the one you linked would have awful aero at the back end. They probably get away with it on a car hood because it sits in the highest pressure point on the car. If it were open back there in an automotive install, it would get almost as much flow into the scoop as the forward facing end.

One thing to keep in mind is that most of the documented high-performing diffusers are 'one dimensional'. The inlet is the same width (or height) as the core, and the expansion is in the other dimension. You can make it work with expansion in both dimensions, but it is a *lot* harder to design and build one that avoids internal flow separation and poor efficiency. Ask me how I know....

On fuselage buried heat exchangers: It's not terribly complicated to do on a tube/fabric airframe, but be very careful with aluminum. Punching large holes in a stressed skin fuselage can have big negative structural consequences. The P51's structure was designed to accommodate the heat exchanger duct. Might be a little easier with composites, but I'd still be very leery.

Air exit. The CAFE foundation did some testing on exits a number of years ago, and they found that a 'bluff body' was the most aerodynamically efficient way to get air out of a cooling system. Visualize a teardrop, slice it in half vertically, then cut off most of the tail. Basically a quarter of a sphere. The interior, feeding into the 'bump', needs to have smooth, large radius lips. You can read the CAFE studies on their web site. Lots of great info there.
CAFE Foundation

Hope that's of some use...

Charlie
Going back over some of calculations I read (and posted previously)

"A fully deployed streamline diffuser is long 3 times the height of the radiator divided by two (YB)."

So to keep all other constraints, a 200hp radiator would be 10x20x1 (H W L) meaning that you only need a diffuser of ((3*10)/2) 15 inches.

Using a 30% inlet opening ratio (60 sq/in) 20in wide x 3 in high.

That's going from 3inches to 10 inches in 15inches of depth. This is only in one dimension (height).

But "The plenum should not diverge significantly... less than 20deg or less is optimum "

To resolve these differences, the only thing to do would be a mock up with manometers to measure pressures?
OR
Would you just use tufts and go for the most laminar flow you could get into the radiator, and tell Mr. Meredith 'not today, good sir'.

My preferred install is a chin mounted scoop, radiator under the motor, exhuast to the rear of the cowl. I'd rather scrap the idea of any meredith effect before I tried a belly scoop.
 

TFF

Well-Known Member
Joined
Apr 28, 2010
Messages
13,405
Location
Memphis, TN
Drag car scoops are not designed for off axis airflow. They are designed for essentially straight directions. Design one like an Indy car or F1 would be closer. Off the shelf ones are not usually advanced. The likelihood of one even being remotely close to fitting would be a miracle.
 

Russell

Active Member
Joined
Nov 29, 2015
Messages
43
Location
Ft Worth, Texas
So to keep all other constraints, a 200hp radiator would be 10x20x1 (H W L) meaning that you only need a diffuser of ((3*10)/2) 15 inches.
Using a 30% inlet opening ratio (60 sq/in) 20in wide x 3 in high.
Pfarber, your calculations for a 200 HP radiator match fairly closely with the one flying in my Glasair / Subaru EG33. My rad is 16” x 15” x 2.75”.

However, I find that in my case, your 30% inlet size is too large. My opening is about 33 sq/in …. In fact when we race this plane or during the winter, I have a sleeve that goes over the front opening and reduces it to about 27 sq/in.
Russell Sherwood
 

Toobuilder

Well-Known Member
HBA Supporter
Log Member
Joined
Jan 19, 2010
Messages
4,858
Location
Mojave, Ca
Is this 33 inches your total thermal inlet area? Do you have a separate oil cooler duct and/or cowl vents?

Trying to compare because I have 46 inches TOTAL on the Rocket. This is primary cooling, oil cooling and "cowl ventilation". I'm effectively cooling 300+ HP in some ungodly high ambient temps.
 
Top