What's new
Van's Air Force

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

Exit diffuser on oil cooler to increase flow through oil cooler

scsmith

Well Known Member
I want to do some cowl work to reduce my cooling exit flow area and at the same time reduce overall frontal area.

But before I can do that, I have to do two other things. One is make a seal around the crankshaft/spinner gap. As you close off the cooling exit, raising the pressure in the lower cowl plenum, you want the higher pressure to result in higher exit velocities out of the cooling exit, not just spill out around the spinner. So I have to do that.

The other thing I need to do is increase the flow through the oil cooler. The angle-valve engines put a lot more of the heat load into the oil, and efficient oil cooling is crucial. I've been very happy with my firewall-mounted cooler arrangement with the large 20006A cooler, I have good oil temperatures even during a long climb at Vy on a hot day. But as I close off the cowl exit and increase the pressure in the lower plenum, the oil cooler flow will decrease because of the reduced pressure differential across the oil cooler. So..how to get that flow back? put a diffuser on the cooler exit. The trick is to do this smoothly enough get good pressure recovery without a lot of length that wouldn't fit in the cowl area. Short, wide-angle diffusers can work pretty well. The pressure in the oil cooler flow at the exit of the diffuser will match the local static pressure in the area around it. Ideally, if the diffuser is flowing well, this results in lower pressure at the exit face of the cooler itself. Its like immersing the cooler in the throat of a venturi.

I made a mold out of wood, painted and waxed it, and laid up a fiberglass diffuser. The diffuser geometry has an area ratio of 1.45. This will also be a good trial of the high-temperature hardener system (3136R) I got for the Jeffco 3107 (Rhino) epoxy we use. The diffuser was room temperature cured, then removed from the mold and post-cured, 1/2 hr at 150F, 1/2 hr at 175F, then 2 hr at 200F. Hopefully this will give a high enough glass-transition temperature (T_g) for the epoxy to survive the hot environment in the lower cowl. Here are a couple of pictures of the installation:

img_1056.jpg

img_1057.jpg


You can see, especially in the second picture, that I made the diffuser a little bit too small. I had measured the mounting flange length and assumed that the cooling channels were the same length, when they actually extend about a 1/4" beyond. So I will add a 1/2" plank on the mold and make another one. In the mean time, I can test this one.

Another issue is that I have an engine mount tube crossing the exit face of the oil cooler. Without the diffuser, this likely reduced the flow through the cooler a little because of the "blockage". With the diffuser, there is some suction pulling on the exit of the cooler that will help pull flow around that tube. Hard to say if the unsteady flow around the tube will help or harm the flow in the diffuser itself. It may cause an oscillating flow separation on the diffuser walls, but hopefully the average pressure recovery will be OK. When/if I put a slightly wider diffuser on, I will also put a spiral wrap of wire on the tube to break up the coherence of the oscillating separated flow downstream of the tube. I'll report back when I get a chance to fly with this.
 
Last edited:
Based on nothing more than wrapping a cord around the roof rack bar on my truck, a single wrap or maybe a 1.5 turn wrap is sufficient.

The truck goes generally less than 75 mph, the bar is about an inch diameter, and the cord was 1/4" marine line. I expect the portion of the bar between supports was four to five feet long. The bar was close to the truck roof, relatively, several inches apart, and the effectiveness of the wrap was evaluated by the presence or absence of thrumming - oscillating aero forces coupling with natural frequencies of the roof, when there was nothing attached to the bar.

Dave
 
....Another issue is that I have an engine mount tube crossing the exit face of the oil cooler. ....

Could get Mr. 3-d Printer (AKA Steve Melton) to make a snap-on airfoil shape to go over the tube (shaped similarly to "streamline tubing" that we make pitot tube mounting extensions from).
 
Put an airfoil-shaped fairing on the engine mount tube to cut its drag. The engine mount tube doesn't care at all but the airstream does.
 
Could get Mr. 3-d Printer (AKA Steve Melton) to make a snap-on airfoil shape to go over the tube (shaped similarly to "streamline tubing" that we make pitot tube mounting extensions from).

Thats a pretty good idea!
 
Put an airfoil-shaped fairing on the engine mount tube to cut its drag. The engine mount tube doesn't care at all but the airstream does.

Thanks. Thats a pretty good idea. It would actually help with the diffusion too.
 
CFD plots I have for a 6A show essentially the same pressure at the spinner base as the the cooling inlets. My OAT sensor is placed just inside the spinner gap area and reads ambient temps any time I'm flying, proving that air is flowing from outside to inside. I use this air to cool my redrive case as it comes in nicely from all directions. You could instrument the gap and your inlets with a couple of piccolo tubes and Magnehelic pressure gauges to confirm.

I'd also instrument the oil cooler exit as it is now with a pitot tube to look at the present exit velocity before you spend a lot of time making a diffuser. You may find it is going very slowly after traveling down the SCAT hose, the wedge duct and through the thick cooler.

My guess after doing a lot of this through flight testing and velocity/ pressure measurement- you won't see any measurable gain in speed with this effort but would be interesting to see the numbers in any case.

I doubt if the tube will hurt much in the slow airstream there, most of the momentum loss is upstream of that and you have lots more area at that point too.
 
Steve,

I'm always trying to learn more stuff about cooling vs drag (water cooled alt engine), so...I'm curious about your info source indicating that a diffuser on a heat exchanger exit will improve flow through the heat exchanger. Normally, you'd see a diffuser on the inlet side, to slow the air and raise the pressure.

I realize that a lip on the exit end of a *duct* will lower pressure at the exit, but my understanding is that this is caused by accelerated flow on the *outside* of the duct (accelerated free-stream flow around the a/c).

Thanks,

Charlie
 
CFD plots I have for a 6A show essentially the same pressure at the spinner base as the the cooling inlets. You could instrument the gap and your inlets with a couple of piccolo tubes and Magnehelic pressure gauges to confirm.

I'd also instrument the oil cooler exit as it is now with a pitot tube to look at the present exit velocity before you spend a lot of time making a diffuser. You may find it is going very slowly after traveling down the SCAT hose, the wedge duct and through the thick cooler.

Ross, it is a gross simplification to state one pressure for the spinner gap. If you go back and look at your CFD results, you should see that the pressure along the bottom of the spinner where it fits against the lower cowl is well above free stream, almost stagnation. In that area, I would expect slight flow into the lower cowl from the outside. (I don't really want that -- it reduces cooling mass flow for a given lower cowl plenum pressure). If you look at the area along the top of the spinner, where the cowl lines are faired nicely with the spinner, you should see that the pressure there is lower than free stream. Air is spilling out of the spinner gap there. If you raise the lower cowl pressure, it will spill worse. ( I don't want that). The pressure at the spinner gap adjacent to the cooling inlets is part way in between the upper and lower -- And for most of us, the inlets are spilling out there too, so the flow and pressures right there are complicated. Flow likely going into lower cowl from outside. I don't want that.

The aluminum wedge-shaped plenum that mounts the cooler to the firewall is almost at upper plenum pressure, close to stagnation. There is some loss through the SCAT tube. I will likely increase the scat tube diameter too. Whatever speed the flow is through the cooler now, its going to be faster with the diffuser.

A lot of time? Gosh, it took about 4 hours total. I figured before I spend a lot of time acquiring and installing any measurement system, and even more time getting measurments, I would just go ahead and make one and see what the oil temps look like.
 
Last edited:
Steve,

I'm always trying to learn more stuff about cooling vs drag (water cooled alt engine), so...I'm curious about your info source indicating that a diffuser on a heat exchanger exit will improve flow through the heat exchanger. Normally, you'd see a diffuser on the inlet side, to slow the air and raise the pressure.

I realize that a lip on the exit end of a *duct* will lower pressure at the exit, but my understanding is that this is caused by accelerated flow on the *outside* of the duct (accelerated free-stream flow around the a/c).

Thanks,

Charlie

Charlie, I guess the original info on an exit diffuser increasing the velocity through the heat exchanger is Bernoulli's Equation.

And yes, we all have a very nice diffuser on the inlet side, it is the cooling inlet ducts, throttled so that much of the diffusion is external, and the rest of the diffusion happens as the flow comes in through the cooling inlets and essentially stops in the upper cowl plenum, reaching essentially stagnation pressure. In my case then the flow re-accelerates some through the SCAT tubing (and no doubt I would benefit from a larger diameter SCAT tube) and then slows back down to almost stagnation pressure in the plenum box that the cooler is mounted on. Yes there is some pressure loss through the SCAT, and also some pressure loss where the SCAT dumps into the plenum box. Those losses are modest compared to the loss through the cooler itself.

Whatever the velocity is now thru the cooler, it will be faster with the diffuser.
 
Last edited:
In flight measurement won't take 4 hours and you'll know the whole truth without any guesses.

Like I said, on my airplane, air is certainly flowing from out to in on the spinner gap, otherwise my OAT sensor would be reading cowling air temps instead of ambient. Your plane could be different which is why you should instrument and find out as Dan Horton and I both did.

Post your results whatever you find out. Always enlightening.
 
In flight measurement won't take 4 hours and you'll know the whole truth without any guesses.

Like I said, on my airplane, air is certainly flowing from out to in on the spinner gap, otherwise my OAT sensor would be reading cowling air temps instead of ambient. Your plane could be different which is why you should instrument and find out as Dan Horton and I both did.

Post your results whatever you find out. Always enlightening.

I guess since you have mentioned it, I can't resist satisfying my academic interest too. My original plan for the diffuser and the cowl nose seal was just to do what I know would make it "better" without really knowing how much better. But it would be fun to know.

So -- where do I get a measurement system that can read something like 5 pressures and maybe a temperature or two? For how much money?
I'd have to mount it in the cockpit somewhere, and get power to it, and a ground. I'd have to route some pressure tubes through the airframe somewhere, penetrate the firewall (Or I suppose I could route them outside throught the canopy skirt and tape them onto the outside of the airframe with some duct tape).
This all seems like a substantial project on its own -- I can't image getting it installed in under 4 hrs, but maybe.

The dividend would be that once installed, it is there to make other measurements too, so for instance, when I start closing off the cowl exit, I can measure lower cowl pressure and exit flow velocity. Again, here I was just going to do things I know will make it "better" without being able to say really how much, other than airspeed change, if any.

So I'm open to recommendations on an instrumentation system. I don't really want to build a multi-tube manometer. Lord knows I've sprayed enough Meriam Fluid on the lab ceiling already in my life. ;) And besides, that would take more than 4 hours.

I don't doubt that the "net" flow around the spinner is probably from outside in. Just saying it is a fairly complicated flow and not going inward everywhere. Spillage out the top is bad. Flow bypassing the engine cooling baffles to go into the lower cowl from outside is bad. A nose seal will be good, regardless of which way the unsealed flow is going.
 
So -- where do I get a measurement system that can read something like 5 pressures and maybe a temperature or two? For how much money?

Ahh, the Chinese menu dilemma: fast, cheap, good...pick any two.

FWIW, I have a row of miniature bulkhead fittings installed in my firewall. Buy a length of 10-32 threaded rod. Cut it into 1-1/2" lengths. Chuck it in the lathe and machine a half inch of threads off each end, leaving a half inch of threads in the middle. Center bore with something small. Now drill a 3/16" hole in the firewall and fasten with a plain nut on each side. 1/8" tubing pushes over the threadless ends.

I have something similar installed in the rear baffle wall, four lengths of brass model airplane tubing soldered in a brass plate.

In an attempt to read static pressure only, I usually go with piccolo tubes, but you know more about that stuff than I'll ever learn.

This cheap electronic manometer is all over eBay:

https://www.ebay.com/itm/US-Ship-LC...4:g:OdAAAOSwTxhcQZuB:rk:1:pf:1&frcectupt=true

Easy temperature probes are LM34-ND top hat cans soldered on the end of a shielded 3-wire cable. The shield is purely physical armor, no electrical value, since the result is a tiny temperature probe on a wire you can move anywhere in the engine compartment.

The nice thing about an LM34 is the 10mV per degree F output. For example, 1.45 V = 145 F.

http://www.ti.com/lit/ds/symlink/lm34.pdf

All you need is power, ground, and your digital voltmeter. Or you can grab a handful of these:

https://www.ebay.com/itm/5PCS-Mini-...=item41896d9197:g:DFIAAOSwdsFUMg1i:rk:20:pf:0

When I built the airplane, I ran six blank wires through the firewall with the rest of the instrument leads. Those six go to a terminal block on the firewall so I always have an easy experimental hookup back to the cockpit.

The dividend would be that once installed, it is there to make other measurements too, so for instance, when I start closing off the cowl exit, I can measure lower cowl pressure and exit flow velocity.

Come on, do it, you know you want to......;)

A nose seal will be good, regardless of which way the unsealed flow is going.

Soft foam block seal if you have the propshaft space. The Rocket gang knows about them. The flap seal I use is still fine after 850 hours. I've thought about detailed pressure measurements in the gap behind the spinner, but it's pretty far down my list for the very reason you state.

Mini bulkhead fittings for manometer tubing, tygon small engine fuel line from NAPA:

Bulkhead%20Fittings.jpg


LM34 on a wire, here determining the air temperature near the variable exit servo:

2012-11-10_16-44-54_99.jpg


Trusty Fluke in the cockpit. LM34 10 mV per degree means 140.4 F. Orange tubing is for manometer:

DVOM.jpg
 
Last edited:
Ahh, the Chinese menu dilemma: fast, cheap, good...pick any two.

FWIW, I have a row of miniature bulkhead fittings installed in my firewall. Buy a length of 10-32 threaded rod. Cut it into 1-1/2" lengths. Chuck it in the lathe and machine a half inch of threads off each end, leaving a half inch of threads in the middle. Center bore with something small. Now drill a 3/16" hole in the firewall and fasten with a plain nut on each side. 1/8" tubing pushes over the threadless ends.

I have something similar installed in the rear baffle wall, four lengths of brass model airplane tubing soldered in a brass plate.

In an attempt to read static pressure only, I usually go with piccolo tubes, but you know more about that stuff than I'll ever learn.

This cheap electronic manometer is all over eBay:

https://www.ebay.com/itm/US-Ship-LC...4:g:OdAAAOSwTxhcQZuB:rk:1:pf:1&frcectupt=true

Easy temperature probes are LM34-ND top hat cans soldered on the end of a shielded 3-wire cable. The shield is purely physical armor, no electrical value, since the result is a tiny temperature probe on a wire you can move anywhere in the engine compartment.

The nice thing about an LM34 is the 10mV per degree F output. For example, 1.45 V = 145 F.

http://www.ti.com/lit/ds/symlink/lm34.pdf

All you need is power, ground, and your digital voltmeter. Or you can grab a handful of these:

https://www.ebay.com/itm/5PCS-Mini-...=item41896d9197:g:DFIAAOSwdsFUMg1i:rk:20:pf:0

When I built the airplane, I ran six blank wires through the firewall with the rest of the instrument leads. Those six go to a terminal block on the firewall so I always have an easy experimental hookup back to the cockpit.



Come on, do it, you know you want to......;)



Soft foam block seal if you have the propshaft space. The Rocket gang knows about them. The flap seal I use is still fine after 850 hours. I've thought about detailed pressure measurements in the gap behind the spinner, but it's pretty far down my list for the very reason you state.

Thanks Dan,

All pretty good ideas. Certainly would have been tons easier to do all this while building. On a finished airplane, ARRGH!

So either one buys several of the manometers, or you sit and switch tubing around in flight (not simultaneous readings, but OK) or make sort of a manifold with manual pressure switches on each input to select which input goes to the manometer, and always be sure only one switch open at a time. Need a co-pilot to log data.

Wouldn't it be great if there was a small box with 5 pressure transducers and 2 thermocouple junctions, plus power supply and signal conditioning, and a USB output to a thumb drive. All you have to supply is the power, ground, and hook up the pressure tubes and thermocouples. I imagine that such a box is available, probably in the $1500 range?
 
So either one buys several of the manometers, or you sit and switch tubing around in flight (not simultaneous readings, but OK)....

I switch tubing around. Hey, you asked for cheap ;)

Wouldn't it be great if there was a small box with 5 pressure transducers and 2 thermocouple junctions, plus power supply and signal conditioning, and a USB output to a thumb drive. All you have to supply is the power, ground, and hook up the pressure tubes and thermocouples. I imagine that such a box is available, probably in the $1500 range?

Just finished installing a single differential pressure sensor, but not for ease of use or less installation time (frankly both are worse), but because I want to look at intake tract and injector bleed pressures at 2400 RPM. In other words, I need a very high sample rate. So, I hope to do it with a laptop, a Dataq DI-1100, and somebody to manage it in the back seat.

DeltaP%20sensor.jpg


Not as all-encompassing as the desired box, but it looks like some of the Dataq kits have channels programmable for thermocouples.

https://www.dataq.com/data-acquisition/starter-kits/?source=homepage
 


These are the Magnehelic gauges, can often find them used on Ebay but new they are cheap anyway. Available in different models and scales. 0-15 and 0-20 in H2O are probably the most useful scales for RVs. These are way more stable and practical than liquid manometers, especially in rough air. Record with your phone or on a kneepad. Old school but worked for me.

A couple of indoor/ outdoor thermometers, but need to find some which go to 100C which is more difficult. An ASI will give you a quick check on exit velocity (corrected for temp), should be lots of these around used with most folks using Glass today.

As an engineer, you'll find it interested to collect and crunch the data I think. :)

Dan's solution is more elegant, always nice to data log on a PC. You don't miss anything and you can concentrate more on accurate flying.
 
In the rocket there just is not a lot of room to make an exit diffuser going to the bottom outlet. Ideally you would have a inlet and outlet that are dedicated to the oil cooler. These could be mounted in the lower cowling area. I saw this idea in practice on one of the MX aerobatic planes. The inlet was on the lower forward surface of each side of the cowling, with a small inlet and a long tapering inlet to the coolers mount on each aft lower cowling. The exit diffusers cleanly exited the cowling lower aft side. This completely isolated the oil cooler from the engine cooling air system.

I would like to do this some time but I decided to do a sort of hybrid system.
My oil cooler is mounted do the back baffle drawing air from the upper plenum, as per normal installations. I built a exit diffuser that curved 90 degrees and exited the cowling just ahead of the side hinge line, in the lower cowling. I then experimented with various "bluff bodies" until I got the lowest profile that would draw air from the cowling. As the shape of the bluff body could draw air from the cowling I did not have to be too fancy with my outlet diffuser, again I was challenged with the available room.
The last picture shows one of the prototype exits. It had the right size opening but "appeared" a bit draggy. The final version is much longer and more streamlined. Over all this solved my oil cooling issue and allowed me to continue experimenting with my cowl flap and exit air fences on the lower cowling. At the end of the day I never felt that I gained much, or any direct speed increases with these modifications. They did however, let me run at full power in a race without high oil temperatures, thus allowing for higher speeds.
120pjep.jpg


2mct4lh.jpg


e9jbma.jpg
 
Interesting and timely thread since I'm working on my cowl right now. Looks like I'm not as clever as I thought I was considering the same ideas I'm implementing are being discussed here.

Mr Martin: I have a similar oil cooler problem and I was looking to penetrate the cowl side just like you have done. I intend to duct the outlet back to the wing root near the point of max thickness to harvest that fat slice of low pressure that always shows up so well in CFD plots. A very simple flanged 3 sided duct could be bent up and attached with speed tape to the fuselage to validate the theory in no time. If you want to do that and report back, be my guest! ;)

Also looking at the prop seal thing. Though Dans flywheel seal works, the Rockets have an extra long hub which provides plenty of real estate to seal. Not only is the overall sealing surface smaller on the Rocket, but the surface speeds are much lower on the small diameter hub. Both are good things. I plan on teflon tape around the hub and a dry felt seal media contained in a floating seal retainer.

Finally, I have added a underbody fairing which closes out the cowl "chin" on the Rocket. This fairing system includes a functional cowl flap as well as exhaust augmentor tunnels. I'm looking into a fully developed augmentor system to eventually replace the articulated cowl flap, but I need to get flying first.

Good stuff.
 
OK, I'm going to demonstrate my untrained ignorance. If Bernoulli says that speed and area have an inverse relationship, and speed and pressure have an inverse relationship (redneck engineering expressions), then I'm struggling to see how putting an expanding cone on the exit of a heat exchanger (when it's located in a volume of basically static air) will reduce pressure at the heat exchanger exit.

When I struggle through books like Kuchemann & Weber, and Kays & London, 'diffuser' is the term used to describe an expanding duct that *slows the air and increases pressure*.

I totally get what Tom Martin did (matches descriptions in the above books, plus docs found on the CAFE website on cooling), but I'm struggling with an expanding cone that's contained within a cowl.

Please educate me.

Charlie
 
I intend to duct the outlet back to the wing root near the point of max thickness to harvest that fat slice of low pressure that always shows up so well in CFD plots. A very simple flanged 3 sided duct could be bent up and attached with speed tape to the fuselage to validate the theory in no time.

I have a box full of wedge, sheet metal outlet ducts from various experiments over the years... Tape 'em on and fly, observe pressures.

Yup, this is a good thread. Interested to see what Steve finds out here.
 
prop shaft seal

Interesting and timely thread since I'm working on my cowl right now. Looks like I'm not as clever as I thought I was considering the same ideas I'm implementing are being discussed here.

snip

Also looking at the prop seal thing. Though Dans flywheel seal works, the Rockets have an extra long hub which provides plenty of real estate to seal. Not only is the overall sealing surface smaller on the Rocket, but the surface speeds are much lower on the small diameter hub. Both are good things. I plan on teflon tape around the hub and a dry felt seal media contained in a floating seal retainer.

snip
I bought 1/2" thick felt on Amazon - I get the F7 1/2" x 12" x 12". I think I got these part numbers from Dave Anders?

I closed off the cowl front bulkhead area, and cut a hole 1.5" larger than the shaft dia in the fiberglass. I cut a hole in the felt about 1/2" smaller than the shaft, then cut the felt in two (rounded to match the cowl bulkhead). I made an alum panel the retains the felt. I did put some liquid Teflon into the edge of the felt. I see that the felt polishes the shaft like Dave said it would; the shaft needs no protection (tape etc). I doubt this setup will need any maintenance, as it has not changed since I put it in.

I see the felt is pushed fwd around the shaft, so it must be seeing increased pressure too...

I recall the felt costing about $20...

My change to the round inlets was the most dramatic change I have ever made. CHTs dropped to low 300s and oil went to ~160F - so I had to tape off an area of the cooler. Previous temps were 400 CHTs and 240 on the oil - not the best numbers...

The strange and confusing part was that the aircraft speed did not change. So, it seems that I should remove some outlet area, and add a cowl flap..and have my own Incredible Shrinking Outlet.
 
Great minds Mark...

I have the big inlets now but after looking at the benefits of a sealed lower cowl and the dynamic extraction of an exhaust augmentor, there is compelling evidence that adequate Pd can be maintained between the upper and lower deck with significantly reduced inlet size. Look at the inlets Andrew F is using on his Lancair - he is cooling 700 HP with less than 40 inches of inlet area! I'm set up for 8 inch round inlets now (100 inches total) but plan to do significant flight test to choke those down to optimum. As a contrast, my outlet is only 30 inches with the cowl flap closed.

The inlets can be modified with easily fabricated restrictor plugs and the exhaust can be opened up with a reduced size splitter duct. It was designed this way to be modular and facilitate flight test. Once I get it correct, the new "final" cowl gets built in carbon fiber and I'll be done. In the meantime my Rocket will be sporting rough primer and filler.
 
Last edited:
prop shaft seal

Mark, on Bill's Rocket I did a conical rubber seal (Vans baffle rubber seal material) similar to Dan's scheme that runs on the extended prop hub. It resulted in a slight smudge on the surface, no surface wear on the hub. I planned for a lot of engine movement so the seal had a lot of flex room.
 
My change to the round inlets was the most dramatic change I have ever made. CHTs dropped to low 300s and oil went to ~160F - so I had to tape off an area of the cooler. Previous temps were 400 CHTs and 240 on the oil - not the best numbers...

The strange and confusing part was that the aircraft speed did not change. So, it seems that I should remove some outlet area, and add a cowl flap..and have my own Incredible Shrinking Outlet.

Impressive temperature drops. Mass flow must have increased to drop the temps but without constricting the outlet to regain lost momentum, I wouldn't expect to see much in the way of speed gains. Cowl flap allows you to optimize exit area for climb and cruise. That should gain you something.
 
The strange and confusing part was that the aircraft speed did not change.

Just imagine some softly rounded bazoombas out in front of those inlets. The fact they are made of increasing static pressure need not cool your ardor ;)

Look at the inlets Andrew F is using on his Lancair - he is cooling 700 HP with less than 40 inches of inlet area!

Careful with that line of thought. Andrew has the magic of velocity squared working for him. Standard day dynamic pressure at 350 knots and 5000 ft is 68" H2O. Here in the sluggo world, we need to cool RV's at 120 knots at full power for climb...where available dynamic pressure is 8".

68 is better than 8 ;)
 
Just imagine some softly rounded bazoombas out in front of those inlets. The fact they are made of increasing static pressure need not cool your ardor ;)

Sometimes I wish this forum had a "like" button! Made me smile..... a lot!
 
OK, I'm going to demonstrate my untrained ignorance. If Bernoulli says that speed and area have an inverse relationship, and speed and pressure have an inverse relationship (redneck engineering expressions), then I'm struggling to see how putting an expanding cone on the exit of a heat exchanger (when it's located in a volume of basically static air) will reduce pressure at the heat exchanger exit.

When I struggle through books like Kuchemann & Weber, and Kays & London, 'diffuser' is the term used to describe an expanding duct that *slows the air and increases pressure*.

Please educate me.

Charlie

HI Charlie, lets see if this helps:
The oil cooler is dumping its flow into the lower cowl plenum, and as always with subsonic flow exits, the pressure in the oil cooler flow at the exit must match the pressure in the local environment around the exit - which is the lower cowl plenum pressure. It doesn't matter what the relative flow velocities are, the pressures must match. With no exit diffuser on the cooler, this results in a bunch of little jets of flow coming out of all the cooler passages, at whatever velocity is needed to match the pressure to the local environment at the exit. If instead, you put a diffuser on the cooler, and lets just say it flows ideally with no separated flow, all those little jets of flow out of the cooler coalesce into a flow that can now slow down and increase pressure in the diffuser, until once again that flow must match pressure with the local environment around the exit of the diffuser. In both cases (with and without a diffuser on the cooler) the lower cowl pressure is about the same. So if the pressure at the diffuser exit matches the lower cowl plenum pressure, then the pressure must be lower at the entrance to the diffuser, at the cooler exit. So what the diffuser has done is lower the pressure at the exit of the oil cooler, and now with more pressure differential across the cooler from inlet to exit, there must be more airflow through the cooler.

Hope that helps. Keep banging on us - -there are no doubt others that want to understand, just as you do. Happy to try to help.
 
...Careful with that line of thought. Andrew has the magic of velocity squared working for him. Standard day dynamic pressure at 350 knots and 5000 ft is 68" H2O. Here in the sluggo world, we need to cool RV's at 120 knots at full power for climb...where available dynamic pressure is 8"....

That's true but he also has a bit of help from the aurgmented exhaust too. As evidence I have had some extensive conversations with a well known high HP RV driver with tiny inlets that can demonstrate a climb from sea level to near the flight levels at a constant 110 indicated and never even get close to 400 CHT. He attributes this performance significanly to the augmented outlet.

If you can cool 230 HP continuously with upper deck dynamic pressure in the low single digits then there is some magic at the outlet to create the delta P through the fins.
 
Augmentors are really magic when done right. Look at the size of the inlets on Andy Findlay's Lancair. He's cooling a lot more than 700hp, though some of that is happening with spray bar water.

The big deal with augmentors is you can regain all or most of the cooling air momentum lost passing through coolers and fins. Of course there is a lot of energy available at these power levels from the exhaust.

If the aim is to boost flow through coolers, that not so hard with some attention to detail, if we want to reduce drag at the same time though, we need to increase the velocity at the exit and direct it aft. That is much harder to do on an RV with the limited space available under the cowling and layout of the parts.

Ideally, if we had room, we'd want to do something like this with the oil cooler ducting:

 
There's quite a bit of real estate aft of the firewall on the RV that can be used to corral and accelerate the exit air. Here is a high fineness ratio (4:24) aft body stuck on my Rocket:


ae1eex.jpg



Only had to drop the chin of the cowl an inch, so minimal increase in frontal area but should really help out the airflow under there.

With the built in exit tunnel on the -8, a proper augmentor system would be even easier, me thinks. If I wasnt so burried in airplanes I'd slap something on the -8 and run some tests.
 
Last edited:
I like this Michael. Thinking outside the box er cowl. Avoids a some obtuse angles to align flow when done inside the confines of the normal cowl space. High fineness is the key to smoother flow from my tests.
 
There's quite a bit of real estate aft of the firewall on the RV that can be used to corral and accelerate the exit air. Here is a high fineness ratio (4:24) aft body stuck on my Rocket:


ae1eex.jpg



Only had to drop the chin of the cowl an inch, so minimal increase in frontal area but should really help out the airflow under there.

With the built in exit tunnel on the -8, a proper augmentor system would be even easier, me thinks. If I wasnt so burried in airplanes I'd slap something on the -8 and run some tests.

Long gradual (high fineness) shapes are best for preventing flow separation, but long passageways with high velocities have a lot of skin friction loss. A really good augmentor (you mentioned Andy Findlay's) is a little shorter than theory would predict because the reduction in skin friction is worth some loss of complete mixing and momentum transfer. Without an augmentor, you really want to get the flow turned and accelerated, and then just end it. You don't want to have the high velocity flow running in a channel for no reason.

I see you did get access to an L-39! While those tip tanks are off, perhaps you would like to fit a set of our CEllipse tips?
 
Long gradual (high fineness) shapes are best for preventing flow separation, but long passageways with high velocities have a lot of skin friction loss. A really good augmentor (you mentioned Andy Findlay's) is a little shorter than theory would predict because the reduction in skin friction is worth some loss of complete mixing and momentum transfer. Without an augmentor, you really want to get the flow turned and accelerated, and then just end it. You don't want to have the high velocity flow running in a channel for no reason.

My flight and lab tests generally showed you need some decent length to avoid separation so it's one of those compromises again. Just long enough to avoid separation and no longer. Would be very interesting to measure the velocity at various points on this duct. Are you up for that Michael?
 
In this case the "ducts" are 12 inches long (aft of the firewall). These are augmented, as the exhaust will terminate right at their inlet within the cowl. The centerbody and yet to be installed L/R skins are an extension of the cowl so are not technically part of the ductwork. As said before, much of this is test only and will likely make way for proper augmentor tubes and fairing bits (like Andrew's).

As for velocity measurement, sure. Testing is what this is all about

Yes, Bob K has hit me up for your wingtips for the jet and I'm still thinking it through. Focus right now is to get the CI accomplished and get it flying.
 
Last edited:
Were getting off track with your thread, Steve, but I wanted to give some context to the cowl shape:

t5kf3d.jpg


Hard to see in this picture, but the bottom of the cowl from the induction inlet back to the lower fuselage skin at the spar is almost a straight line. There is a slight kink at the firewall and the drive to keep that line flowing is why I have such a sweeping aftbody fairing.

And of course the long, flat lower cowl was driven by my need to have a long, "technically ideal" divergent duct airbox:

2v3frbq.jpg



2ensq6u.jpg



fbvpd3.jpg



351enap.jpg


Bottom line: one "simple" change drives a whole lot more.

Back to your regularly scheduled thread...
 
Last edited:
Hey Steve, the inlet "diffuser" to that cooler looks pretty square, any chance some pressure/flow gains available there?

Off topic: maybe DR could/should make a topic category for "Data Acquisition" or "Instrumentation" or such, to capture specific ways we all gather data. Pressure, temperature, and frequency. It would be great to get all those ideas on one section. For instance: I use a 4 channel, stand alone, thermocouple data logger just to avoid the extra ancillary equipment of the DatQ box. The DatQ boxes are much more versatile. Also, the clever/compact little pressure sensor DanH had is likely one of many things we could use. Data is sorta easy to obtain, cost effectivity and accurate is another matter.
 
...Off topic: maybe DR could/should make a topic category for "Data Acquisition" or "Instrumentation" or such, to capture specific ways we all gather data. Pressure, temperature, and frequency. It would be great to get all those ideas on one section...

I support this too. In fact I have pushed for and received verbal agreement to start a "DT/OT" (Developmental test/operational test) on another forum. Each thread will document an idea from concept, fabrication, test and conclusion. It's intended to become a repository for ideas that are validated or disproved via the scientific method.

Not sure we have enough interest here in the overall VAF group, but I'd sure like to see it.
 
Hey Steve, the inlet "diffuser" to that cooler looks pretty square, any chance some pressure/flow gains available there?

In hindsight, you are probably right. At the time, my thinking was that the velocity coming into the plenum box thru the SCAT hose is rather low, and so any loss is very small. However, I also used 3.5" SCAT, mainly because the butterfly valve fit on the rear cooling baffle better, but also because the slightly smaller hose fit through the engine mount tubes better. When I add up the effective flow area inside the cooler, through all the fins, etc., I find that the area is not much smaller than the flow area in the 3.5" SCAT. So I probably have more velocity, and more loss there than I originally imagined.

That being said, If I had used one of the commonly used fiberglass adapters that form a blended, but very short, wide-angle diffuser, with the moderate velocities involved it probably would also separate and cause just about the same loss.

I'm thinking pretty strongly about increasing to 4" SCAT. But since Ross and Dan implored me to take some measurements, I will measure what I have first, then see how much it changes with bigger SCAT.

If I had it all to do over I would probably mount the cooler perpendicular to the firewall with a mount that also forms the exit diffuser, and then make a smooth inlet adapter as well. Pretty hard to do that now on the finished airplane. But if the engine ever comes off.....
 
This will also be a good trial of the high-temperature hardener system (3136R) I got for the Jeffco 3107 (Rhino) epoxy we use.

Steve, quick side note. Recall I built fairly extensive oil cooler ducting. It's all plain old West 105. The loads are so low so I don't think Tg is a concern; no evidence of a problem after several hundred hours.

On the flip side, the lower cowl exit panel and door is all built with the Resin Services stuff we talked about a few months ago. Recall the RS spoxy has funky specs, so please do report if you can get the Jeffco 3136R loaded and hot. The glass down there is subject to a lot of radiant pipe heating.

OC%20Duct%20Layup.JPG
 
Last edited:
My thoughts from the beginning of the thread match BillL's comment.
I have a firewall mounted cooler and had oil cooling issues from the start. I have been through a few different modifications to get the temps down.
I was limited to 3" SCAT.
I started with the square'ish transitions sold by Van's. This was horrible and I assume only a 3" round portion of the cooler was really getting air. I made a fiberglass cooler inlet first which helped a lot. Made again with better entry angle. Third I made a connector for the back of the baffle which transitioned from larger rectangle to the 3". No exit diffuser but temps are now great.
 
I made a fiberglass cooler inlet first which helped a lot. Made again with better entry angle. Third I made a connector for the back of the baffle which transitioned from larger rectangle to the 3". No exit diffuser but temps are now great.

Got any photos of these parts?
 
I do not normally post pics of my fiberglass work.
Mine is not the value porn as Dan's work.
Kinda like Roseann next to R Welch.

I only have pics of the inlet end. By the way each change in configuration made a 5-10 degree difference in temps.

I chose to improve the flow through the cooler rather than replacing with a larger cooler or trying to make a larger SCAT fit.

I found these pictures from my first hand model photo shoot :)
5zha9g.jpg
[/IMG]
mjyg7k.jpg
[/IMG]

This is the original cooler/hose transition.
2sao084.jpg
[/IMG]
6ga25j.jpg
[/IMG]
 
this is how simple it is to do the cowl spinner seal.
9

if there isn't enough space to seal to the hub, you can easily make a 90 degree circular flang and place the felt aft to seal on the angle of the flywheel.
 
this is how simple it is to do the cowl spinner seal.
9

if there isn't enough space to seal to the hub, you can easily make a 90 degree circular flang and place the felt aft to seal on the angle of the flywheel.
Dave, anything special about that felt? I like this better than mine, it is baffle material and touches the flywheel. That makes it hard to remove the lower cowl. Thanks for posting!!
SDHuxD5Or_2-CfQk9z1FcDMZY_T0lHRf-91oCJPe1W4yuPCmZgX6LHvttDXd7bEbrnpm3uoSEYaHM8yvfJ1aTUjngERoCkCIvmgE-eKy6gD-SuiRqdg2B96dXn46HnHtExjM4pTqSWPLU48gc_ayFQdrM_13twa-zRKlyYjdtLd8m2TvCY67RUF2wE0d12sI0Mn7sUQgWfDY9QLTlBAZoHl48xfXhGmcMAf_X5ak2GUvujv721Kyet16vVJF9tekfFlUWmc6rT3tHMUzpV2Twwwkgo8rNfCjO8W8TC4fYOYd8Wrk92kdMCD2wlNk_m9g27R7EH-17Lz8j7BIk51OWUrhW6LCoPwIBHge2ovBBwRvJ96vSXBBj8C10zA03xJW5Pt0xFunoJ4Pv6BiVpMoqKNcCCdTflUe7RSYk1A9_UiMJMW4Y2EFPS8ngnmnpCD5oAfGA5llMf28dpjNu1vJmVE6VlY2rblh-qPztzM2YDULPFYzS88gEAiu68kJfmpjysQviafRIB2fRGp5cAckSIkFbA_DTLrWYiSomIcQczAc46zXpZ4eofr07QzS1H3njUp7wT8be_MfuLJujrLiQr0Z9UBeIlf2fXjmNXEN-_AjYtBW23LWfWwkrRN33-BYsl0VrGQzSWJNvu0hZcQDYCqfK3JlS3mUEMMRi5JBBJoxIj96jsr2peaJmctGfjFrVgEzS6fOoe6gfv4JayXAwlPa=w800
 
Last edited:
bill i wish i has your computer knowledge. i couldn't even post my own pic. you can see how much defection there is in the felt seal at the prop hub. this is with about 2000 hrs on the same seal. all it does is polish the hub. it was cut with a bevel at the surface where it contacts the hub. the felt grinds pretty good 90 degree die drinder and a coarse disk then it was soaked with silicone spray along the contacting surface and the more P2 pressure the better it seals.
 
bill this felt from mcmaster carr is quite flexible and non abrasive. the thing i like is it produces a fairly large sealing area. i can't even feel it when i remove and put my cowl back on. the benefit of the soft wide seal area is, i think it, seals even in turbulent conditions. mike was correct regarding allowing about a 1 1/4 - 1 1/2" space around it. the guys at reno that have and were using it haven't had any problems even under those conditions. i've never had it contact the aluminum backing plate even with my crummy aerobatics.
 
That looks like a great way to do it with the hub Dave has.
It looks like you cut off the normal flange on the cowl that extends back toward the flywheel at 45? or so.

My hub does not have a round surface like that to seal to unfortunately. McCauley hub for WW 200RV.

But the felt might still be a good choice to seal to the flywheel.

I might also look at sealing to the carbon-fiber spinner bulkhead. I seem to recall it is kind of conical, but I need to look at it.

One of the things I've worried about is that my cowl seems to shift around some. When I remove it and install it, the cowl edge is a perfect fit to the spinner. After the engine runs, the cowl seems to shift a bit so the cowl edge sits about 1/8" up above the spinner. I have MilSpec 1/4 fasteners on the cowl attachment to the firewall, and I think they allow some float. So, unless the seal has about a 3/16" compliance, it is going to gap on one side or the other as the cowl shifts.

With the felt, I think I can let it rub like Dave has with enough compliance so it won't gap.
 
Back
Top