Engine Cooling Revelation

[swpilot3]

Hi Everyone. I've been doing some testing of various configurations to try and help lower my CHTs. I've recently had a breakthrough so I thought I'd share.

First, a little background. I'm out in Arizona and in the summer I've struggled to keep my CHTs at anything reasonable. I have an O-540 turning 2700 rpm making 260hp. After reading Mike Busch, I've adopted 415F as my "do something now to fix it" temp even though Lycoming lists 500F as the max temp. To be honest, I'd be elated if I could keep them below 415.

Up until now, I haven't done a great job documenting with concrete numbers. I've read forums, talked with people and implemented what seemed to make sense. I double-checked all the baffeling. I increased my outlet area, then increased it again, then increased it again! I placed a lip at the front of the outlet area, then made that lip bigger. I added a "fence" from the cowl lip to the tunnel. I added a small 1" aluminum round (don't know what to call it, if anyone can tell me, I'll edit this description.) at the bottom of the firewall, at the front of the tunnel. Then I made it bigger. With still unacceptably high CHTs I decided to add louvers. I decided to either go big or go home. So I made large louvers on the sides of the cowling based on lots of reading.

After all of that, I could sometimes manage the CHTs... Until the summer days came.

So I decided to get scientific. I had a homemade manometer, but it's hard to read and bounces around a lot. So I found a monometer on Amazon. https://www.amazon.com/gp/product/B07K7HT3XJ/ It's not too expensive and measures the same as my homemade one. However, this one has an average feature. Which is perfect for our use. Hold an airspeed for a few minutes and look at the average reading.

For my baseline, I used the current configuration with a 1.5:1 outlet to inlet ratio, a 2" lip, fences, large firewall round, and large louvers.
At 92 KIAS it was pulling 3.50 inches of water.
At 105 KIAS it was pulling 5.10 inches of water.

I found that the size of the firewall round didn't make any difference at all. However, what was interesting was that my louvers were making the differential pressure worse. With the louvers removed:
At 92 KIAS it was pulling 3.69 inches of water.
At 105 KIAS it was pulling 5.15 inches of water.

So now on to what I discovered. I decided to replace the cowl opening and lip with a one-piece fixed cowl flap. Same exact outlet area.
At 92 KIAS it was pulling 4.20 inches of water.
At 105 KIAS it was pulling 5.70 inches of water.

For my installation, a large cowl flap seems to do a substantially better job of cooling. Especially at lower airspeeds. At a cruise yesterday at 2500ft, 23 squared, with 108F OAT, this configuration lowered my CHTs from 430F to 380F. This substantial!!! On the descent, the CHTs were falling to around 310F.

Since this fixed cowl flap is proof of concept, now I'm going to build a proper cowl flap that I can control.

If you're dealing with high CHTs with a big engine, don't take this as my recommendation on what to do. Just take it as a datapoint while you research solutions. YMMV. Figuring out what the air is doing inside your cowl is difficult. However, with my particular installation, this seemed to work!

IMG_3602.jpeg For reference, you can see the louvers that I had before (that only hindered) and the fixed cowl flap. Please excuse the dirty airplane!

[jaredyates]

What a fantastic post Bobby! Thanks for the data and please keep us updated on the progress.

For what it's worth, here's what Jimmy Tubbs had to say in his very good article on the topic:

..Lycoming Engine Installation Manual does specify that the O-320 engine requires 5-1/2 inches of water while the O-360 engine should have 6-1/2 inches
pressure drop for good cooling...

[davzLSA]

Hey Bobby thanks for the post, in my opinion this is what EXPERIMENTAL aviation and all about. Solving problems and you knocked this one out of the park!!! I believe this info can be applied to any airplane with an ICE engine.

[swpilot3]

Thanks Davz! It was a fun process. Especially when you have an "Ah-ha!" moment!

Thanks, Jared. I'll go back and re-read that article. I do get around 6-1/2" at a fast cruise. I'm gonna do some homework and see if I can find that same chart for my engine.

After I build a proper cowl flap, I'm planning to build a plenum for the baffling, just to rule out any issues with that.

[Rollie]

Good stuff, I'm in the process of building a new lower cowl, hoping for better cooling along with some other mods I'm doing. Did you consider trying different size inlet openings? I've been thinking smaller openings in the nose bowl might be helpful. I see these tiny round inlets on fast planes, and maybe they only work well going fast, but they don't seem to overheat on climb out. We can only go so big on the outlet area so another way to get the pressure differential might be to go with smaller inlets. I'm going with the idea that air movement is more important than air volume.

Once I get mine back in the air if I don't see the results I'm looking for I'll try changing the inlet size and shape but I figured since you are actively testing this sort of stuff right now I would suggest it.

[swpilot3]

I've noticed the same thing. It's an interesting point. I read a pretty good AOPA article about cooling and inlet size. Here: https://www.aopa.org/news-and-media/...powerplant-(4)It says the smaller inlets were done on more modern aircraft to reduce cooling drag. They were able to do this because they've become better at optimizing engine cooling. So it doesn't seem that a smaller inlet actually aids in cooling. They were just able to get away with it because they were able to cool by other means... redesigned cowling, modified cylinder fins or placing the inlets farther away from the prop.

"Minnis says the biggest reason for the reduction in the size of cowling air inlets is an improved understanding of the effects of the propeller hub and propeller spinner on airflow. It's now known that the airflow in this area is very turbulent and greatly disturbs the orderly flow of cooling air. Current designs position the cooling-air inlets away from the propeller hub area."

Additionally, the first couple inches of the propeller doesn't create any lift, so it just blocks airflow. If the inlets were place farther outboard, then perhaps they would be behind the lift generating portion of the prop and that could be beneficial.

Oddly enough, the author seems to insinuate that cowl flaps aren't needed in modern designs. Maybe I need to rethink my directon.

The other concept that I keep reading about is that cooling is a function of time. It takes time for the heat to transfer from the cylinder to the cooling air. If a small volume of air is going by the fins very fast, it's not as effective as a larger volume of air going by slowly. Obviously, too slow and it's ineffective. There's a sweet spot. That's why I think the Lycoming installation manual calls out a specific volume at a specific pressure.

Tony Bingleis (old school but tried and true) recommends multiplying your HP by 0.35 to determine inlet size. So for my engine that comes out to 260 * 0.35 = 91sq in. The BH has about 84sq in of inlet.

So perhaps, if the nose bowl was going to be altered, the best thing to do is move the inlets outboard, away from the spinner. Then you can decide if you want to play with making the inlets bigger or smaller.

[John Bickham]

Thanks for that info Bobby. I'm just a little east of you in NM. Figures I will be doing my initial break-in and Phase 1 in August, I hope. I may need some of this. I think I have the same engine, Barrows carburated O-540. I might add one other little nuance that can have an affect that is counterintuitive. I retired from the petro chemical industry with some practical experience with heat exhangers. One thing that took me a long time to accept was there is an optimum velocity for heat exchange. Most of us believe more and faster is better. There is a point that too high a velocity actually has an opposite effect. I've seen first hand where moving the rear baffle wall back to create more clearance and slow the air past the rear cylinder helped a lot. That is only helpful if a particular rear cylinder is your temp problem.

Maybe we can visit in fall??

[swpilot3]

Hi John. That's good insight. Absolutely, I live at an airpark on the east side of Phoenix. Fly out when you get your plane going.

As a matter of analogy, I recently had a pool built in our backyard with a spa. After it was all done the pool builder was programming the app that controls the spa. He said one of the big advantages of a variable speed pool pump is that while the spa is heating, he can program the pump to run at a very low speed.

Counterintuitively, this actually heats the spa faster. This allows the water that is being pumped through the heater time to pick up more heat, then be delivered to the spa. If we ran the pump at full speed while the spa was being heated, it would take longer. Fascinating! This is all starting to make sense now!

That's also a good point about moving the back baffle wall. I hadn't thought of that. I might try that also. How far did they move the back baffle wall?

[John Bickham]

I'll double check. I've been hitting my head on the plane a lot lately and not sure about my memory. I'm pretty sure it was under a 1/2". Doesn't take much over 6 or 8 inches to slow things down.

[Russellmn]

As a matter of analogy, I recently had a pool built in our backyard with a spa. After it was all done the pool builder was programming the app that controls the spa. He said one of the big advantages of a variable speed pool pump is that while the spa is heating, he can program the pump to run at a very low speed.

Counterintuitively, this actually heats the spa faster. This allows the water that is being pumped through the heater time to pick up more heat, then be delivered to the spa. If we ran the pump at full speed while the spa was being heated, it would take longer. Fascinating! This is all starting to make sense now!...

This is the same principle that allows a liquid cooled engine to often cool better (run cooler) WITH a thermostat than without. The old farmer trick with an overheating tractor of removing the thermostat often doesn't work anymore.

[yateselden]

Removing a thermostat works if you replace it with an orifice. This has worked well In race cars over the years. Start with a 3/8 - 1/2 hole and drill as needed.

[Russellmn]

yup, gotta slow down how fast the water moves through the radiator, orifice does that pretty well too, just takes more experimenting to get the right size.

[John Bickham]

Just checked with my mentor on the exact measurements of his rear baffle move. On an IO-390 in a Carbon Cub, he fought high CHT for two summers. He moved the rear baffle wall back 1/8" on that one and hasn't looked back. He has a Murphy Rebel with an O-320 and on that one moved it back 2" because he had room. Fixed high CHT's on #2, and #4.

That all the practical data I can muster. He mentioned using spacers with ALUM pull rivets till you get the results you desire, then go solid rivets. Hope this somewhat helpful. Back to swaging aileron cables for final rigging.

[Pbruce]

I’ve owned two 0-320s, a 150hp Model on a 172 and a 160hp on an RV-9. I fought high CHTs on both, particularly on the rear cylinders, especially the right rear. I tried a bunch of things: baffle seals, caulking all holes holes and cracks, forming better flow behind the spinner inside the RV9 cowl with foam, and nothing really worked. I understand the parallel-valve 0-360s are similar in terms of cooling needs, at a slightly larger scale.

I also have read of guys who moved the rear baffle back, giving 3/4 “ of space behind the cylinder and they swore it cured the issue. I went to local experts for help in doing this mod, and they all refused, saying it was non-sensical, that I needed to squirt cold intake air at high speed past those cylinders. I no longer own either plane, but I sure wish I had done the mod on the RV9, so I could better judge whether it was worth doing on my Patrol. I may go with a plenum, and if my building skills allow, I will try to find a way to make the rear baffle position adjustable without tearing the whole thing apart.

Some other thoughts I’ve been pondering: I wonder if vortex generators or other “turbulators” inside the cowling might be of use in mixing and slowing down the airflow inside the cowl. Bernoulli also suggests a plenum with more volume on top of the engine would provide higher pressure. Most cowls do the opposite: small volume on top and large volume below. I’m interested in any opinions or experience on this topic.

[zkelley2]

Awesome. I'm looking for lower CHTs at the moment. I might give this a shot.

[swpilot3]

Fantastic info John. Thank-you! I will give this a try for sure.

[whee]

Awesome info and interesting discussion.

I think it’s very interesting that your louvers made things worse. While not the same, my cowl flaps are positioned on the cowl cheeks like your louvers and they significantly aid in cooling. Opening them can take my CHTs from 410 and rising to a steady 370ish.

While velocity (residence time) does play a role in heat exchangers I not sure doing things to alter the air velocity will make a significant difference in our situation. I know I’m stating the obvious but cooling or heating a liquid isn’t the same as air cooling an engine. But discussing similarities does help develop ideas.

I like having cowl flaps. Sure you can get along without but in some climates that means having a hot engine in the summer and/or a cold engine in the winter. I can’t flying my airplane in the summer without the cowl flaps open during some phases of flight. During the winter if I leave cowl flaps open my engine is way too cold.

Probly worth noting the Conti IO360 has a reputation for running hot and being hard to cool. I haven’t found that to be the case and think it probly gained that reputation in the turbo applications and the rear engine of a Cessna 337.

[swpilot3]

410 to 370ish is amazing!!

Can you show us a pic of your cowl flaps?

[whee]

I thought I had pictures but I haven’t found them yet. If I do I’ll post them.

[swpilot3]

Quick update. While I'm waiting for materials to arrive from Aircraft Spruce to build my cowl flap, I decided to make my fixed one bigger. It now opens 7" rather 4". Add this to the tunnel and it now opens a full 10". It feels ridiculously huge! However, flight tests revealed that it improved things a little more.

Big Cowl Flap:
At 92 KIAS it was pulling 4.45 inches of water.
At 105 KIAS it was pulling 5.90 inches of water.

This is a big improvement seeing as how I started at:
At 92 KIAS it was pulling 3.50 inches of water.
At 105 KIAS it was pulling 5.10 inches of water.

I think there's more work to be done figuring out what helps to cool CHTs, but at least now I can fly my plane in the summer!

[svyolo]

Hey Bobby, in regards to your louvres, they might work better you swapped the "aspect ratio". Yours are narrowand long with respect to the direction of airflow. You might have better luck if they were wide and short instead. Most of the louvres I have seen for this application are wider than they are long.

I also remember reading a while back about closing off part of the inboard part of the cowl opening. If it is too close to the spinner, it can function more as a leak than an intake. In other words, without the ram effect, air can more easily spill out of this area, reducing pressure under the cowl.

[JimParker256]

When Roy Lopresti took over as chief aerodynamicist for the Grumman AA-5 Traveler (developed into the AA-5A Cheetah and AA-5B Tiger), first thing he did was put a bunch of tufts on the cowl, including the inlet area. Pictures showed significant "reverse flow" where the air was flowing from inside the cowl toward the outside. Uh oh! Moving the cowl inlets outboard helped, and making them smaller helped as well. In the end, they also increased the size of the "exit area" at the aft lower cowl, and that was the final piece of the puzzle at the time. Nowadays, a lot of Grumman owners with "less than perfect" baffling use a "spoiler lip" in front of the lower cowl opening, which prevents upflow through the bottom of the cowl in high AOA situations. Seems to help a LOT! Perhaps applicable to the Bearhawk line, or perhaps not...