Weird CHT issue

[EdFred]

See picture for 1000's of words.
This is at cruise.
Cylinders #1-#5 were all put in brand new in April/May.
ECI Titan Nickel Bore

I should be past the break in period. 25+ hours
Break-in instructions followed meticulously.
Boroscope looks "perfect."
Oil consumption is normal/stabilized.
Oil temperature is normal - on the low end of the green range.
Oil pressure is normal.
Cylinder #6 (OH'd, been in for a year acts normal, never gets to 380)
Cylinders #2 and #4 act like #1 and #3 respectively but at a cooler temp. I've never seen #4 get above 400. #4 Usually stays well below 380. #2 is about 25 degrees cooler than #1.
The graph slides left/right a little with altitude changes.
No roughness to engine until I lean past peak.
EGT's act in relation to CHTs as the books say they should.

What say you?

 

[Archammer]

Is Cyl #1 closest to the oil cooler return line?

 

[flyingron]

#1 is the front right cylinder.

The first thing I'd do is check the baffling. It's amazing how just getting one piece MISSING or bent can mess up the cooling. There was a hunk of baffle on my engine that was out of place that you really couldn't tell unless you pulled the nose bowl off.

 

[Archammer]

Yep or a limited injector, and one thing that I know gets missed is probe position/fitting in a lot of cases. It may not be the actual temp but an inaccurate reading.

What about a cowl flap problem on that side?

 

[RotorDude]

See picture for 1000's of words.
This is at cruise.
Cylinders #1-#5 were all put in brand new in April/May.
ECI Titan Nickel Bore

I should be past the break in period. 25+ hours
Break-in instructions followed meticulously.
Boroscope looks "perfect."
Oil consumption is normal/stabilized.
Oil temperature is normal - on the low end of the green range.
Oil pressure is normal.
Cylinder #6 (OH'd, been in for a year acts normal, never gets to 380)
Cylinders #2 and #4 act like #1 and #3 respectively but at a cooler temp. I've never seen #4 get above 400. #4 Usually stays well below 380. #2 is about 25 degrees cooler than #1.
The graph slides left/right a little with altitude changes.
No roughness to engine until I lean past peak.
EGT's act in relation to CHTs as the books say they should.

What say you?

I don't trust Cyls #1 and #5 readings. Unless those are absolutely known to be correct, any further analysis is flawed.

 

[Henning]

Induction leak?

 

[EdFred]

#1 is the front right cylinder.

The first thing I'd do is check the baffling. It's amazing how just getting one piece MISSING or bent can mess up the cooling. There was a hunk of baffle on my engine that was out of place that you really couldn't tell unless you pulled the nose bowl off.

First thing I checked. Especially with it limited mostly to the right side of the engine. Nice thing about the Comanche is I can get a good view. Nothing is missing, but I could probably check the rubber/poly material again and see what's going on.

Yep or a limited injector, and one thing that I know gets missed is probe position/fitting in a lot of cases. It may not be the actual temp but an inaccurate reading.

What about a cowl flap problem on that side?

Carburetor. #1 is a ring probe because the baffling rod is in the way to put the 'normal' CHT probes in. Could be an issue with #5 though. Also, no cowl flaps.

I don't trust Cyls #1 and #5 readings. Unless those are absolutely known to be correct, any further analysis is flawed.

#1 will be easy enough to check. I'll just put it on #3 and see if it follows the trend on 3. And if both #3's check out, I'll put #3 in #5 and see what happens there.

However, with my luck, the probes are probably good.

Oh, also, I am running REM40E plugs, and I am going to swap those out for 38's tonight. I'm up for trying just about everything.

 

[EdFred]

Well, it's not the probes. Messed with those tonight. The ring probe on 1 indicates 20 hotter than the 'stick' probe on 3. (ran them simultaneously). Double checked the baffling. No gaps, all looks good there.

 

[EdFred]

An update on this.

As the cylinders got some more hours on them 1-4 all stabilized and dropped. Now in cruise, 1-4 all run under 340 degrees, often down into the low low 300s. But number five remained a pain in the ass and unless I feed it 18+gph was always above 400 - like 430 in cruise. ECI said just keep running it, that's not that bad. But it kept doing so even up to 50 hours, and it shouldn't.

So, we did the following:

All new right side baffling - didn't help (more on that).
Disconnected the standby vacuum line and sealed off the fitting - didn't help.
Swapped probe channels - not a JPI issue.
Swapped probes - not a probe issue.
Pulled off #5 and #3 cylinders and swapped positions - problem stayed with the position, not the cylinder.
Upon further inspection new baffling at the back was not sealing up due to a channel that runs the length of the cowl. Made slits in the baffling. Baffling now sealed up against the cowl - didn't help.

So finally, yesterday we did the following:
Swapped intake tubes between 3 and 5, disconnected the air inlet tubing to the carb heat box/shroud and blocked that off.

Last night I took it up at 4500, 5500, 6500, 7500,and 8500 feet and flew it for 10 minutes at each altitude. WOT, 2300 RPM, leaned to rough, and then just back to smooth. (varied between 14.5gph --> 11.2gph at all altitudes)

4500 - 405-410º (enriching higher than 15 helped here through 6500 - it would go as high as 440 before)
5500 - 400-405º
6500 - 395-400º
7500 - 385-395º (enriching from 11.2 - 14.5 didn't really affect CHT here)
8500 - 377-390º (or here)

If I leaned to just rough - all cylinder temps dropped pretty much in unison so that at least tells me that #5 isn't peaking early or late, and fuel flow is pretty even. #5 even dropped down into the 360's if I was just barely rough - of course, 1-4 were under 300 at that point as well and 6 was half way between them and #5.

#3 didn't change from what it was so we can rule out a leak in the intake tube/fitting. After the flight I did notice the rear baffling didn't stay curved forward, as the air pressure was enough to push it back - and there still looks to be some corners that are open due to some jogs.

So at this point I got a > 40º CHT drop just by blocking off the carb heat box air inlet. So, that seems to be the culprit on this so far. We are going to work on the baffling/air flow some more. RTV, silicone tape, longer baffle seals, whatever else we can think of. If I can get another 20 degrees of temp drop, we will then reconnect the carb heat air intake and see what the difference is. Then work on fixing that - or rerouting the intake air.

 

[Chip Sylverne]

My number 5 runs hotter as well, though the difference is closer to 15-20 degrees, but I have a cowl flap. I think part of the reason is that the boss for the intake side of that cyl. butts up right against the rear baffle restricting airflow, and since it's not finned it doesn't shed heat well. One of the guys on the Comanche board fabbed up something he called the "baffle bump" that creates a little space back there, increase airflow over that part of the head. Other guys have reported some success by wedging a little spacer in there to create a little gap and do the same thing. Others still report some success by extending the baffle in front of number one up a little to direct flow up and over to get more air to the rear.

Don't forget to seal real well around the engine mounts.

 

[EdFred]

My number 5 runs hotter as well, though the difference is closer to 15-20 degrees, but I have a cowl flap. I think part of the reason is that the boss for the intake side of that cyl. butts up right against the rear baffle restricting airflow, and since it's not finned it doesn't shed heat well. One of the guys on the Comanche board fabbed up something he called the "baffle bump" that creates a little space back there, increase airflow over that part of the head. Other guys have reported some success by wedging a little spacer in there to create a little gap and do the same thing. Others still report some success by extending the baffle in front of number one up a little to direct flow up and over to get more air to the rear.

Don't forget to seal real well around the engine mounts.

Was out at the mx hangar this morning. We moved the rear baffling back, by adding a strip of aluminum to where the rear baffling plate screws/bolts to the side plate, and essentially moved those holes. The plate was tight up against the cylinder before, and now there's a gap that's about the same size as the one between the cylinders. Hopefully this helps some more. Also went RTV crazy, and added some more silicone pieces in the corners. Unfortunately, the ceiling isn't that high, and the wind is gusting to 44 in places, soooooooo, testing will wait another day.

On the "odd" side the engine mount is completely behind the baffling.

 

[nrpetersen]

Bear in mind that the air pressure drop across the baffles is about 1 ft of water at 112 mph, rising by the square of the speed ratio. That would be 2 ft of water head at 158 mph etc...

 

[EdFred]

Bear in mind that the air pressure drop across the baffles is about 1 ft of water at 112 mph, rising by the square of the speed ratio. That would be 2 ft of water head at 158 mph etc...

But if the air isn't going to where it needs to go, the pressure doesn't matter. Path of least resistance and all. That's what this seems to be about - coercing the air around the number 5 cylinder rather than to wherever it's going currently.

 

[Chip Sylverne]

On the "odd" side the engine mount is completely behind the baffling.

Yeah, but since it's one big plenum leaks anywhere don't help, right or left side. I also sealed the gaps between the inter-cylinder baffle plates and the case with RTV so air has to exit through cooling fins.

I believe it/ called "frog- butt tight".

Real interested to hear how you make out. This is on my list too.

 

[EdFred]

Yeah, but since it's one big plenum leaks anywhere don't help, right or left side. I also sealed the gaps between the inter-cylinder baffle plates and the case with RTV so air has to exit through cooling fins.

I believe it/ called "frog- butt tight".

Real interested to hear how you make out. This is on my list too.

Good point, I may have some more work yet to do with that.

 

[nrpetersen]

Yes the cooling air has to be forced thru the fins - and not just around them. That's why the baffles are as rigid and tight as they are so air pressure can't force them away from the fins.

Dirty fins don't help, nor do some rough cylinder fin casting flaws. I have seen some cylinder head castings where the casting pattern had slipped when the heads were made, making the air flow thru the fins partially blocked.

 

[EdFred]

Yes the cooling air has to be forced thru the fins - and not just around them. That's why the baffles are as rigid and tight as they are so air pressure can't force them away from the fins.

Dirty fins don't help, nor do some rough cylinder fin casting flaws. I have seen some cylinder head castings where the casting pattern had slipped when the heads were made, making the air flow thru the fins partially blocked.

Yeah, I get that, but when the opening on the back side of #5 is much smaller than the openings between 1,3,5; 2,4,6; and the back of number 6, and there aren't cooling issues on those cylinders, I need to coax that air to the back side of number 5 somehow. So, we moved the baffling back a little bit with some modification to the "L"

 

[Henning]

Yeah, I get that, but when the opening on the back side of #5 is much smaller than the openings between 1,3,5; 2,4,6; and the back of number 6, and there aren't cooling issues on those cylinders, I need to coax that air to the back side of number 5 somehow. So, we moved the baffling back a little bit with some modification to the "L"

You could create a seperate split duct to go over 1&3 and plenum over 5 with clean flow.

 

[nrpetersen]

Yeah, I get that, but when the opening on the back side of #5 is much smaller than the openings between 1,3,5; 2,4,6; and the back of number 6, and there aren't cooling issues on those cylinders, I need to coax that air to the back side of number 5 somehow. So, we moved the baffling back a little bit with some modification to the "L"

I know what you mean. The cooling fin gaps don't go all the way through on the intake side because of the intake manifold boss on (at least per my C-85). I don't have an answer other than to space the rear baffle back slightly as you suggest, but whatever air gets thru, force it once again to go through the fins.

I'm fighting CHT on my 1940 J4A & have noticed the same thing but I have not yet tried that. Gotta look hard again,

Actually though I think that there is comparatively little temperature gradient across an aluminum cylinder head. IOW cool it anywhere you can but don't worry about "hot spots". Easy to say but not to do......

 

[Henning]

Does the cowl flap position alter the distribution pattern? I forget, are you FI or carb?

 

[Anthony]

Maybe all new cylinders would help?

 

[EdFred]

Maybe all new cylinders would help?

Yeah, I'll get right on that.

 

[Henning]

Yeah, I'll get right on that.

I was just thinking, #5 is in the back correct? Is there any way for you to measure the pressure differential inside and outside the cowl above #5? If you have positive differential outside in the cowl zone, you could add a NACA duct above #5 (opening aft) to increase the airflow across it.

 

[EdFred]

Problem solved for now.

Saw a 30-35 degree drop in temps on the number 5 cylinder at 5500 when I went up today. Slightly less at 6500. Even leaned out in the climb I never went above 375º. Came back down to 3500, WOT (well above 75%), leaned to peak CHT - and was at 365. I don't know if we are at the sweet spot for spacing on the back baffle plate or not, but I got more out of it that I expected.

We did not hook the carb heat inlet back up, and it does not appear to be necessary as the carb heat without air coming off the cylinders. Now, for my next trip, I'm off to ___________. (I don't know!)

 

[Anthony]

^^^^^^ Great! Guess you avoided the firewall forward overhaul like I was going to suggest.

 

[EdFred]

^^^^^^ Great! Guess you avoided the firewall forward overhaul like I was going to suggest.

I had a different plan, that has been used by at least one other member here, in mind for solving everything if this didn't work.

 

[Anthony]

I had a different plan, that has been used by at least one other member here, in mind for solving everything if this didn't work.

A match?

:wink2:

 

[GMascelli]

Ed,

I had an issue with one cylinder starting to run hotter than the others in climb out but was close in cruise. I pulled the top cowl at annual and could see where the baffle was bending back and air blowing by. I added a new section and interlocked it with the sections on each side providing an overlap. Temps rock solid now in climb and cruise.

Amazing what a small gap will rob from the airflow. Glad to hear you have it squared away.

 

[nrpetersen]

Ed - By chance are your cowl sealing flexibles made from felt? Some earlier Pipers were made that way & I wonder how much air they allow to pass thru the felt.

Glad you found a fix. I'm gonna look for that on my J4A. THX

 

[Henning]

Problem solved for now.

Saw a 30-35 degree drop in temps on the number 5 cylinder at 5500 when I went up today. Slightly less at 6500. Even leaned out in the climb I never went above 375º. Came back down to 3500, WOT (well above 75%), leaned to peak CHT - and was at 365. I don't know if we are at the sweet spot for spacing on the back baffle plate or not, but I got more out of it that I expected.

We did not hook the carb heat inlet back up, and it does not appear to be necessary as the carb heat without air coming off the cylinders. Now, for my next trip, I'm off to ___________. (I don't know!)

Could very well be; the spacing effects where the back wall vortex affects the flow over the cylinder. Too close and the cylinder can be affected by reverse flow, move it further away and now the vortex down stream flow joins the the downstream path through the fins.

 

[EdFred]

Ed - By chance are your cowl sealing flexibles made from felt? Some earlier Pipers were made that way & I wonder how much air they allow to pass thru the felt.

Glad you found a fix. I'm gonna look for that on my J4A. THX

No, not felt. The "old" ones on the 2,4,6 side are the orange flexible silicone. the new ones are black silicone. I also went RTV crazy anywhere I could see an opening. On the 2,4,6 side there is still a decent sized opening where the motor mount is and at annual (April/May) we will look at closing that off, but for now things seem to be under control. I do know that if I ever designed a cowl for an air-cooled engine, all the cooling would be directed air rather than, "lets just push all the air into here and hope it goes out where we hope it is supposed to."

Or as another PoA member said when he texted me back saying, "I used to think concrete was the dumbest thing I ever worked with. But now that I fly, it's got to be airplane engines."

 

[Chip Sylverne]

Problem solved for now.

Saw a 30-35 degree drop in temps on the number 5 cylinder at 5500 when I went up today. Slightly less at 6500. Even leaned out in the climb I never went above 375º. Came back down to 3500, WOT (well above 75%), leaned to peak CHT - and was at 365. I don't know if we are at the sweet spot for spacing on the back baffle plate or not, but I got more out of it that I expected.

We did not hook the carb heat inlet back up, and it does not appear to be necessary as the carb heat without air coming off the cylinders. Now, for my next trip, I'm off to ___________. (I don't know!)

Great news.
I'll need a billing address...

About how large a gap did you make, 1/8" or so?

 

[nrpetersen]

Actually the cooling details are supposed to be worked out by the airframe manufacturers. But there isn't a lot of accommodation done by the engine manufacturers.

 

[Henning]

Look at some of the cooling plenum designs on the RARA Sportsman classes, you see some interesting stuff. One of the really interesting things you can observe is the realities of shock cooling wen you notice the spray system by the heads and the large tank of water in the pax seat that they spray on them during the race.

 

[EdFred]

Great news.
I'll need a billing address...

About how large a gap did you make, 1/8" or so?

In the neighborhood. I eyeball it at probably a heavy 3/16"

Actually the cooling details are supposed to be worked out by the airframe manufacturers. But there isn't a lot of accommodation done by the engine manufacturers.

Supposed to. They still have to put it together right.

 

[Chip Sylverne]

Did this mod over the weekend. Instead of jacking around with the "el", I cut a couple of 1" squares out of alclad and folded them in half and in half again so the look like 3/16 thick aluminum matchsticks. After loosening the nut on the threaded rod that holds the baffle in place, I coated them with a little high-temp RTV and inserted one vertically between the baffle and head close to where the baffle curves around the barrel, and the other outboard beneath the screws at the "el". Tightened up the threaded rod and reinstalled the screws to clamp them in place and now have a a nice, square 3/16" air gap behind the head.

Flight test OAT was in the 40's so hard to tell if here's any great temp reduction at this point, but seems to be about a 20 degree reduction in the climb on #5.
Worth doing' IMHO.

 

[SCCutler]

Any chance of photos depicting what you guys did? I don't have a PA24, but it is an interesting exercise.

 

[GeorgeC]

I get the impression that baffling is a huge PITA to get right. Why don't folks duct each cylinder and do individual ram air cooling via scat tube?

The io233 is the only engine I've ever seen that had ducted cylinders. Are there others?

 

[Henning]

I get the impression that baffling is a huge PITA to get right. Why don't folks duct each cylinder and do individual ram air cooling via scat tube?

The io233 is the only engine I've ever seen that had ducted cylinders. Are there others?

I've seen it on just a couple of planes.

 

[Chip Sylverne]

Spike,
A pic of the spacers I installed to push the baffle back from the cyl head, now you can see light between the baffle and the head. Finally, I glued the corners of the baffle seal together with RTV to prevent them opening in flight.