High EGT on 1 Cylinder

Hello Andre and Fellow Tuners,

I am encountering a problem with a full race boosted methanol drag engine, and will appreciate any solid advice.

The EGT on Cylinder 3 starts rising much higher than the others at around 7500 rpm & up. The other cylinders will peak at around 1300 F but the problem cylinder will shoot pass 1600 + F if allowed. The higher the rpm the higher the temperature.

Even though you can physically see the runner is a different color due to the heat, I still replaced the EGT probe out of safety precaution with no change.

The engine also has a 4 channel AEM wideband installed, and the Lambda readings are all equal. The cylinder has and additional 5% Fuel trim added with minimal difference. If I advance or retard the Ign Trim on that particular cylinder it makes no difference.

Both a leak down and compression test were carried out and all cylinders are within range (Marginal Differences).

Weak valve springs, or burnt valves are strong possibilities but the engine still runs very clean with no miss or hesitation. Has anybody ever encountered a similar problem? If so, can you share how it was rectified? I will greatly appreciate it.

Thank you

Apparently the answer is no, nobody has encountered this issue. Seems pretty strange especially since you did some due diligence checking into things and didn't find any indication of something wrong.

Did you ever measure the temperatures of the outside of the exhaust runners after a run with something like a thermal camera or laser thermometer? Also did you ever tear down this engine or figure out what was causing this?

Try swapping external components -- for example, swap injectors to determine if the #3 is not delivering as much fuel and running lean. If your ECU allows individual cylinder trim, what happens if you give more fuel to the #3 cylinder?

I find individual EGTs very handy for diagnosis of any running problems -- got a misfire -- the EGTs will tell you instantly which cylinder is the problem.

That's a pretty weird situation you've got yourself there. To be honest, in 20 odd years I haven't come across the exact scenario you're experiencing although it's not too often that we see both EGT and individual cylinder lambda data at the same time. One of the common issues I see with accuracy of the EGT readings is the specific location of the sensor. For reliable results the EGT sensor needs to be fitted exactly the same distance from the exhaust valve on each cylinder and with the same protrusion into the exhaust runner.

If the AFR is consistent across each lambda sensor in the manifold then the other aspect that will influence EGT is the ignition timing although it sounds like you've tried this. Generally advancing the timing will cool the EGT however you obviously need to monitor knock and torque output - There's not much value over advancing the ignition timing past MBT. As a bit of a stretch, the other aspect that could give the results you're seeing would be a camshaft where the lobe timing is inaccurate on that particular cylinder.

I think that I have a modified Subaru SOHC EJ253 with the same problem. The engine includes a longer stroke and greater valve overlap.

At 5,757 RPM, 92.3 MAP, Lambda .746, 21.4° Advance the EGTs in numerical cylinder order were 1422 1491 1398 1404. At 4,120 RPM, 72.0 MAP, Lambda .96, 30.3° Advance EGTs were 1416 1486 1371 1367. Apart from EGT sensors per cylinder there is a lambda sensor for each bank which runs the bank closed loop and always shows fairly closely matching closed loop trims.

A previously installed engine which was close to OEM 2010 Forester using the same intake manifold and same exhaust system gave pretty even EGTs around 1300F at full power so I doubt uneven charge is the cause.

I have swapped injectors and EGT sensors with no transfer of readings. I have also applied fuel and timing trims to cylinder 2 with only minimal change and in the case of a 12% variation in fuel between cylinders 2 & 4 (to enrich 2 by 6%) there appeared to be increased vibration.

The only explanation that I can think of is that the exhaust valve for cylinder 2 is opening sooner than it should.

My questions are should I run the engine without trims and accept the uneven EGTs and can I expect a valve failure or other problem as a result.

Just an update.

The "hot" #2 cylinder was originally showing a simple compression reading of 215 psi compared to cylinders 3 & 4 at 210 and cylinder 1 at 197 psi. A recent test showed a drop to 207 with no change to the others. There were also two knock count increases emanating from #2 but at a time the engine was running at a steady state of around 50% MAP.

Maybe the exhaust valve has been damaged by still burning charges and is sticking. Perhaps the noise is the valve being struck by the piston ?

I am really just guessing at possible cause and would welcome any feedback. I think that maybe replacing the engine would be sensible.

Re: the original posts (although they are zombies), I would have been interested in whether they tried moving the injectors around and if the problem followed it or stayed at the same cylinder? I would suspect a partially blocked filter mesh which was fine with low flow rates, but was restrictive at higher rates, or a blocked nozzle, or two with the remaining ones able to supply sufficient fuel at low load but as the flow increased they couldn't.

Re: the last one, more likely there is a little more ring leakage, although if it was running slightly lean it may have resulted in sliht wear, or oxidation, of the valve and/or seat and very slight leakage. I'm not sure about the 'noise'- if it's referring to the knosk sensor, that won't be affected by mechanical noise, so I wouldn't worry about that. I understand that one cylinder of the Sub' engine is characteristicaly hotter running than others, due to the cooling routing, could it be that? I understand there are fixes for it.

This was so long ago I almost forgot the root of the problem. And yes the injectors were swapped resulting in no change.

The issue was actually a strained crimp on the trigger (ecu) side wire of the secondary injector. It was difficult to pick up as it was housed in a Deutsch round connector with a Raychem heat shrink boot. It also passed the basic Ohm test (static). It was not until we carried out voltage drop tests on the injector harness under dynamic conditions, that we were able to pinpoint the problem.

Dang, that would have been a bugger to find - high fives all round when you did, though, I expect - well done!