1G Cylinder Pressure Waveform interpretation

[barza21]

Hello All,

I need some help interpreting some of the waveforms that I was able to capture using my labscope. The vehicle is a 1991 TSi AWD. The waveforms pertain to in cylinder pressure test while running (at idle only. I have snap throttle events if requested). Based on what I see on the captures, it seems like the exhaust cam is advanced by certain amount of degrees. I believe this since there is a rise in pressure towards the end of the exhaust stroke and the pressure in the cyclinder stops dropping off too early in the expansion stroke). I thought that I had timed this engine correctly, but it seems like I didn't according to this. I am wondering if my interpretation of this is correct. If anybody could take a look at the images and share what they think, it would be really helpful seeing as I would like to know if I have to tear this head apart or if it is just a simple re-timing.

A little background as to why I am here. I have been battling a misfire that does not pop a CEL; however, I can hear it at idle consistently. I am running ECMlink on speed density. I have tested and changed injectors (cleaned and tested PTE 1000cc), coils, PTUs, and wires. I have verified fuel pressures, boost leaks, compression checks, and leak down tests. All of those tests come back good. Cylinder leak check came back at 98% and compression checks came back at around 160-165 all around. I have checked the wiring throughout and it checked good. Grounds for ECU and sensors have been verified clean. The signal coming out of the CAS has been verified and smooth. I have look at ignition patterns on primaries, secondaries, etc., and they look similar across all cylinders. However, they didn't look as smooth as I would like, hence, what led me to dig deeper.

I am learning to interpret these waveforms, so if I am misinterpreting them, please feel free to correct or share knowledge. The images were picked up with a pico 2400a with the maximum amound of captures possible and a SSI pressure transducer was used to collect the pressure data. If this helps me solve this issue, I will probably share the set up here for future reference and troubleshooting.

Thank you in advance,

Oscar

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

 

[We're on Boost]

@steve and @barza21
Thanks for all the explanations about Pin 13, because I had found a diagram for it in the 1990 FSM, but I only half get it, and my FSM copy is the usual made from a scan thing with some things like subscripts barely legible.
You know, Boeing is another place where engineers sometimes work from crappy scans. They started scanning important old things way back when scan technology was still pretty lousy. Too bad. Their earlier system which used photography was actually better. I'm starting to forget now how all that stuff evolved there.
Anyway, from page 14-54 of the 1990 FSM.
They didn't put numbers on the ecu inputs here but the one that comes from S sub 1 must be ecu input 13.
The input labeled "Fuel Pump Control" must be ecu #56, connected to control relay pin #7.

You must be logged in to view this image or video.

 

[steve]

Pin 13 showed up on the 1990 MPI diagrams but was omitted on the 91 FSM and caused me some confusion early on when people were reporting the 42 Fault code. I looked at my 91 FSM and pin 13 wasn't shown but I could see in the ECU code what condition triggered it. Looking at the PCB I could trace it back to the pin.

Here's the page from the 90 Eclipse FSM. The diagram you posted I think is from the 90 DTM not the FSM. The 1G DTM is a great wealth of information that isn't covered in the FSM and everybody should have read it. The 2G DTM isn't quite as informative but again explains some of the changes between the two generations.

There is also a 1991 Stealth Technical Information Manual that has even more background info that can be useful.

You must be logged in to view this image or video.

 

[We're on Boost]

Here's the page from the 90 Eclipse FSM. The diagram you posted I think is from the 90 DTM not the FSM. The 1G DTM is a great wealth of information that isn't covered in the FSM and everybody should have read it.

Ok that helps explain what this pdf is that I have for the 1990. I downloaded it from Tim Zimmer's google drive where he put it in 2018 and he links to it here in post #2. He calls it the Mechanic's Manual. The actual title page says "Technical Information Manual". You are calling it the DTM. What does the "D" stand for?

 

[barza21]

Ok that helps explain what this pdf is that I have for the 1990. I downloaded it from Tim Zimmer's google drive where he put it in 2018 and he links to it here in post #2. He calls it the Mechanic's Manual. The actual title page says "Technical Information Manual". You are calling it the DTM. What does the "D" stand for?

Yeah. Working on this platform and getting a bit deeper of an understanding of certain systems requires looking at different documents at the same time... Sometimes.
Here is a link that is shared on here for all the documentation for these cars

service manuals

i thought i would share my collection of service manuals with everyone. I noticed that the faq already has links to the manuals i posted on scibd but i have way more on mediafire.You must be registered to see element.. ive got 1g, 2g, 3g, 4g, 3000gt, evo8, evoX and some other random stuff...

www.dsmtuners.com

Oscar

 

[We're on Boost]

Here is a link that is shared on here for all the documentation for these cars

Right. I don't think I see what I'm looking for there though. I'm looking for the 1990 FSM, the "regular" FSM which is called "Volume-1 Engine, Chassis & Body". All in one file, not in pieces.
I see it uploaded by Brian just a couple months ago Here. But it's in about 20 files.
@19Eclipse90 do you have this thing all in one file?
Oh wait, those appear to be not searchable.
Never mind Brian!

 

[steve]

You are calling it the DTM. What does the "D" stand for?

D for DSM.

 

[barza21]

I don't think it should unless you have configured ECMLink to prime the fuel system. Normally you would see anything there until the key was in START or after the CAS started pulsing.


It should be clean and under no case be 16v or more. How is your fuel pump electrically connected? Do you see any of his noise at the battery terminals?

Well.... After doing some digging and looking around. Those pulses that we see there are the pulses on the power side of the ignition coil. I was able to check other pins, but everything looks normal.

I checked the input of the O2 into the ECU and it follows exactly what link is plotting/displaying. The airflow sensor and CAS sensor have spikes around where the popping happens, but I don't know if that is from the hiccup of the engine and the alternator reacting or if it is an actual issue with power. I looked at the ground and it shows about 300mV on sensor grounds to frame ground. However, when I check it with a multimeter, it only shows 5mV. I don't know if there is some type of filtering I am supposed to apply to the readings.

Another piece of data that I gathered is that the popping only happens when the O2 signal starts cycling. It doesn't matter if it is simulated or not. It doesn't matter if the engine is cold or warm. If the O2 starts cycling when the engine is 100*F, it starts popping. If it starts cycling at 196*F, it starts popping then.

I have some oil pooling on the exhaust manifold #1 runner (around the top). It seems to be seeping through the bolt. I think this is an indication of a valve guide seal going out. Can a valve guide seal leaking realistically cause a misfire like this? This needs to be corrected, so I will potentially be replacing those seals sometime this summer.

Any additional ideas are appreciated.

Oscar

 

[danl]

It sounds like a lean misfire. If you lock it into OPEN loop in dsmlink, does the problem persist? If so, what if you add 10% fuel with the fuel sliders? I would not trust what the wideband is telling you, it could be the source of your problem.

 

[barza21]

@danl Locking it in open loop doesn't do anything once it has started popping. I don't know why.

However, I did mess around with the fuel settings. I kept richening up the global until I overcame the adjustments made by the ECU with the trims (-29.5 CFT). Once I was able to overcome that (the global was at only -9% and I am running 1000cc PTEs), the misfires stopped. The O2 sensor was pegged high and the AFRatioEst was also pegged high. However, the wideband was reporting back richer readings. This is looking like a an exhaust leak pre O2 sensor or am I thinking about something the wrong way?

I will try to post a log later with all the messing around I did. I will try to pressurize the exhaust and see if I can find a leak somewhere.

Oscar

 

[We're on Boost]

I kept richening up the global until I overcame the adjustments made by the ECU with the trims (-29.5 CFT). Once I was able to overcome that (the global was at only -9%

I can't follow this.
Could you lay it out a little more clearly, like:
When my Global Fuel was -xx% my Combined Fuel Trims were -29.5%.
When my Global Fuel was -9% my Combined Fuel Trims were xx%.
And which one of those was before, and which one was after you got the misfires to stop.

 

[danl]

It sounds as if you force it rich enough, the misfire stops? If you get it to misfire, what happens if you deactivate each cylinder one at a time in dsmlink? Can you identify which cylinder is misfiring? If you identify a cylinder, what happens if you swap that cylinders injector to another cylinder, does the misfire follow the injector?

 

[barza21]

I can't follow this.
Could you lay it out a little more clearly, like:
When my Global Fuel was -xx% my Combined Fuel Trims were -29.5%.
When my Global Fuel was -9% my Combined Fuel Trims were xx%.
And which one of those was before, and which one was after you got the misfires to stop.

My bad. So, I was trying to force the system to run rich. I started messing with my global setting. For my size injector, the global was set to -48% prior to any changes. I started moving towards 0% to give it more fuel. It wasn't till I reached -9% global fuel setting that the AFRs being displayed by the wideband got below the high 14s to low 15s. Once the AFRs went below 14s it started clearing up. Eventually the widbend started reporting 13s and it was clear at idle (no more popping). Once I reved up to 2k it would also remain clear.

 

[barza21]

It sounds as if you force it rich enough, the misfire stops? If you get it to misfire, what happens if you deactivate each cylinder one at a time in dsmlink? Can you identify which cylinder is misfiring? If you identify a cylinder, what happens if you swap that cylinders injector to another cylinder, does the misfire follow the injector?

I have tried deactiving a cylinder at the time and reving it up to 2k. It sounds choppy because of it is running on 3 cylinders, however, it sounds smooth. It doesn't break up. Based on what I have seen, there doesn't seem to be a change in choppiness regardless of what cylinder I have selected for deactivation.

So... I had leak tested the exhaust in the past. I had smoke tested it and it had checked fine.... There was no soot, and there is still no soot. However, I went with the pressure approach this time around. I blew air up the tail pipe and sprayed soapy water all over the exhaust. Lo and behold: leaks. There is a leak on the upper corner of the #4 cylinder exhaust manifold/head gasket. There is also a leak, a big one, at the junction between the manifold and turbo. Finally, there is one between the O2 housing and the turbine housing. I am hoping that this is the reason why this is happening....

I am sure smoke tests work, but they have not worked for me anytime I have needed them. This is a great jump forward.

Oscar

 

[We're on Boost]

My bad. So, I was trying to force the system to run rich. I started messing with my global setting. For my size injector, the global was set to -48% prior to any changes. I started moving towards 0% to give it more fuel. It wasn't till I reached -9% global fuel setting that the AFRs being displayed by the wideband got below the high 14s to low 15s. Once the AFRs went below 14s it started clearing up. Eventually the widbend started reporting 13s and it was clear at idle (no more popping). Once I reved up to 2k it would also remain clear.

Let's see if I can answer my own question now, that is in post 60:

When my Global Fuel was -48% my Combined Fuel Trims were -29.5%. (Right?)

Then when I changed my Global Fuel to -9% my Combined Fuel Trims became ... what?

 

[barza21]

Let's see if I can answer my own question now, that is in post 60:

When my Global Fuel was -48% my Combined Fuel Trims were -29.5%. (Right?)

Then when I changed my Global Fuel to -9% my Combined Fuel Trims became ...

Sorry. With my global at -48%, the CFT were at 0 ±5%. It would bounce around consistently. Once I switched the global to -9%, the CFT went to -29.5% (which I believe is the max since it wouldn't go past that). At that -9%, the wideband actually started richening up and reporting readings below 14.0 AFRs.

Is it the definite problem? I don't know, . I do know it is a big problem to have a leak pre-O2 sensor and I hope the repairs puts an end to this pain, .

 

[We're on Boost]

Sorry. With my global at -48%, the CFT were at 0 ±5%. It would bounce around consistently. Once I switched the global to -9%, the CFT went to -29.5% (which I believe is the max since it wouldn't go past that). At that -9%, the wideband actually started richening up and reporting readings below 14.0 AFRs.

Is it the definite problem? I don't know, . I do know it is a big problem to have a leak pre-O2 sensor and I hope the repairs puts an end to this pain, .

Ok that makes sense. Thanks!

It would still be a good idea to post a log of it running with Global Fuel at -9% so we have it for any other questions.

For a log of it when you had Global Fuel at -48%, well, we have your log from May 28 which was 2 weeks ago now.
Maybe that log is good enough for showing how it was at -48% but if there is a later one that is more representative somehow, then post that too.

I think these lean spikes are an interesting problem. They are a little weird. There must be something good to know there.

 

[barza21]

Ok that makes sense. Thanks!

It would still be a good idea to post a log of it running with Global Fuel at -9% so we have it for any other questions.

For a log of it when you had Global Fuel at -48%, well, we have your log from May 28 which was 2 weeks ago now.
Maybe that log is good enough for showing how it was at -48% but if there is a later one that is more representative somehow, then post that too.

I think these lean spikes are an interesting problem. They are a little weird. There must be something good to know there.

Yeah. I agree. I am not sure if those lean spikes are a response to a condition set by the O2 signal that is not accurate or if it is something else. I am going to replace the stem seal along with the gaskets. We shall see what that does. I found a definite problem that requires addressing, though.

 

[We're on Boost]

I guess we don't know the end of the story yet for your lean spikes, but I have sort of figured out the lean spikes that I was having. My lean spikes were different, probably a different cause, but there must be something to learn from it, so here it is.

I was having lean spikes in the AFRest at idle and light cruise, so I looked back through my logs to see when they started. They started in December of 2021 when I replaced my omnipower 4 bar MAP sensor with an AEM 5 bar stainless MAP sensor.

So a few days ago I decided to replace the MAP sensor again. I replaced the AEM 5 bar with an AEM 3.5 bar, stainless. Thinking it was a long shot, because really why would the car be so sensitive to the MAP sensor when running in closed loop?

Anyway, it worked. Now with the AEM 3.5 bar, there are no lean spikes at idle or light cruise anymore.

I don't notice any difference in the way the car runs or idles. And I didn't notice any change back in 2021 either when I went from the 4 bar to the 5 bar.

The only thing I can think of is that the resolution, psi per volt, is so different. AEM 3.5 bar and Omni 4 bar have fine resolution. AEM 5 bar is more coarse. That's maybe not very helpful for your case since you are already using an omnipower 4 bar.

When at idle or steady light throttle, my MAP sensor readings will occasionally bump down or up for one sample here or there. I noticed in my logs that the lean spike would usually (about 90% of the time) happen exactly at the low point of a bump down in MAP. Now it's normal for a 1g to have a lean spike when you take your foot off the gas during cruise. So my thought was that if the MAP sensor resolution is coarse enough, those random little bump down events at steady throttle could be seen by the ecu as me taking my foot off the gas a little bit. And maybe a finer resolution MAP, when it bumps down, maybe it's not enough of a pressure change to throw a lean spike. It's going to be a .02 volt change regardless because that's our ecu input resolution, but that .02 volts is only 0.5 inches of Hg with an AEM 3.5 or an Omni 4 bar, and it's 0.7 inches of Hg with an AEM 5 bar. Maybe the bigger pressure change fools the ecu.

Now, with the 3.5 bar, I still have these little bumps down in my MAP logs, but there aren't lean spikes happening there.

So maybe this is what it is with my car. I don't know enough about what's going on in the ecu to know if this is even plausible. And I think yours is different. For one thing, I could make my idle spikes stop by forcing open loop. You have said yours don't stop when you try that. Also, my spikes were very frequent, many per minute at idle. Your spikes are very few at idle. My car runs a little rowdy all the time, pops and crackles on decel for example, which I figure is due to so many mods, so this type of thing just seems normal to me when I'm in the car. I might notice it better with an audio recording of it idling. I do have a couple of recordings of it with the AEM 5 bar. Probably will make a recording at some point for the AEM 3.5 bar.


Good places in my logs to see my idle spikes, or lack of spikes:
After 1793 in the June 19 log (last drive with the 5 bar)
After 1133 in the June 29 log (2nd drive with the 3.5 bar)

And I bet there are some 1g guys using the AEM 5 bar who don't get the same spikes I was getting, so there's probably more to it!

 

[barza21]

Update... @steve @danl @We're on Boost
It has been a while since I have updated anything, but it is summer up here and the car kinda fell on the back burner for some time. I went ahead and replaced all the gaskets on my exhaust system (after I found those leaks pre O2/wideband sensor). I have verified there is no leaks on the exhaust between the head and the wideband now. Also, I verified and there is no leak from the any of the valve guide seals (it was a concern of mine). I verified the mechanical timing again and it checked out fine; if I want to be critical, it might be off by 1/4 of a tooth at the crank... Maybe... Additionally, I replaced the OMNI MAP sensor with an AEM 3.5 bar one to see if this helped with the AFRatioEst spikes (this was done based on what @We're on Boost had said he experienced and the history behind OMNI sensors) . With that being said, the AFRatioEst spikes are pretty much gone for the most part post MAP replacement; that's a win! However, the exhaust popping is still there. I got around to messing with the VE table and the global settings to see if I could adjust the mixture that is being read by the wideband. No adjustment that I made would let the wideband go richer than 15.2-15.5afr at idle. The AFRatioEst would change from 12.xx to 15.xx afr as I messed around with the VE table and the deadtime adjustments. However, the ECU, while in closed loop, would consistently try to adjust it back to whatever the O2 sensor thinks is stoich. Therefore, the fuel trims would change accordingly to keep it at what it thinks to be stoich and giving me the reading of 15.2-15.5afr at the wideband reading during idle. The whole time I can hear a pop here and there from the exhaust. The combined fuel trims hover around 0% +- 3%.
So, as I am learning to troubleshoot with link, I discovered that if I lock it in open loop and adjust the cell being used in the VE table at idle, I can manually change the mixture without the automatic adjustment to the mixture from the ECU. I didn't know until I tried it last night. That is my bad. So, once I discovered that, I proceeded to richen up the cell, therefore richening up the AFRatioEst and this time around the wideband afrs followed! I was able to adjust the wideband afrs to 14.7. Once I was able to get it to that, the popping stopped. Anything richer than that, there is no popping; just unburnt fuel smell. Furthermore, if I leaned it out to 14.9-15.5, the popping from the exhaust would resume.
My question is: why is this happening in closed loop? Why is my O2 sensor thinnking it is running richer than what it is and therefore causing the ECU to remove fuel? My idle vacuum is 19/20 in/hg, which seems pretty normal. Several different injectors have not changed this. My fuel pressure is set to 43.5psi on a Fuelab AFPR. I am scratching my head on this. I really am. Any ideas, once again, would be greatly appreciated.

Oscar

 

[danl]

It’s because o2 sensors and the controller are not perfect. You could also have a cylinder to cylinder imbalance meaning that 3 cylinders are 14:1 afr and the 4th is 17:1 and misfires.

If the problem is just at idle, change the target afr in link. I’m not sure what it’s called, but you can tell link to target a richer afr at idle. This helps on cars with cams and is common enough that link added the feature.

You will want to do a plug read after a wot pull to make sure your fueling to all cylinders is even during high load conditions. If it’s not and you target 12:1, you could have one cylinder at 15:1 actual while the others are 11:1.

 

[We're on Boost]

I finally got my audio clips done for A-B comparing my idle sound with and without lean spikes in the AFRatioEst, so, might as well post them! I'm not so sure I can even tell any difference. It might idle a teeny bit cleaner without the spikes. Either way there are lots of little bumbles going on with how it runs at very light load like this. The higher the LoadFactor goes, the smoother it gets.
I put more words in the video description fields. I've posted the full logs below too.

Lean spikes at idle with AEM 5 bar MAP sensor, speed density, ECMlink, 2025-06-01

No lean spikes at idle with AEM 3.5 bar MAP sensor, speed density, ECMlink, 2025-08-10


For perspective, it's fun to remember the old road/race sports cars with carburetors and points-coil-distributor type ignition systems. Like this 1957 Maserati! I remember these 50's Masers and the Birdcage Maserati of 1959-61 which was also a 4 cylinder. Most of us never knew what our actual AFRs were in those days, but we knew that engines ran better when slightly richer than stoic.