1G Unending Misfire

[barza21]

Hello All,

I am here once again seeking help with my issue I can't figure out. This is a follow up to last year's post regarding the same issue. I had time to take a look at this car again this year and nothing has really changed. The issue at hand is a misfire that happens pretty frequently at idle and when reved up while in neutral. There are no codes for misfires, but I can hear and feel them when I hold the throttle open and rev it to 1500-2000rpm. A WOT pull doesn't reveal any misfires that I can feel or hear.

While I'm driving and under load, the car drives fine. I don't feel any bucking or misfire. Once again, no codes for anything. I did notice this year that if I am going downhill and engine braking, that if I give it a little bit of gas and get it to about 19in/hg on the gauge, it bucks and doesn't like it. That same 19in/hg is at what I idle at and what it sits at when I rev and hold. So, I suspect this is related to my main problem.

Other than that, everything is the same. I am on ECMLink running speed density. I've checked for leaks and recently rebuild the throttle body thinking that maybe there was something in there I couldn't see and causing a problem. It didn't change anything. Additionally, I also rechecked compression this year. I am sitting at about 160 across the board. A leak down test reveals 98% seal across all cylinders. I've checked the base ignition and mechanical timing as well. I've checked the mechanical timing in all the ways possible listed on this site. It looks fine every time. Plugs have been replaced, ignition cables swapped around, PTUs swapped, I've experimented with different fuel pressures and adjusted in ECMLink. Nothing has helped. I bought a vacuum gauge to see more accurately what the engine is going, but nothing. The vacuum gauge moves around a bit because of the misfire, but sits around 19in/hg. I've discovered that even when making it run in open loop, it'll still hiccup, but not as bad as it does when in closed loop.

To me it seems like a cylinder is running lean. I can't figure out which one. I've pulled cables and turned off injectors (in different combinations), but the moment that I do that, it stops misfiring. It just runs rough because it is running on 3 cylinders. The injectors installed have been cleaned and flow tested.

I've looked at signals with an O-scope and they seem normal, but I am not 100% sure what normal is in certain instances. The cylinders look similar to each other. The signals from the different sensors seem to be getting to the ECU and verify against what I see on the scope.

At this point I wouldn't be surprised that I have overlooked something incredibly simple and have totally dismissed it.

I am seeking assistance because I am at the limit with it. I really am. I am getting to the point where I am not enjoying this car since I can't tune it because of the issue it has. It is incredibly frustrating. I like working on the thing, but not when it is am issue that has repeatedly been kicking my behind.

I am seeking some much needed assistance from anybody. Almost any suggestion will be tried, .

Thank you!

P.S. Here is the link to last year's post.

1G - Cylinder Pressure Waveform interpretation

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...

www.dsmtuners.com

 

[barza21]

You are correct. I have disconnected the connector from the PTU and probed pin 3. The other lead is hooked to ground. The ground used is on the firewall. The moment I turn the ignition key all the way to on (engine not running) the pulse starts showing up. The spike is not large, about 100mV. It disappears once the key goes to off and the MPI relay unlatches. I can also make it go away when I disconnect the middle connector from the ECU. Maybe it's nothing. Maybe it's something. I checked the sensor ground and the signal is there too. It is also present on the ECU ground.

The moment I turn the ignition off or disconnect that connector, it goes away. I saw that same pulse on the sense wire of the TPS when the car was running. That's the only reason why I kept digging. Keep in mind I have probably gone down the deep end, but I don't have experience with automotive signals. I am just going based on what I've found or read.

 

[luv2rallye]

Some ideas to check/consider:
- Not totally familiar if the 1g has it but on a 2g when you turn the key to ON, it goes through an IAC recalibrate cycle which could cause 100mv pulses on the ground while it's going on (since the IAC has coils in it).

- Is your TB Idle switch set correctly? Read "Idle Switch" and its adjustment in here: https://www.sixsigmatuning.com/dsm-idle-control

 

[steve]

Addressing the potential floating ground scenario that could be taking place, where is the ground wire for the PTU? I thought it was the one at the firewall on the driver side; near the clutch's reservoir. Am I correct?

The PTU Ground is #1. That's the central firewall like you describe.

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

It disappears once the key goes to off and the MPI relay unlatches. I can also make it go away when I disconnect the middle connector from the ECU.

C-60, the middle connector on the ECU, among other things has the signals for the MPI relay. So disconnecting it will turn off the ECU power by deactivating the MPI relay.

 

[luv2rallye]

Does the 1g PTU have a better ground on the back of its case like the 2g one has? If so, you should have the PTU mounted in it's bracket to the head for 1) a heat sink and 2) because the bracket gives a much better high current ground than does the ground wire.

 

[steve]

I'm not sure any of the PTU versions use the metal heat spreader as an electrical ground.

I've been using heatsink grease on the backs to the metal mount and other than better heat transfer it hasn't change the behaviour which it would if it was intended to be a ground path.

Given the amount of corrosion I usually find between the two it wouldn't be very conductive either.

 

[We're on Boost]

Want to show you a shot of a log that I made in 2009.
At this time my ecu was stock 1990 but it did have fresh caps in it.
I was logging it with an ECU+ which has good input channels for measuring voltage. These channels are purpose built to measure voltage (not like our repurposed ones with ECMlink) and they have an input impedance of about 1 million ohms.
I was only actually using a few of these voltage measuring channels. The rest of them had their wires taped off, not connected to anything.
The not-connected ones are labeled:
Analog Input 0
Analog Input 1
Analog Input 2
O2 Voltage (Rear)

Look how noisy they are!
Lots of little spikes all over the place.
Magnitude - about 100 mv.

I never thought much of it at the time. It didn't seem to hurt anything. Never really looked at it until yesterday, because we are talking about stuff that seems related. So yesterday I scaled these windows to show from 0 to 1 volt, to make the small voltages stand out clearly.

At the time in 2009 I just thought the disconnected wires were picking up noise from the atmosphere.
But maybe the noise was in the ground? The ECU+ does have one wire that connects to our sensor ground, and another wire for power ground, and that's how it was connected. Plus all the other connections that are shown in the installation diagram which I'll also post.
Don't know if this helps anything but I thought I'd throw it onto the pile!
BTW I'm still using that same ecu but now of course it has been converted by ECMtuning.
And I'm still using the ECU+ but I'm only logging oil pressure with it.
If I ever get around to putting my stupid AEM X-series wideband on the car, I will log that with the ECU+ also, because those things are known to be not very compatible with our ecu.

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

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

 

[barza21]

Some ideas to check/consider:
- Not totally familiar if the 1g has it but on a 2g when you turn the key to ON, it goes through an IAC recalibrate cycle which could cause 100mv pulses on the ground while it's going on (since the IAC has coils in it).

- Is your TB Idle switch set correctly? Read "Idle Switch" and its adjustment in here: https://www.sixsigmatuning.com/dsm-idle-control

It could be the case. I will check the driver side of the ISC signal and see what it's doing while just with the key on. That article that you linked has a lot of good info. regarding the idle switch, I adjusted it a few times on two different throttles. I've also simulated the idle switch via link and it hasn't affected the misfire.