9x8

Actually the PO replaced a lot of ECU units (around 3, I think) in a desperate attempt to fix the problem. So I'm fairly sure it's not the ECU problem.

9x8

Update: did some wiring digging, found that bypass not too far from the plug itself. Changed it to the way it should be, the car doesnt start without anything plugged in anymore . What happens now:

1) With IMS relay plugged in, the car starts properly, idles smoothly for 1-2 mins, then goes into 4-cyl mode with green light on.

2) With the bypass in, it works okayish for now, didnt get it to "die" today, but might be due to somewhat low ambient temps, the problem seems to be more noticable with higher ambient temperatures. Idle seems to be a bit rough tho, with some minor rpm fluctuations.

Coils and distributors are actually not too old, replaced 'em around 30k km ago, wonder what's causing triggering of the IMS.

Also, found out that the car previously had cats, but they've most likely been deleted in Germany.
Oh, we found out that WOT switch doesnt work (should activate at about 2/3 of throttle pedal travel - well, it doesnt trigger at all) - but this shouldnt really cause ignition problems? The "closed throttle" switch does work.

Hilton

If the cats were deleted, did you check to make sure the appropriate changes were made to make it a non-cat car?

Non-cat 928's have a potentiometer attached to the LH (pin 23 from memory - would have to go check sheet 10 of the wiring diagram to confirm) which is used to calibrate the MAF signal to a known starting point.

The LH coding plug would also have to be changed to tell the LH the car is a non-cat car too (involves removing a resistor and changing the pin configuration). There's a page with the coding plug configurations in the WSM at the end of the Fuel section in volume 1.

Edit: I bring this up because someone I know locally removed the cats from his '90 GT and had a dying issue until fitting a potentiometer - would die sometimes when changing down gears at low rpm's. He also had issues with idle sometimes - although that may have been ISV-related.

jon928se

If the cats are deleted but the O2 sensor is still there and working and the coding plug configured for cats the car should still run in Cat mode.

Alternatively if the O2 sensor isn't there and the coding plug is configured for Non cats then the potentionmeter needs to be plugged in to the three way plug with two wires attached that hangs around somewhere above the CE panel.

In a Cat car equipped with IMS this will continue to function even if the cats and O2 sensor are removed and the coding plug changed to Non Cat operation. The only way to stop it working is to remove the relay and bridge pins 2,4 and 5.

9x8

Ok, this might be the cause. Coding plug is set to non-cat mode, but there is no potentiometer plugged in. So then 2 possible solutions are:
- either change the coding plug back to cat-mode: "trick" the car into thinking that cats are still there;
- or, fit the potentiometer.
Gonna try the first one, I guess.

worf928

There's nothing wrong with that. There's no reason you cannot run the 02 control loop without cats. Or IOW, you can run lambda control without cats. For a normally-aspirated street car that would seem like the thing to do.

jcorenman

Is there still an O2 sensor? If you change the coding plug to "cat" mode, then there also needs to be an O2 (lambda) sensor present, otherwise you haven't changed much. The O2 sensor may have been removed along with the cats (it lives just upstream of the cats, in the short cross-over).

With the jumper in non-cat mode the LH runs open-loop (no adjustments or adaptation based on the O2-sensor readings), and the pot sets the idle mixture (which effects mixture up to around 2500 RPM). With no pot, the idle mixture is way off (rich I think).

With the jumper in cat mode but no O2 sensor then the LH is still in open-loop mode with no adjustment/adaptation, but in this case it won't be looking for the pot. The idle mixture will still be off but not as bad.

The LH either needs the pot (in non-cat mode) or the O2-sensor (in cat mode) to get the idle mixture right.

So if there is an O2 sensor then change the jumper to "cat" mode, otherwise fit a pot and set it to mid-range for starters (it is a 20-turn potentiometer).

I am not sure that explains why the green LED is coming on with the relay in place-- that relay responds to a difference in temperature between the two banks.
With the idle mixture messed up the car should be running equally crappy on all cyl's...

worf928

You'd think so. However, I've seen a car running poorly enough that the IMS triggered. The car had a bad o2 sensor and a bad MAF. It was running super-rich and the difference between the cylinders was enough to trigger the IMS.

9x8

Set the coding plug to cat-mode (closed loop) and really liked how car runs now. Lambda is indeed present there, was also replaced by PO (lol, why would he do that with car running open loop? ). It didnt eliminate the "dying" problem tho, and I begin to think that its irrelevant to the IMS. I'm actually pretty sure that it's not IMS's fault, and that green light is just a failed thermal sensor apparently, 'cause the car generally runs very smoothly, without any noticable ignition misses.

I'll describe the problem again:
Car starts, warms up, runs very strong, without any problems. But in a few mins (maybe around 20-40) after startup it slowly dies: it feels like fuel supply is cut off (although fuel pump does run, I can hear it), idle RPMs slowly drop, pressing throttle pedal actually makes it even worse. If we immediatelly restart the car, same thing happens. Although if we wait 5-10 mins, it starts again and runs strong for a while, then this happens again. It also doesnt happen much in cold weather, so it really looks like something's overheating. The question is: what?

Tried replacing most of the "53" relays (fuel pump, LH, EZK) in relay board, tried replacing ECUs, tried different MAF too.
Both distributors, coils and O2 sensor have been replaced around 30k km ago.
BTW, the "internal" fuel pump is disconnected, maybe its causing this prob? But I still can hear the main pump running even when the car is "about to die".

worf928

If the green LED in the IMS is lighting up ... are you using the IMS bypass? Or is the IMS plugged in?

9x8

The thing is, the above problem happens with either IMS relay, or the bypass. Also, if it dies with bypass in, and we restart it right after (and put IMS relay), the light doesnt seem to light up at all for those 2-3 secs before car dies again.

As for the cold start with the IMS relay in, the car goes into 4-cyl mode with green light on after about 1-2 mins of idle, which probably indicates faulty sensors. It doesnt die too, can run for a while in 4cyl mode.

Thats why I came to conclusion that IMS is not really the cause for the dying problem.

worf928

Ok. The thing is.... the IMS relay interrupts ground to the fuel injectors when it activates. So, with the LH harness plugged-in to the IMS relay, if one LED is lighting up then two injectors on each bank are not firing because the IMS isn't letting them.

From a cold start it will take a minute or three for the IMS to decide what to do. If the car runs poorly with the IMS bypass and runs even more-poorly with the IMS relay plugged-in (and lighting up) then the IMS is operating correctly. This is - I think - the summation of your description. So, it sounds to me like the IMS is doing its job - more-or-less - and the engine is running poorly.

So, the IMS is, probably, not the problem as you suggest. But, the IMS lighting up when in the circuit, doesn't mean, necessarily, that it is activating due a bad EGT sensor.

At this point I would hook up my Hammer or diagnostic software to the diagnostic port and see what the errors the ECUs are reporting. You could have an ignition problem. Or the fuel/air control could be poor-enough that combustion isn't uniform across the engine (as in: bad 02 sensor, MAF, or wiring to and from).

One trick for the first failure mode is to swap the connections to the ignition amplifiers and see if the IMS's red LED lights up instead of the green LED. Next swap the coils.

61tony

do you get back firing noises in the intake when it is starting to die out? do you notice any other new electrical problems lately?

9x8

I probably didnt describe the problem properly:
the car runs just fine for a while with the bypass. It starts acting up only into well 20-30 mins of running good. And it's totally different from what happens when IMS kicks in.

That's what I would do, if it was that easy. No diagnostic tools for 928's are available locally: we have only around 10-20 sharks running here in the whole city.

It we press the accelerator when it starts to die, then yes, it seems to mis- or backfire. The single electrical problem that I can recall now is the malfunction of throttle position sensor (idle switch works fine, but WOT switch doesnt work, gonna replace it soon'ish).