Mondial 3.2 decides not to idle

My understanding of the fuel distributor test is that what mattered was the behavior with the pump running—zero rise of the fuel column in the open port. While perhaps a little suspicious, I wasn't inclined to be too worried about the fuel level creeping up over the course of several minutes after everything was shut off. These systems are so dependent on pressure differences and pressure drops across various orifices that I wouldn't put too much stock on anything going on when it's shut off—so long as the accumulator holds some pressure and the injectors don't leak it out!

But the real proof, from my perspective, is the duty cycle of the frequency valve going full rich (as if the oxygen sensor wire were shorted to ground). If I had a mechanical issue making it run rich, the system would be reacting in the opposite direction. And (to Steve's point that my meter can't necessarily be trusted not to make a signal out of noise) the fact that putting it in open-loop mode with the full throttle switch 'fixes' the issue reinforces the conclusion from my frequency valve measurements. (If it were running rich due to a mechanical issue, forcing the system open-loop would only make it worse.)

 

Yeah, I'm not about to try it with a borrowed ECU unless someone can tell me definitively that they've done this before with no ill effects.

Thanks—we started out measuring a coil wire (and almost came to some misleading conclusions) but then switched to a single plug wire when it became apparent that the timing light would at times flash erratically and not correlate to any apparent change in the smooth running of the engine!

Thanks also for this tip—for what it's worth I've discovered that I get a cleaner FV duty cycle measurement from pin 15 on the ECU (which is evidently there specifically for diagnostic purposes) than from the adjacent pin that's actually connected to the FV. (I have yet to actually haul the oscilloscope out to the garage to see what the difference is, but I've read that pin 15 is a clean square wave and the actual FV isn't, presumably due to inductive effects of opening and closing the valve.) I was still pretty suspicious of the protection relay, until I discovered that shorting the full-throttle switch would restore the idle and prevent the engine from dying—if my symptoms were due to the protection relay cutting out, this presumably wouldn't have had any effect. I might still have been tempted to swap it with the identical relay in my Porsche, but...

...the good news is that I drove the car again today, with the oxygen sensor disconnected (i.e. the other open-loop mode, wherein it cycles the FV at the nominal 50% instead of the 60% full-throttle mode), and it ran totally fine, so if the ECU is screwed up, it's not screwed up so badly that it can't still detect an open-circuit lambda circuit! (But I still need to consider the possibility that Amazon sold me a defective oxygen sensor, although it would be odd for that to temporarily cure itself by turning the car off and on again!)

 

Ah, yes, I should have been more clear—I wasn't referring to stopping the car for any amount of time, in this case—just switching it off and on again (or starting it again immediately after it dies) restores normal operation for at least a couple minutes.

I've heard of this (it's been recommended to me before re: the power seat memory module on my Porsche), but always been afraid to try it, for fear that the heat will damage some component which may have been rated for reflow soldering temps when new, but is more fragile now. The heat gun sounds especially imprecise. (I think I've also heard of people reflowing old circuit boards in their kitchen oven???)

 

I should mention that when I referred to pin 15 here (and previously), I meant pin 17--that's what I get for trusting my memory--pin 15 is the one connected to the frequency valve, and 17 is the otherwise-vacant one that evidently carries this undocumented FV duty cycle square wave. (I say undocumented, but one of my Bosch FI books mentions it specifically in regard to several other K-Jet equipped cars; the book just doesn't refer specifically to Ferrari applications.)

 

Did I never post the exciting conclusion of this saga? (Oops.) Replacing the oxygen sensor fixed it!

 

Yes!

Indeed!

 

Yeah, if it had been the associated circuit in the ECU rather than the sensor itself, I would have just run it open loop rather than try to source a new ECU.

But mine doesn't hunt up and down, and never has. (I'm a little surprised how many people put up with that—clean your grounds, everybody (can't repeat that enough)—and if that doesn't cure it, retrofit a 4-wire sensor from a '90s car, and ground the fourth wire in the exact same place that the ECU is grounded!)