John Speake

The diagnostic tool will adapt idle, but you are still left with O2 adatation to perform while you drive. This can take up to 50 miles to be fully completed. How many miles since the rebuild ?

Ed Scherer

Over 100, but I think there was one power reset in there at some point.

Once I'm sure that I've figured out this stall problem (at this point, that means that confirming that this relay replacement has fixed things), I'll do another idle adaptation and then go on a victory drive. I've got someone's loaner MAF to return, too.

Ed Scherer

On Saturday, just for the heck of it, as I was sorting through all the used parts resulting from this job (and the previous TB/WP job), I spread it all out on the floor and took a photo. There was some stuff that I threw away (like some fuel hoses) or stored (like the air pump and related plumbing) that isn't in the photo, but this is probably most of it. Most of this stuff is now in the trash, but a few items are good enough to live on in other people's cars (because the item(s) were actually still good, but replaced for cosmetic or upgrading reasons).

Ed Scherer

Drove the car again twice over lunch; about 13 miles to lunch, 13 miles back.

100% reliable. No stalls, no weird idling problems.

This (relay replacement) seems very promising.

Mrmerlin

FWIW its a good idea to replace the engine relays if your tearing into other related systems.
but being methodical can also point you in a failed part,
change each relay one at a time then test

I think the arrows on the filter are for indicating this side up

Ed Scherer

I spoke too soon.

It was back to its old tricks on the way home. A stall or two, and some bucking now and then as I was slowing down for stop lights/signs.



I'll try to be methodical as I continue trying to track this down.


Yeah, apparently so.

And so consistent with other arrows in the vicinity.

Landseer

Ed,
I'd be changing a ground strap and checking /looking hard at the electronics ground.

borland

Please re-read and respond to my questions concerning the voltage gage and gage action when the radiator fans kick in at a stop light.

Also, try disconnecting that aftermarket A/F ratio meter.

Ed Scherer

Thanks, guys. I know I need to make a checklist and be very methodical about going through it. I'll be doing that over the next few days or turning the car over to someone else who can do it (I'm really needing to spend more time on other stuff ).

Barry, IIRC, I think it's easy for me to disconnect the ARM1 pretty close to where the O2 line is tapped. And I might just start logging voltage (my DVM can do that, although I think its fastest sample rate is still pretty slow) at some suitable test point as I drive. I idled the car in the garage after I got home this evening and it stalled abruptly after a few minutes when I wasn't in the car. I went back and looked at the dash (ignition still on, but engine not running) and I think the voltmeter was showing right at 12 V. But I'm not sure I trust it anyway; I'd prefer using an accurate DVM on a good test point (preferably inside the car so I can keep an eye on it). I've been running with my CE panel exposed, so it's easy to get to everything.

I'm obviously trying to detect significant correlations (like to electrical loads like the fans, as you've noted, Barry) and haven't noticed any yet, but it's kind of hard to monitor all this stuff concurrently. It shouldn't be difficult, though, to at least log voltage and see if there's any correlation of low voltage to stalling/bucking episodes. FWIW, I can detect idling problems even before I've waited at a light; it sometimes starts bucking or otherwise not running well as RPMs drop as I slow down to stop (i.e., I don't even need to be sitting there with the engine starting to heat up and the fans coming on).

And there are still a lot of obvious tests that I haven't performed, including temp II and both coils, neither of which has been recently replaced. I did check resistance on the easy-to-get-to-left-side coil this morning, though, and got 0.70 Ω on the primary side and 6.65 kΩ on the secondary side.


FWIW, I cleaned (and DeoxIT-ed) the two grounds at the back of the engine and the coil grounds. They all looked OK, but are even better now. I have not cleaned any grounds inside the car recently.


I've got a pile of 253.B relays coming, so I can replace as many of those as I want.


I've recently (a few months ago and not that many miles ago) replaced the external fuel pump. I've got an in-tank fuel pump available but haven't replaced it yet (that'd actually be on my short list of stuff to do anyway, as my current in-tank pump is about 7 years and about 35,000 miles old, I think).

borland

If your fuel pump was going, you'd have other performance problems too. The other mode of fuel pump failure is the car won't start or run at all.

Forget the DVM logging. If the voltmeter is showing 12V with engine not running and ignition on, then it's accurate enough. What reading is showing on the voltage gage when the engine is at normal operating temperature and at a steady idle? When the car is at rest, you should be able to hear the fans kick on, as they typically do at a traffic light. When the fans kick on, does the voltage drop any, and if so, how much?

If your idle voltage is abnormally low, the fans turning on will make the problem worse. If the voltage is normal, but the fans still dramatically drop the voltage reading, you could have a loose supply wire (loose nut/bolt at the positive battery terminal).

Bill Ball

You said that fuel pressure is maintained despite the stalling, so I don't think I'd be looking at the fuel pump, fuses or relays. Relays are essentially off/on switches. When they fail, the car dies suddenly. lt doesn't just develop a low idle, bucking, stalling, etc. Despite the fuel pressure being steady, you said you see the AFR meter goes full lean when the low idle and stalling happens. If the meter loses the O2 signal, it goes full lean. Where is the meter tapped into the O2 circuit? I wonder if this is an odd O2 circuit failure from an intermittent break/short in the harness.

Ed Scherer

OK, more data. Including one really significant observation.

The mundane stuff first:

• In answer to Borland's questions (and suggestions), I disconnected the ARM1 from the O2 sensor today (it's connected via two quick-connects that have spade terminal receptacles) to get that out of the equation. No difference. I watched the voltmeter on the dash more carefully today. Usually was running around 13.5 V or close to 14 V with engine running and a little over 12 V with ignition on but engine not running. I got a chance to watch the voltmeter when fans cycled on and off; no discernable voltage change.
• I watched the fuel pressure gauge on the right side fuel rail again and got to observe a stall. About 50 PSI Definitely didn't drop at stall time. Increases to about 58 PSI after stall; I assume that's due to fuel pressure regulator and/or dampers losing vacuum.
• I've been wondering for awhile if the idle switch always works correctly. A few times on the highway, I've let off the pedal completely. Digital dash shows 90.0 MPG within a second usually but not always. So maybe, just maybe, the idle switch isn't always triggering. I'm contemplating hooking up an LED to monitor that signal so I can see if I can correlate a possibly-incorrect idle signal to the idle/bucking/stall episodes.

OK, and here's a really interesting finding. This evening after I got home, I just let the car idle in the garage, and sure enough, it stalled after awhile. So... I figured it was about time to see how repeatable this was. So I tried it again, and again, and again, and I timed it. And guess what: every time, the stall was about 75 seconds after the engine was started. What happens after 75 seconds? Closed loop operation? Engine was at operating temperature for all of this. It never made it past that 75 second time period. This seems like a pretty big clue to me.

Maybe it's time to try running with the O2 sensor disconnected again. And I think I'll test the Temp II sensor, too.

I'm off to read up on this.

dprantl

75 seconds when fully warmed up sounds like way too long for the LH brain to begin closed-loop operation. I observe on the ST2 in my car that it is less than 30 seconds. But yes, do the same timed test with the O2 sensor disconnected and see if it dies in the same amount of time.

Dan
'91 928GT S/C 475hp/460lb.ft

Bill Ball

The ignition monitoring system has a delay as well, but it's shorter than 75 seconds. If it trips, the car will usually still run, although the idle and power drops. You would know if that was the problem, so I'm not suggesting it is involved, but just adding it to the list of things that can screw up the idle and general operation.

It sounds like the delay, which is about the O2 sensor warm-up time, points to the O2 sensor as Dan said, and you mentioned earlier your AFR meter goes full lean, that would be another sign, as I noted before, that you have no O2 signal.

Temp II would make the car hard to start as well, but go ahead and test it.

The drop to full lean (90 MPG) on deceleration is RPM dependent - below 1400 RPM it doesn't happen. Above 1400 it should, so it would be worthwhile to check that your TPS is consistently working. If the cable is a smidge too tight or needs some lube or if the switch itself is iffy, it might not seat and trip each time you decelerate. I have found that to cause the idle to be high and surge a bit.

Ed Scherer

Well, since I've got easy access to everything now and the engine was already at operating temperature, I did this moments ago:

• Tried starting engine and letting it run again. Timed it. Stalled after... you guessed it... 75 seconds.
• Disconnected battery ground strap, disconnected O2 sensor, waited a minute or so, and reconnected ground strap.
• Started car. Started timer. Timer now reads 9:50 and engine is still running just fine.

So what do I conclude from this? I'm going to take a damn close look at where the O2 wire is tapped for the ARM1, as well as the harness side of the O2 connector (part of the rubber connector pulls back a little, but I haven't tried pulling it back even more to expose what might be poorly soldered or crimped connections in there). But I'm also guessing that having the O2 sensor in/out of the picture involves a bunch of other stuff, too. BTW, O2 sensor is brand new.

But this seems like a very productive start at a new round of things to inspect.

Advice is welcome.