944crazy

Do this again, and when the idle smooths out, plug the o2 back in. If that causes it to start stumbling again, then you know for sure it's the o2 sensor, in which case, I'd just replace it with a brand new one.

jvandyke

The only thing (I know of) that's '83 is the engine, harness and sensors seem to be '87 924s (which is some cases puts it the same as an '83 944 I guess).
Thing is, this didn't happen last fall and nothing changed besides removing transaxle and bell housing for clutch work which disturbs Ref and speed sensors, ground lugs and O2 sensor. All that looks okay.
Yes, unplug sensor, car smooths out, plug it back it, it gets pissed. I think, just for fun I'll plug it back in once the exhaust (and thus sensor) is good and hot and it "should" be just fine (because this is how it performs on the street with the sensor always plugged in; loss of power 90 seconds into running, power returns about 30 seconds late)r which tells me the sensor is working but the heater circuit to preheat the sensor is not, so I'll check it for voltage but I can't find definitively WHEN the voltage is sent by the DME, I assume right after start, maybe it's continual, maybe it cuts it off, some delay the heater for a while, so does anyone know?
Since two sensors behave the exact same way I'm not buying a third unless I have some assurance that's what it wants. I think the heater circuit is grumpy. Is it fused? We had a couple blown fuses yesterday too. Research says it's the same circuit as the fuel pump so must not be the one or the car wouldn't be running merrily down the road.

Or it could be the DME looks at the sensor after 90 seconds, get's weird info, tries to adjust mixture, car runs horrid, then after 30 seconds it gives up and goes back to open loop?

V2Rocket

DME doesn't give up on closed loop once its up to the closed loop temperature, with the O2 sensor connected.

924S is kind of a mishmash car...early body and most wiring but late suspension, DME etc.
On the early model 944 the O2 sensor heater circuit is two wires (power and ground) fed IIRC from the fuel pump power circuit, after the DME relay. Either that, or its off the ignition switch black wire.

I imagine it's similarly laid out on the later cars, but might want to check that your heater wires actually have power or not.

jvandyke

I'll check for power on the heater wires on the plug, assuming it's supplied by DME right from start (most likely?). I can cut out my solder joint too although I'm doubtful that's got anything to do with it as they were soldered before and soldered worse now () which should make it work better since supposedly it's air in the wire strands that the sensor needs and solder stops that, some say it's the change in resistance from solder that makes the sensor give erroneous information but the issue is shrouded in mystery. My guess is heater circuit. '87 924s is three wire sensor BTW.

jvandyke

confirmed power on two whites (heater wire) with key on run, cut out soldered wires and crimped/shrink wrapped instead, only one test cycle but it DIDN'T do it, will test further tomorrow, it would be crazy if it were soldered joints causing this but why would it "cure" itself when hot? hmmmm

jvandyke

nope, still doing it, nuts.

jvandyke

ordered dirt cheap O2 sensor since I'm not too thrilled to buy $90 parts I'm not convinced I need. Sure seems like the heater isn't working but it has voltage. Weird. I'll try the new sensor if it also acts the same way I'll pretty much know it's not the sensor, if the new one works, better or differently then I'll pony up for the correct on. I'll hold the thing in my hand and see if it heats up or not, should be able to do that I believe.

jvandyke

took an 8 mile run yesterday, about 2 miles in we pulled over and plugged in the O2 sensor, idle changed a bit, took off, car runs much better with sensor in operation, drove another 4, stopped (to buy more fuses) and left again, ran 6 miles back home (hammered on her a bit, went like stink-which is good)

​​​​​​​
today;
$9 sensor arrived, crimped it in, two runs, no recurrence, will continue to monitor but it sure seems like both old sensor behaved the same way, weird, no? Assume it was just sensor the whole time unless I come back here to whine.