Oxy moron (premature oxygen sensor failure)

I've been going thru O2 sensors. They quit cycling and then after about 40 minutes give some crazy voltage reading. I had a Bosch which lasted several months, but I had been mixing leaded racing gas in which I know is not good for it, so when it went out a couple months ago I figured "Oh well", drove with it going all wierd for a while, and when I started getting codes got a new one. Replaced it with an NTK (NTK=NGK) and after all of two weeks it quit cycling and acted like a throttle position sensor, but read steady between 0.85V and 0.90V on WOT. I figured I just got a bad one so I got another Bosch and right from the start it doesn't cycle on idle or part throttle, varies only with changes in throttle, and reads fine on WOT.

So I checked the O2 sensor plug at the two leads that connect to the two white heater wires. One reads system voltage while the other does not. Should they both be reading system voltage? (If so, why are there two of them?) I'm guessing this is why the sensors aren't cycling. Where does the O2 sensor heater pick up its power? Maybe there is a bad connection at the other end wherever that is.

Sorry for the long post. Any advice is appreciated.
Mike R.

 

[G]

I'm going to say bad connection somewhere on the harness side (brittle wires).

 

I can believe that. There are a lot of rather brittle looking wires in the car which is starting to worry me a little (as in: what's next?).

So in making this suggestion you are indirectly confirming that both white wires need to read at system voltage? (If so, why are there two of them for the heater? Why couldn't one suffice? Could I jumper the two whites together?) Also, if I were to make a new O2 sensor harness, where is the connector to tie in to? Thanks.

Mike R.

 

[iceman69510]

I don't think the heater wires should have any affect on operation once it is warmed up.

I don't know the answer at the moment about voltage on both whites, but will look into it if I get a chance.

 

[Rausch]

i had a very similar problem. after monts of chasing it, and a few O2 sensors, i found that everything worked perfect when i re-ran the ground for the O2. for some reason it was not getting a good ground, and acted very similar to what you are describing....mine would seem fine for a bit, and then just stop functioning correctly. it would read about the same at WOT as you described.

 

That's interesting. So you just spliced an external ground line to the sensor plug recepticle?

 

I seem to recall a thread saying the O2 ground was attached to the ecu ground and needed to be for proper operation; does anybody remember that or know what I'm talking about?

 

[jepherz]

To get the most accurate reading, you probably have to have it wired to the ecu ground. I don't recall any thread about that though.

 

[Rausch]

Quote:
That's interesting. So you just spliced an external ground line to the sensor plug recepticle?

The last time i replaced the damn thing, i re-ran the signal and ground wires, as i thought maybe there was an issue there...nothing changed....i wound up rerouting the wires through the driver's side of the vehicle, and grounded the sensor out near the kicp panel on the driver's side. worked like a champ.

 

[powerplay]

This might help, since it shows the two wires for the heater. The ground is pin 105 and the power comes from the MFI relay. It isn't the easiest schematic to read, but it might help.

 

Quote:
I don't recall any thread about that though.



It wasn't about it, but it was mentioned. Searching for a thread involving 02 sensors and grounds w/out remembering the specifics is almost impossible though. I hope the OP reads this thread and pipes up. /ubbthreads/images//graemlins/laugh.gif

 

^^ Right. I did search around. There was an O2 thread a couple months ago but it gave me more questions than answers. Info on VFAQ was helpful. It confirmed that if the ECU reads two consecutive 20-minute sessions thinking the O2 sensor is bad it is going to start throwing error codes. Interestingly, mine does start doing that after 40 minutes or so. But it does this even with a new sensor which raises suspicions about the integrity of the wiring harness. Then I noticed only one of the heater wires was hot, and started speculating that was the problem. But if it only needs the heater on initial startup then that is a red herring. I like the idea of running a fresh ground wire and will do that first. If that doesn't fix it, I'll run a fresh signal wire too. Meantime, any other insights would be helpful. Thanks.
Mike R.

 

sometimes I have gotten direct fit o2s where the wires where crossed for my application. might want to double check that first.

 

[Rausch]

Quote:
To get the most accurate reading, you probably have to have it wired to the ecu ground.

When i re-ran the wiring the first time, i used the ground cluster by the ECU. There are a bunch of grounds there, mostly for the ECU itself. That, for whatever reason, did not work for me. After running all the wiring new, and through the driver's side, using the ground point at the kick panel there. i got a clean signal, and a good reading. I would speculate that the primary concern for signal quality would be the signal wire itself. The voltage that the signal wire carries it the important portion of what the ECU needs to see, whereas the ground just completes the circuit. A good ground is necessary for accurate operation, and any additional resistance in the circuit could cause an inaccurate reading. I believe that my issue was with a clean ground point, which i don't believe has to be specific to the ECU grounding point, as all the grounds essentially go to the same point anyways. I could be wrong, but i think any good ground would work. In my case, i think i just had a bad connection somewhere, even after the new wiring. I would be willing to wager that if i pulled the ground to the ECU grounding point, it would function the same.

edit: after some additional though, i think that grounding closest to the ECU would essentially provide the cleanest, and shortest, path to close the circuit loop.

 

[bustedsm]

There might be a possibility that you're running your car too rich and killing o2 sensors. My car-car was doing that for a while until I figured out the problem. It'd cycle for about 20-30 minutes run-time and then sit steady at .82 and higher.

 

^^Interesting info's above. Regarding the report of a direct-fit O2 sensor with crossed wires, there are six possible combinations of 2 wires crossed. How did you narrow it down? Just keep swapping wire pairs until it worked right?
Regarding running rich killing the O2 sensor, I had never heard of unleaded gas harming the sensor but the car could be running rich. If I turn up the boost...
Mike R.

 

[bustedsm]

When you run rich the sensor gets clogged up with unburnt fuel and will foul out the sensor. If you have DSMlink or any other sort of fuel tuning device you can test it by adding some fuel. It will eventually just stop oscilating and stick. If you reverse the process though, the sensor will clean itself basically and start it's normal activity.

 

[Rausch]

Quote:
Regarding the report of a direct-fit O2 sensor with crossed wires, there are six possible combinations of 2 wires crossed. How did you narrow it down?

I have been thinking about this on the way home. IIRC on a 4 wire Bosch unit the wires are coded in the following manner:

Black: 1 wire - signal wire
White: 2 wires heater wires (orientation does not matter)
Gray: 1 wire signal ground/ground

Hope that helps.

 

Quote:
Black: 1 wire - signal wire
White: 2 wires heater wires (orientation does not matter)
Gray: 1 wire signal ground/ground



Oh sure I knew the colors. So you got a sensor where the wire was in the wrong part of the plug. I thought you meant they had done something like made the ground wire black and the a heater wire gray, or something. I don;t think this is what's happened here but thanks.

Quote:
When you run rich the sensor gets clogged up with unburnt fuel and will foul out the sensor. If you have DSMlink or any other sort of fuel tuning device you can test it by adding some fuel. It will eventually just stop oscilating and stick. If you reverse the process though, the sensor will clean itself basically and start it's normal activity.




OK I'll look at that this weekend if not too cold outside...But in this case the new sensor was noncycling right from the start. It didn't really have enough run time that it should have gotten fouled. The old one had a white coloration. Is there a color that would indicated a fouled condition?

One other question. VFAQ says the cycling is from the ECU hitting high then low trying to optimize the fuel mix. Could noncycling be a symptom of an ECU problem?
Mike R.

 

[Rausch]

Quote:
Oh sure I knew the colors. So you got a sensor where the wire was in the wrong part of the plug. I thought you meant they had done something like made the ground wire black and the a heater wire gray, or something. I don;t think this is what's happened here but thanks.

I figured you did..I was just throwing them up there, in response to the mixed wires comment. Have you pulled the sensor out? Does it look rough? Really dirty for the time it's been in there? I assume the last one has not been in the car long. That would give you a good indication of if the sensor is getting fouled.

Straight from Bosch:



Mind you that if the ECU is not getting an accurate signal, the O2 sensor will most likely look like the white one in the pic above. That does not mean that the sensor is fouled, per say, especially if it has not been in the car very long. The others I would be concerned about.

I still suggest trying new wiring. I believe the non-cycling would not necessarily indicate a problem with the ECU, but you'd most likely need a known good sensor/wiring to tell for sure anyways.