Fun with oxygen sensors

2000 XJ, 4.0, Federal emissions.

Got a break in the weather between cold days today, so decided to replace the MAP and oxygen sensors on the XJ. The MAP sensor went fine, as did the downstream oxygen sensor. The upstream oxygen sensor, however, has been problematic to say the least.

Long story short: the old, dead sensor is still in the pipe. It's unplugged from the connector going to the XJ's harness; there's no electrical connection between it and the XJ. However, driving the XJ with that dead sensor in there and no wiring in place does not throw a CEL - much the same as when it was plugged in.

The CEL is functional, so it's not a case of a dead bulb. I'd really expect a code to be thrown for at least a failed oxygen sensor heater circuit if not the (lack of) reading a signal from it, but... Nothing.

Any ideas? It seems very strange to me that with the sensor disconnected entirely this would be the case.

That doesn't surprise me at all. I didn't get a CEL for my upstream oxygen sensor when I unplugged it on my '96 either, but I did get the CEL when I disconnected other sensors such as the TPS and ECT.

I have a feeling the P0171 code I've been having problems with lately may have been due to a loose connection at the plug for the upstream oxygen sensor. If so, the oxygen sensor not being connected properly continually threw a P0171 code but never turned on the CEL so I was unaware of the problem until I hooked up the scanner to check for the code.

XJSteven said:

That doesn't surprise me at all. I didn't get a CEL for my upstream oxygen sensor when I unplugged it on my '96 either, but I did get the CEL when I disconnected other sensors such as the TPS and ECT.

Click to expand...

Understood. FWIW, the way I've usually found out that I've got a dead oxygen sensor isn't because the CEL goes on or there are any stored codes, but rather because the XJ fails an emissions test.

I have a feeling the P0171 code I've been having problems with lately may have been due to a loose connection at the plug for the upstream oxygen sensor. If so, the oxygen sensor not being connected properly continually threw a P0171 code but never turned on the CEL so I was unaware of the problem until I hooked up the scanner to check for the code.

Click to expand...

Thing is (and I should've mentioned this earlier), there are no stored codes. It's almost as though there's a simulator in there instead of an actual sensor, but that's not the case.

I'm running low on ideas with this one - there really isn't anything between the connector for the sensor and the PCM other than the relay for the sensor heater, and that at least appears to be working OK. I'm really hoping this isn't the first sign of a PCM going out.

Looking at the wiring diagrams for my 98 I was unsure how the hell the ECU actually detects a failed O2 heater at all, I seem to recall it being relay controlled. I also got no codes... until the CTS reached op temp, at which point the ECU went into closed loop mode, the motor started heaving and jerking, and I got a bunch of O2 sensor codes and misfire codes due to continually guessing wrong, running lean till it started misfiring, then running really rich, etc etc.

Is it possible that you haven't reached full operating temp and are still in open loop? I'd still think you'd get a heater circuit malfunction, but I guess not...

looks like great minds think alike..... and so do Ken and I

on the other hand, fools seldom differ :looney:

I'd have to look at the diagrams again to be sure, and they are sitting at home.

kastein said:

Looking at the wiring diagrams for my 98 I was unsure how the hell the ECU actually detects a failed O2 heater at all, I seem to recall it being relay controlled.

Click to expand...

Yep, there're (IIRC) two relays in the PDC that handle the sensors; they've been swapped with other working ones and same result. Can't figure that one out either.

I also got no codes... until the CTS reached op temp, at which point the ECU went into closed loop mode, the motor started heaving and jerking, and I got a bunch of O2 sensor codes and misfire codes due to continually guessing wrong, running lean till it started misfiring, then running really rich, etc etc.

Click to expand...

Drove for 20 minutes to test, so it was well and truly warmed up and should have been reading the sensor. Never got a single code, misfire, or anything else of the sort - it just ran pretty much like it always does. CTS was replaced about 8 months ago and was showing 195degF-210degF on the temperature gauge :dunno:

shorty said:

Is it possible that you haven't reached full operating temp and are still in open loop? I'd still think you'd get a heater circuit malfunction, but I guess not...

looks like great minds think alike..... and so do Ken and I

Click to expand...

Heh Just to clarify: it was almost 50degF out last night, so in the 20 minutes I was driving it should definitely have got out of open loop - at least, that's what I'm assuming.

Wow, that's freaky. There is no way it should run smoothly with the upstream sensor disconnected.

kastein said:

Wow, that's freaky. There is no way it should run smoothly with the upstream sensor disconnected.

Click to expand...

Yeah, I'm at a loss on this one as well. It's literally as though I never removed it, which has me more than a little puzzled.

casm said:

Yeah, I'm at a loss on this one as well. It's literally as though I never removed it, which has me more than a little puzzled.

Click to expand...

nothing to see here, move along... these are not the O2 sensors you are looking for :conceited

kastein said:

nothing to see here, move along... these are not the O2 sensors you are looking for :conceited

Click to expand...

Hah

I'll pick up a couple of relays while I'm out and see if that makes any difference. The fun part is that I can't really throw the new sensor in there in any sort of meaningful way; the old one is more firmly rusted into the bung than I can get out with what I've got here, so until that's cut out and a new bung welded in I'm kinda screwed.

Feels odd to be actually hoping for a misfire and/or lousy running as a test condition.

Got any ATF and acetone on hand? Try mixing some 50/50 and applying it... after it is visibly dry, apply flame wrench liberally, then quench with WD-40 or pb-blaster (don't use the ATF/acetone, it burns very nicely.) Do that a couple more times then crank on the sensor as hard as you can with a box wrench.

I had a sensor so rusted in that the box wrench was putting purple marks on the palms of my hands and the bung was rippling the exhaust tubing that I unscrewed with a 3/8" ratchet and a deep socket after doing this. Once you get it to budge, it's all over.

X2^^ on every bit of it-- they can be a S.O.B. to get to move, but once it does, it'll zip right out.

Make sure the replacement has a good bit of the anti-seize goop that came with it on the threads when it goes back in-- that stuff's different than regular store bought anti-seize. It has silicone glass beads that keep things from sticking in the extreme heat (is it possible the last one didn't have any?)

kastein said:

Wow, that's freaky. There is no way it should run smoothly with the upstream sensor disconnected.

Click to expand...

My guess is it would be running smooth in open loop mode. Perhaps another sensor (bad MAT?) is keeping it in open loop? Does a scanner show live O2 sensor data? what about open-closed loop?

If their is a short in the wiring harness to the O2 sensor, it may be getting false readings that are keeping it from throwing a code.

kastein said:

Got any ATF and acetone on hand? Try mixing some 50/50 and applying it... after it is visibly dry, apply flame wrench liberally, then quench with WD-40 or pb-blaster (don't use the ATF/acetone, it burns very nicely.) Do that a couple more times then crank on the sensor as hard as you can with a box wrench.

Click to expand...

Interesting - I'd never heard of the acetone/ATF mix before. I'll have to give that a shot at some point, but:

I had a sensor so rusted in that the box wrench was putting purple marks on the palms of my hands and the bung was rippling the exhaust tubing that I unscrewed with a 3/8" ratchet and a deep socket after doing this. Once you get it to budge, it's all over.

Click to expand...

That was last night on the downstream sensor. Upstream, I just managed to completely mangle the hex head on it down to something more like a circle, which is when I finally gave up (and hence why the cut & replace idea is looking likely).

Shorty said:

Make sure the replacement has a good bit of the anti-seize goop that came with it on the threads when it goes back in-- that stuff's different than regular store bought anti-seize. It has silicone glass beads that keep things from sticking in the extreme heat (is it possible the last one didn't have any?)

Click to expand...

Oh, the problem isn't that the threads are seized - it's that last year's snow and salt escapades caused the mating surfaces of the sensor and exhaust to rust together in an amazingly solid fashion. They are literally one piece right now.

casm said:

Interesting - I'd never heard of the acetone/ATF mix before. I'll have to give that a shot at some point, but:



That was last night on the downstream sensor. Upstream, I just managed to completely mangle the hex head on it down to something more like a circle, which is when I finally gave up (and hence why the cut & replace idea is looking likely).



Oh, the problem isn't that the threads are seized - it's that last year's snow and salt escapades caused the mating surfaces of the sensor and exhaust to rust together in an amazingly solid fashion. They are literally one piece right now.

Click to expand...

Mine was probably the original sensor (from the fact that I gained 3-5mpg immediately upon replacing it) and was from '91. It spent its entire life in Alaska and Massachusetts on a total rot-box MJ getting bathed in calcium chloride every winter. I'd wager mine was at least as solidly rusted as yours.

The real problem is that it rounded off... I HATE it when that happens. After that, it's damn near impossible to get out without either welding a nut onto it (tough to do with an O2 sensor) or literally cutting/boring it out.

Ecomike said:

My guess is it would be running smooth in open loop mode. Perhaps another sensor (bad MAT?) is keeping it in open loop? Does a scanner show live O2 sensor data? what about open-closed loop?

Click to expand...

Well, the good news is that I finally got the CEL for it earlier today. No idea why it didn't flag it last night (the XJ had been parked for a good four hours before I test-drove it, so both would've been cold starts), but it was the code for Bank 1, Sensor 1 high voltage; it hit right at the time the coolant was up to temperature per the gauge. Odd that it didn't mention the heater circuit, but my guess is that if the relay is operating properly it may not throw a code for that one. Still going to yank the relays, toss 12V at them, and see if they work or not, though.

BTW, it's running perfectly smoothly - no misfires, stumbling, or hesitation. I'd bet that I'm in run-rich mode right now, though, so it definitely needs to be taken care of ASAP. The IAS was new around the same time as the CTS, incidentally.

If their is a short in the wiring harness to the O2 sensor, it may be getting false readings that are keeping it from throwing a code.

Click to expand...

Makes sense. AFAIK the wiring harness is good, so for now I'm just putting it down to the PCM being slow to respond for some reason. It still hasn't been reset yet, so I'm thinking this may have been a case of new values taking time to replace old values in the PCM before the code was finally thrown.

What's odd to me is that no code has been thrown for the rear sensor - I'd expect it to be complaining about catalytic converter efficiency in run-rich mode. Not going to question that one too closely, though, because if the cat really is keeping the exhaust that clean I'm just thankful that it's one less thing to have to replace.

kastein said:

Mine was probably the original sensor (from the fact that I gained 3-5mpg immediately upon replacing it) and was from '91. It spent its entire life in Alaska and Massachusetts on a total rot-box MJ getting bathed in calcium chloride every winter. I'd wager mine was at least as solidly rusted as yours.

Click to expand...

What's funny is that even though I bought this XJ in Los Angeles, it had come there from Northern Michigan by way of Oklahoma. The first time I replaced the sensors five years or so ago, they were nowhere near as horrible to do as these have been - and these are only on their third Winter out here.

The real problem is that it rounded off... I HATE it when that happens. After that, it's damn near impossible to get out without either welding a nut onto it (tough to do with an O2 sensor) or literally cutting/boring it out.

Click to expand...

Yep. And since I lack the ability to weld, it's left me with little choice other than to take it somewhere and have it cut out. I even took a shot at Dremeling out some of the rust between the bung and sensor mating surfaces, but it pretty quickly became apparent that I was essentially trying to cut through 22mm of (mostly) solid metal.

Well, I could cut it out with the angle grinder - it'd just sound really interesting until I got it to the exhaust shop and had the new bung welded in