Failing O2 sensor?

My 99 XJ started having intermittent surging at idle and power loss on acceleration. The surging at idle was rhythmic at about 1 cycle per second, which made me think one of my O2 sensors might be giving a bad signal. And the power loss on acceleration felt a lot like a problem that was fixed with new O2 sensors a couple of years ago.

I unplugged the upstream O2 sensor, and took a few short trips (with the engine already hot). The problem seems to be gone: I don't have any hesitation on acceleration and I haven't had any more surging at idle. Also, normally when my AC compressor cycles my RPMs bump up by about 250 and the Jeep will often overcome the brakes and start moving forward if I'm not really stepping on them. That seems to be completely gone -- with or without AC I'm pegged at 750 idle.

The Jeep just passed CA smog a few weeks ago, so the O2 sensor couldn't have totally failed (?).

Does this sound like an O2 sensor? If not, what else could it be? I also considered that my TPS might have a dead spot, but I replaced it within the last few years with a dealer part. The last time I had my exhaust manifold in my hands, there was a bit of a crack at the collector. Could that, or a bad mating between the exhaust or intake manifolds and the head, be giving the O2 sensor a bad reading?

tetrad said:

..... The last time I had my exhaust manifold in my hands, there was a bit of a crack at the collector.

Click to expand...

That would be the most likely cause of the symptoms. If the crack lets fresh air in, the O2 reads a lean condition. Test the O2 sensor function, its not difficult.

Logically your first step is to fix any known problems, such as the cracked manifold.

Upstream o2 feeds a air/fuel ratio reading to the PCM. Here it is mapped on a scale with other variables to determine richness and leniency of the mixture prior to the cylinders firing... A crack in the manifold, as previously stated by Tim_MN, will let air travel in and the o2 will then see the mixture as being too lean and the system then compensates (not knowing any better) by adding more fuel into the ratio causing the bad idle. Its probably within tolerance values just enough. just about every 4.0 manifold has a crack, some just cause an annoying tick, some don't have any noticeable negative side effects and some let enough air in to cause issues, which i believe is your problem...

another highly unlikely but plausible issue would be a failing MAP sensor, but within tolerance enough not to trigger limp mode... By unplugging your upstream o2 sensor, you caused the PCM to put the jeep into "limp mode" which goes off of a set air fuel ratio map and bypassing the MAP sensors. This would make the jeep appear to run better but definitely not at its best.

Best way to test an O2 sensor, especially a pre-cat O2 sensor would be with a scan-tool system capable of looking at the O2 sensor voltages. I won't go too much into why, but with the cat and O2 sensors warmed up, the pre-cat O2 sensor should be producing a fluctuating voltage. Propane enrichment test might also help a bit, but again, you need to be able to look at the O2 sensor signal.

Thanks all. I'll see if I can get a good reading on my (cheap) analog multimeter.

If that checks out I guess I'll either try to weld up the manifold or get one from a salvage yard -- it's just such a pain to get to either way.

Repairs seldom last. If you are going to spend time/effort to remove the exhaust manifold, you might as well install a new one.

The APN manifold at iNeedParts.com works well.

Well the nicer APN is not CA legal, so that's off my list. There's a salvage shop out here that specializes in XJ and YJ teardowns, so I'll see if they can beat $130

I would attach an analog multimeter to the O2 sensor at your own discretion... I have always been told NOT to hook analog stuff up to electronics and sensors as the low-impedance can screw stuff up. It's also worth noting that O2 sensors typically have low-voltage to begin with (like 1 volt max) which can make it difficult to measure with most multimeters. Like I said though, at your own discretion. You can Google "testing O2 sensor." I know I did and some of the results say you can use a multimeter. If everything's working right, pre-cat O2 sensor should have a fluctuating voltage and post-cat O2 sensor should have a very steady voltage.