• Welcome to the new NAXJA Forum! If your password does not work, please use "Forgot your password?" link on the log-in page. Please feel free to reach out to [email protected] if we can provide any assistance.

Intermittent rough run, any clues?

D

dutchjeep

NAXJA Forum User
Location
Half Moon Bay, CA
My XJ has a weird habit. It will sometimes run very rough, usually smoothing out when warmed up (but sometimes this takes much longer than other times). What often helps is stopping the car, restart, and....smooth run! I don't know what the sound is, I'm not familiar with piston slap so could be. It almost sounds as if the timing is off just a little (had that once on a diesel engine and is somewhat similar, not identical), the engine is overall much more noisy (valves?) during the rough run, and I really dislike the sound of it. I am really clueless what this could be.That restarting helps suggests a computer/sensor controlled issue maybe? Anyone experience with this? These post-renix XJ's don't have a knock sensor right? I already replaced the cracked exhaust manifold and recently did coil/plugs/wires/cap.

Anything?
 
It might help if you tell us if it always fixes itself when warm, or just sometimes, how warm, does restarting always help, etc. How often is sometimes in other words?

If it reliably goes away instantly with a restart, it isn't purely mechanical. Probably electrical sensor related. The first thing that comes to mind is the O2 sensor, especially since you probably handled it in the manifold work. It comes alive after it warms up, and has a heater inside to quicken that. Do you get an emissions light? If it is starting to signal, but is flakey (like the heater is out in it), it would cause bad mixture and possibly a whole bunch of odd noises too like knock or misses.

As it heats up from exhaust it starts to work properly... If it's cycling early, it's probably grossly inaccurate, and the ECM thinks it's live and using the data.
 
Sorry for the inaccuracy there. It always smoothens out. When the truck is warm it runs great, no exceptions. It will always, in the warm up cycle, run rough for a little while, also no exceptions. How long it runs rough varies and I have been unable to find any consistent patterns there.

To give an example. It will usually run ok when just started, but will usually run rough when ran for about 2-3 minutes, this will last another 2 or 3 minutes and then it smoothes out, and run fine. This is in the warm up cycle and I always drive th truck real gentle when it's warming up. Sometimes it will even run rough directly when started, which generally smoothes out at some point also (usually the rough run picks up a little later though). If I find the smoothing out takes too long I found that if I just stop and restart (when the engine is a little warmer already but still running rough) the engine runs smooth again.

I have had no emission codes. Just passed smog and it did great (can get the numbers of the rapport if that's helpful). If O2 sensor could be a culprit I will replace it asap. What do you think?
 
Well, the O2 sensor is dead when cold, but has a heater so it warms up enough to output in a couple of minutes. One other time the heater is needed is during slower running, when it cools below it's proper temp. The ECM will give a code if it is consistently out of bounds, but if it comes and goes and works fine for periods, it may not trigger a light. Depending on what you did before the test, it probably was hot enough to carry you, which means it may be the internal heater on it, which can die too. It might also just be giving bad numbers for these cooler times - the ECM wouldn't know, it only knows it's getting readings.

It all seems to follow the profile of activity of the O2 sensor. The only odd part is the warm restart, but that does restart the ECM's 'expectations' too, so there are a lot of things that could influence that. Check the connection to it. I'd backprobe the wires while it was running fine, the do the same near the ECM and watch those while it runs rough. Does it change pattern? You can also just compare running fine patterns with rough patterns near the sensor. Given you were around there repairing, it might be a loose connector or wire.

Also try getting it warm enough to run smooth. Let it idle a while - does it run rough then right then rough? Then dribble water on the exhaust near, but not on, the sensor. As it cools, and doesn't steam up anymore, flood it with water more to really cool around the sensor. Can you make it run rough on demand that way? If so, it's there.
 
If you passed SMOG your O2 sensor is working but was that before or after your manifold mork? It's also not a component of cold-start condition so - STEP AWAY FROM THE O2 SENSOR, for now.

My guess is a temperature sensor out of whack, air in fuel system? Or, on my '89 the injector connectors would get corroded and only miss when cold. Try wiggling each injector wiring/connector when it does this.

Noise? Stethoscope the damn thing and tell us where it's coming from. But if the engine is running like shit and rocking around, things are gonna be noisy under there.

Never compare anything about a diesel to gasoline. About the only thing in common is that they both have a crankshaft and rotate. For example when a diesel is "blackstacking" (blowing thick black smoke) it's burning-up inside. When a gas engine blows black smoke, it's running rich & cold.
 
It's not a part of cold operation because a normally functioning one doesn't out[ut anything when cold. As soon as it starts outputing, it goes closed loop. If it is outputing cold, it will be inherently inaccurate, yet be used. Passing the smog test when it was running right just means the sensor was hot enough to be accurate at that moment - it is NOT a clean bill of health on the sensor in general. This is a problem only at certain times and conditons.

The tests are easy and simple and the first place to look. Even if you disconnect the temp sensor, it won't cause these effects.
 
Thanks guys.

So instead of trying to cool down the O2 sensor with water, can I just unplug it see what it does? That would cause the truck to run too rich when warm right (it will not go into closed loop), but shouldn't affect the cold run? Wouldn't that be a good test to see if that affects the rough idling during warm up? Just want to make sure that won't harm the truck.

I also though about the temp sensor (this is the one on the t-stat house right, the one in the back close to the firewall goes to the gauge, right?). So can I disconnect that one to test what happens? What happens if the computer doesn't get temp info? Or...what temp info does it get if I disconnect?

So where is the ECM physically located?
 
XJXJ said:
If you passed SMOG your O2 sensor is working but was that before or after your manifold mork? It's also not a component of cold-start condition so - STEP AWAY FROM THE O2 SENSOR, for now.

My guess is a temperature sensor out of whack, air in fuel system? Or, on my '89 the injector connectors would get corroded and only miss when cold. Try wiggling each injector wiring/connector when it does this.

Noise? Stethoscope the damn thing and tell us where it's coming from. But if the engine is running like shit and rocking around, things are gonna be noisy under there.

Never compare anything about a diesel to gasoline. About the only thing in common is that they both have a crankshaft and rotate. For example when a diesel is "blackstacking" (blowing thick black smoke) it's burning-up inside. When a gas engine blows black smoke, it's running rich & cold.
Click to expand...

I know diesel's are quite different, it's just that most of my adult life I've been driving diesels and noise-wise that's my only reference. I just thought, if someone else knows how a diesel sounds when the timing is off AND he knows what could give a similar sound on the XJ engine (not necessarily the same phenomenon), then he/she might be able to make the link. I can't.

Will wiggle the injectors to see what that does! Thanks for that one. Please see also my questions on the temp sensor (whether or not I can disconnect that for testing) in previous post.

Thanks,

Tim
 
Forgot to answer this one. The rough running in the warm-up cycle has been there, before and after the manifold work. When doing the manifold I hoped that may solve it but it had (in retrospect: of course) nothing to do with it. Needed to be done anyway.

Tim
 
The ECM is on the fender just behind the left front headlight.

On the Renix and most others, the temp info if for fine tuning mixture. Some may have algorithms that will drop it into open loop mode below a certain temp, but I've seen it almost cold and still closed loop, so not sure who does it and when. It will estimate a temp if the sensor is out.

Disconnecting the O2 sensor will cause it to go into open loop. My guess so far has been it's giving BAD data, not no data. That's different, but it can give clues. Disconnected, it may turn on the emissions light, but it uses pre-programmed settings in open loop. If you do it when acting up, and it clears it, that would be partial confirmation. IF you disconnect it, and it changes rough to good, leave it off for a day and see if it maintains some level of consistency - unlike now. If so, it's in there. Clean the connector, try it connected again, and if it goes bad again, it's confirmed. Before replacing it, also see that the heater is getting voltage. If I remember (docs not with me right now) there is a relay that powers that heater. No voltage to it would cause the same thing as a bad heater inside the sensor.

You can also use an ohmeter to test the heater in it, that would be final confirmation. A voltmeter on the signal, correlated to bad/good times would also confirm.
 
Will play with the disconnecting the O2 sensor and tem sensor. Seems like a good place to start.

What resistance values (Ohms) shoud I measure over the O2 sensor (heater)?

How many volts should be on the heater?

Just some values so that I can verify them. I'm pretty sure my Haynes hasn't got these (will check tonight) but maybe such data is in the FSM?

Thanks,

Tim
 
I don't mean to steal the thread, But this describes my jeep to a tee. The difference being that I haven't tried to turn it off and on again to see if it dissapears. One day, mine just started running perfect again right out of the blue, and I even noticed an increase in fuel mileage.
Since then, the sympoms have returned.

Which O2 sensor are we taking about? The one at the manifold or the one after the Cat? Or either could be the culprate? I noticed it running like this after I had done some exhaust work so maybe it's a connector as well?

TEMP SENSOR:
I would like to add that my electric fan was running intermittently and would even run for an exptended period while the jeep was not running. I would turn the key back on and it would stop.
I didn't notince any tep changes but replaced the sensor anyway.
That problem is now fixed, the fan runs only when it's supposed to but hasn't done anything to improve how it runs.
 
Last edited:
In my case we are talking about the O2 sensor close to the manifold (upstream) since in 1992 they only had one. I think the one downstream is a check on the cat, not sure though (I don't have one).

Glad to hear I am not the only one! That way we can both try and learn. So yours was running fine for a while? What changed, different fuel maybe, something else? But like mine, it always runs good when warm right?
 
Ok, an update and more questions. Finally got some time to try some things on the XJ again (it's my wife's dd). When warming up and running rough I've tried fiddling with the injector connectors & wiring, I've checked the connections on the ECM, and looks fine (but there's not much more I can say about that since I only have a multimeter and don't know what to measure there; my Haynes has not much to say there).

To narrow down the search a little I've replaced the O2 sensor (wanted to do that anyway) and the temp sensor on the t-stat housing. Also tried unplugging that one when it was running rough (during warm-up); changed nothing just threw a check engine light at me for doing that.

So, here's what I think. Somehow the computer gets misinformed about the engine's status. I am not sure whether or not it tries to go into closed loop or something else (how does it sound if it tries to inot closed loop when warming up?). Now what sensors can be doing that? Is there any way the computer may be off? Can someone hand me some steps to diagnose the issue and identify the culprit?

Thanks for any more input!!

Tim
 
The thing that tells the ECM to go into closed loop is an O2 sensor voltage starting to swing. That gives it something to work with and adjust timing and mixture. If it's doing this cold, and if the sensor is new, it's probably not that then, although try disconnecting it when running fine and see if it makes it run like problem times. One thing you didn't say that you checked was to see if the sensor heater was getting voltage. There is a relay that powers that, and without it, it extends the time to go into closed loop. It should be fine warm and in open loop though, but maybe you have something that contributes to it not running fine under preset mixtures... The timings you give are very similar to the timings of an O2 sensor with and without heater coming up.

Seeing as how you worked on the manifold, the egr, if a 92 has one, would be my next bet. If it's like the 89, you can play with the shaft. This would be a big bet. It might stick when cold. If it has one, then disconnect the vacuum line to the egr proper (not to a solenoid valve) while it's hot, at idle, and fine. Plug it. See if the next cold start runs any better. Fiddle with it. Something tells me the 92 doesn't have one though, and it's too cold and late to go check mine...

Next two would be to listen for a vacuum leak. If it makes it run rough, and not just raise the idle, it'll be big. The other is the throttle position sensor.

The key symptoms seem to be:
- happens at a little warm (unless the O2 was faulty, it's just gone into closed loop then)
- goes away with rising temp (it's near the block or manifold)
- can go away immediately with a restart (electrically driven or sticking vacuum driven part)

You said a restart can almost always make it go away. Instead, make a list of electrical components and their connectors, starting with relay type stuff. When it does it's thing warming up, quickly try disconnecting and reconnecting each. You may have a few weird other effects, but it might narrow down which thing is responding to having power cut then reapplied. Many items may cause it to run rougher when disconnected, but if it works great when you quickly connect it again, I'd dig further down that path...
 
skipc said:
You said a restart can almost always make it go away. Instead, make a list of electrical components and their connectors, starting with relay type stuff. When it does it's thing warming up, quickly try disconnecting and reconnecting each. You may have a few weird other effects, but it might narrow down which thing is responding to having power cut then reapplied. Many items may cause it to run rougher when disconnected, but if it works great when you quickly connect it again, I'd dig further down that path...
Click to expand...

OK, it was late.... the list should be related to engine operation stuff... and the disconnect should not change things (much) if you hit the bad part, but reconnecting it should make it better if you can do this during the bad running times. Just as if you only turned the bad part off and on again.
 
skipc said:
The thing that tells the ECM to go into closed loop is an O2 sensor voltage starting to swing. That gives it something to work with and adjust timing and mixture. If it's doing this cold, and if the sensor is new, it's probably not that then, although try disconnecting it when running fine and see if it makes it run like problem times. One thing you didn't say that you checked was to see if the sensor heater was getting voltage. There is a relay that powers that, and without it, it extends the time to go into closed loop. It should be fine warm and in open loop though, but maybe you have something that contributes to it not running fine under preset mixtures... The timings you give are very similar to the timings of an O2 sensor with and without heater coming up.

Seeing as how you worked on the manifold, the egr, if a 92 has one, would be my next bet. If it's like the 89, you can play with the shaft. This would be a big bet. It might stick when cold. If it has one, then disconnect the vacuum line to the egr proper (not to a solenoid valve) while it's hot, at idle, and fine. Plug it. See if the next cold start runs any better. Fiddle with it. Something tells me the 92 doesn't have one though, and it's too cold and late to go check mine...

Next two would be to listen for a vacuum leak. If it makes it run rough, and not just raise the idle, it'll be big. The other is the throttle position sensor.

The key symptoms seem to be:
- happens at a little warm (unless the O2 was faulty, it's just gone into closed loop then)
- goes away with rising temp (it's near the block or manifold)
- can go away immediately with a restart (electrically driven or sticking vacuum driven part)

You said a restart can almost always make it go away. Instead, make a list of electrical components and their connectors, starting with relay type stuff. When it does it's thing warming up, quickly try disconnecting and reconnecting each. You may have a few weird other effects, but it might narrow down which thing is responding to having power cut then reapplied. Many items may cause it to run rougher when disconnected, but if it works great when you quickly connect it again, I'd dig further down that path...
Click to expand...

Thanks for the input! Pretty sure my truck does not have an EGR, if it had, that would have been a good bet.

Now, what sort of electrical switches should I and should I definitely not (harmful) touch in this line-up? Any suggestions?

I am not really comfortable with just randomly disconnecting wiring etc. if I don't know what they are doing. I guess for these kind of system-things I really need to get the FSM and read it in detail (That I will do today) to get a good feel for the system as a whole.

Thanks again, and if you have more thoughts please let me know!

Tim
 
I forgot to suggest a few other things. Again, it was late... ;)

When it's running bad, can you use the throttle and raise the rpm to some point and get it to run smooth? Will just 400 or 500 rpm do it? None at all?

Disconnect a small vacuum line (not to a sensor) and how does it change?

Put a digital volt meter on the O2 line and see if it is bouncing around high then low about every couple seconds... see if it first starts doing that when it smooths out again. Putting that cycling in perspective with the problem may help a lot.

Also, just before shutting down when hot, disconnect the idle air control (if it goes away above idle). Leave it disconnected for the next start and see if it goes bad the next time.

I dont' have a 92 FSM, although I wish I had. I can think up generic items to check, but the manual will tell where and all...
 
skipc said:
I forgot to suggest a few other things. Again, it was late... ;)

When it's running bad, can you use the throttle and raise the rpm to some point and get it to run smooth? Will just 400 or 500 rpm do it? None at all?

Disconnect a small vacuum line (not to a sensor) and how does it change?

Put a digital volt meter on the O2 line and see if it is bouncing around high then low about every couple seconds... see if it first starts doing that when it smooths out again. Putting that cycling in perspective with the problem may help a lot.

Also, just before shutting down when hot, disconnect the idle air control (if it goes away above idle). Leave it disconnected for the next start and see if it goes bad the next time.

I dont' have a 92 FSM, although I wish I had. I can think up generic items to check, but the manual will tell where and all...
Click to expand...

Revving doesn't help. It will rev, but will remain rough. I tried that before.

Does it matter which vacuum line? There are quite a few of them and most of them i don't know what they do.

I will definitely measure the O2 sensor. Be interesting to see what it tells me.

Where is the idle air control located? (will see if I can find in the Haynes)

Thanks again! My wife's got the XJ this week on a trip so will get to it this weekend and will report back.

Tim
 
The easiest to get at and try is the one to the power brake booster. HOWEVER, it will give you the ability to let in a lot of air, or just a little, and it's disconnecting has nothing to do with the engine other than causing a vacuum leak - no sensors on teh other end. Just put your thumb over the port right away and ease it away to try different amounts of 'loss'. See how it responds with a little open and how it does as you open more.
 
You must log in or register to reply here.

Similar threads

G
1997 Sport running rough after warming up
Replies
8
Views
2K
Gilpine
G
B
2001 XJ dies at highway speed
2
Replies
20
Views
2K
N8N_99xj
N
A
Intermittent Idle/Bog
Replies
0
Views
282
azhang253
A
B
Jeeper needs Help !!
2
Replies
22
Views
3K
denverd1
denverd1
S
Rough idle, positive pressure from radiator with cap off...
Replies
4
Views
866
Slingshot505
S
Back
Top