Jeep Wrangler 4.0 idle misfire and bad mpg

Hi everyone,

I have 1995. (pre-OBD II) Jeep Wrangler with 4.0L engine and 32rh automatic transmission. Jeep was recently completely rebuilt. Apart from complete engine rebuild, it got new spark plugs, coil, distributor, dist cap and rotor, injectors tested and cleaned, all filters replaced and fuel tank cleaned. I did smoke test with some DIY machine and there were no vacuum leaks present. I suspect there is some minor exhaust manifold leak from crack but it is barely noticeable.

I'm trying to catch some irritating misfire that seems to be at random moments but is mostly present on idle. When engine is cold ( eg has been sitting for 4-5 hours on cold weather) car will start but it will hesitate and buck randomly a couple of times for a minute or two until it warms up (I'm not sure if this is related to going into closed loop). Car drives fine after that but slight to mild misfire is present when idling, and it is not really consistent when it comes to frequency.

When I hook up PCM to my Snap On MT2500 there are no DTCs, all parameters seem to be normal except Fuel Trim which is 3 to 4% (max is 25% per snap on documentation). By looking at this I can conclude that the car is dumping more fuel. Could the reason for this be cylinder bore increase for .020 oversize pistons? IAT, MAP, CTS, TPS - they all seem to be valid. I can trigger injectors one at a time from MT2500 and test IAC.

I checked spark plugs and there are some oil (not fuel) deposits on one of them, but cleaning them doesn't change anything. I'm suspecting deposits came from valve stem seals since compression on each cylinder is 150psi which is maximum per manufacturer.

Fuel pressure test showed some 29-30 psi on idle. Per FSM, it should be 31. Maybe my gauge is not that accurate, but the reason why I don't suspect fuel pump or regulator is because there are no performance issues when it comes to driving this car.

I recently got oscilloscope, thinking this might be better thing to do than randomly replacing sensors. First thing I did - I tested crank and cam sensors. And here is what I got:



This pattern seems to be constant. I understand that there should be four equal peaks for each cylinder. This is definitely not a good CPS signal. I tried to adjust sensor, but no difference, pattern is constant. Should I suspect CPS sensor or flexplate? It could be that the flexplate got damaged while we were mounting transmission to the engine.

My concern is if this is the only issue. Could this disrupted signal cause huge mpg decrease, idle misfire and bucking when cold?

I would suspect a flexplate issue at first glance of your waveform. The big gap suggests a bent reluctor or metalurgy problem between two of the four pulse sets. The other problem is the wider second pulse out of four - kinda suggests someone used the CKP window as a pry point to turn the engine over. Not pointing fingers, just seen it before.

That said, I'm honestly surprised this truck runs as well as you say it does! :silly: It also looks like the distributor/CMP timing is slightly retarded, but that still isn't the issue.

I don't know what the increased displacement will do to the fuel calculation? But I wouldn't sweat the 3-4% fuel trim at all. Same with the fuel pressure. Find out why your CKP waveform is off first, then proceed further.

I'm not a mechanic and this is the first time I did complete rebuild. And you are absolutely correct. I remembered something and now I understand that using a flexplate for blocking the engine was not a good idea. I have spare one so I will drop the transmission and replace it.

As for the fuel trim, I thought that slight increase in displacement might require a bit longer injector pulse because of more oxygen being sucked inside cylinder.

Timing - both sprockets and timing chain are new, just like distributor. When I put distributor in, I remember that I could move it forth and back a bit, really not much, before fully tightening it. Could that be sufficient to cause this sort of signal? Camshaft is not new so there might be some slack between distributor and cam gear.

I guess that timing chain is not an issue for this since compression test showed maximum per FSM.

Rutesic wrote: As for the fuel trim, I thought that slight increase in displacement might require a bit longer injector pulse because of more oxygen being sucked inside cylinder.

You're probably right? I honestly don't know! :silly: In my head, the PCM is basing the fuel calculation on the MAP sensor, and theoretically the vacuum level in the intake should be the same between the original displacement and where you're at now. Someone please correct me here.

When I put distributor in, I remember that I could move it forth and back a bit, really not much, before fully tightening it. Could that be sufficient to cause this sort of signal?

Definitely. As you noticed, it's really not much, just enough to get the waveforms squared. You can do it live with the scope, but it's easier with the scanner if the function is available. I usually see it labeled 'Set SYNC Signal'. You hit the button, the PCM raises the idle to 1000 RPM, and you turn the distributor until the sync value is zero.

If that's not available, just move the distributor until the transition of the CMP is square in the middle of a low section on the CKP. Or as close as you can, anyway. :silly:

Thanks a lot! I will post results here once I'm done with the repairs

One thing that hunts me anyway... When does ECU fire the injector / coil ? I see there are groups of 4 pulses but on which of these is the injector/coil fired?

It's up to the PCM. :lol: The individual pulses represent timing in relationship to TDC for the three cylinder pairings. This picture might help:


IIRC, the coil ends up getting fired around the third or fourth pulse at idle.

Ok, so here are the results.

I replaced the flexplate. New one is correct part number, made by ATP. Seemed to have reluctor windows on same locations as original one. I reinstalled same CPS.

What happens now is that I have the same pattern, but it moved one group to the left. But pattern is the same, one a bit longer window and later huge one. Car hesistates around 2000 rpm. I don't think that this is flexplate related now since the patern is exactly the same.

Can this be somehow related to torque converter? Like it is bent or something? Or can I blame CPS sensor for this? I installed spacer but I'm not sure if it is correct size

Here is the new waveform. Again it seems to be incorrect.

Hey this is Ruben in Phoenix Az, seems to me its a great time to put a smoke generator to the intake manifold. Sounds like a vacuum leak when cold and those intake manifold gaskets leak all the time. you can put a rubber glove over throttle body to increase smoke pressure. Test it when cold. Timing looks in sync to me. Pico has a good library of know good waveforms but they look good. But ofcourse each vehicle is different.

Smoke generator will probably reveal vacuum leaks in the intake gasket or EGR valve port

Thanks Ruben, however as per documentation this waveform is incorrect. I got this with car running on almost constant rpm when it's hot. It should have groups of four per each cylinder, each of equal length

According to your signal and the chart posted by Tyler, after the new flex plate was installed it is happening:
between cylinder 5 TDC and 3 TDC
between cylinder 2 TDC and 4 TDC

Is it possible to rotate the engine over by hand until you get to that extra wide malfunction in the wave form, remove the sensor then take a look at what is near it?

Maybe it's the reluctor teeth ( but doubtful since you just replaced the flex plate )
Maybe the grooves and the sensor are both fine, but some other piece of metal is coming near the sensor.
Maybe something down there became magnetized ( which you of course could not ever visually see ) and is affecting the sensor.

Did you visually inspect the reluctor teeth when you installed it? Are they all exactly the same thickness and shape?
Do they all have the exact same gap between them?

I don't work on Jeeps ( yet ), so I don't know how hard this would be.

I also noticed on closer examination of the wave form that all of the lower or bottom parts of the wave form look correct, but something seems to be forcing it to the high state at certain times.

Maybe a wire or plug is rubbing on something that is shorting it to power somewhere? I don't know what would cause it to happen at an interval that regularly... but something must be causing it.

It also now seems to happen right when the cam signal switches. Is it possible to run the engine with the cam sensor unplugged to see if the cam sensor signal is somehow crossing over into the crank signal?

I will try to turn the engine by hand and check if this error still appears. However, it could be that it doesn't pick up signal only when it's turning fast.

I can run car without cam sensor once it fires up, so I will try that one to. They share the same ground and reference signal from PCM, so this could be an option too. However, I doubt a bit since in previous waveform (before replacing flexplate) that glitch was outside of cam sensor switch

Ok, so this is driving me insane now.

I bought new NTK crank sensor. Now I have three crank sensors. Those sensors are held with one bolt and should come with paper spacer for air gap adjustment. Two of them give the same pattern like I posted above after flexplate replacement. Doesn't matter if I do any adjustments. Those two are some cheaper aftermarket ones. I doubt they are bad since they give exactly the same waveform.

If I push new NTK CPS all the way down in the hole, keep it pressed and hand-tighten it, it works. On idle waveform is perfect. Sometimes it will backfire above 3500 and I'm seeing that some peaks are missing from the waveform or some of them are too long, but this is completely random. Real problem comes if I tighten it with wrench it instantly starts running like crap. (it is not hitting flexplate at that point). What the hell could this be?

BTW Can too much spark advance cause serious damage (pistons, rings, bearings) ? I think sometimes it fires spark too early because of crappy signal

Can you show us these 4 wave forms:

1. Crank vs Cam with it hand tightened
2. Crank vs Cam with it wrench tightened
3. Crank vs Cam with it hand tightened with cam sensor unplugged
4. Crank vs Cam with it wrench tightened with cam sensor unplugged

This will tell you exactly what is going with that sensor.

hakachukai wrote: Can you show us these 4 wave forms:

1. Crank vs Cam with it hand tightened
2. Crank vs Cam with it wrench tightened
3. Crank vs Cam with it hand tightened with cam sensor unplugged
4. Crank vs Cam with it wrench tightened with cam sensor unplugged

This will tell you exactly what is going with that sensor.

I did this test. Waveform doesn't change with cam sensor unplugged at all. This is how it looks no matter cam sensor being unplugged (I tested with multiple crank sensors. .

Since I have FOUR crank sensors, I did a little test. I noticed that hole on NTK sensors is positioned too low, so I drilled it a bit. By doing this I can adjust the air gap between sensor and flexplate. I pushed it all the way down I could and still it isn't touching flex plate when tightened. After doing this waveform looks much better. The glitch that is marked red disappeared, that peak is perfect now. However the one marked with blue color is still there.. It is not as wide as in this picture but it is a bit wider than all the others (maybe twice the normal one).

Since all crank sensors give same signal, and drilled one gives better, my guess is that this flex plate is bad (probably out of round) and only by decreasing air gap sensor is able to pick it up correctly. If flex plate was ok, then all of them should be able to give proper waveform. One more reason to suspect this is because on original flex plate (which was definitely bent by me) glitches looked completely different (image from first post).

Does this make any sense?

Thanks for following up!

I'd remove the sensor and manually turn the engine over with a ratchet.

1. Check the shape of each reluctor on the flex plate
2. Make sure that it is clean ( especially from metal shaving or anything else that would effect its magentic properties )
3. Measure the depth from the bottom of the sensor hole to the flex plate at each position ( you can do this using a pencil if you are careful )

I'm not sure if it makes sense that a bent or damaged flex plate could cause this problem or not.
I'm assuming that:
A. The sensor on the flex plate is your Crank sensor.
B. The Cam sensors are mounted somewhere else.

Yet, you seem to be saying that changing the gap between the flex plate and the Crank sensor is affecting Cam sensor output?
So messing with the Crank sensor is affecting the Cam sensor?
Am I wrong about this?

hakachukai wrote: I'm not sure if it makes sense that a bent or damaged flex plate could cause this problem or not.
I'm assuming that:
A. The sensor on the flex plate is your Crank sensor.
B. The Cam sensors are mounted somewhere else.

Yet, you seem to be saying that changing the gap between the flex plate and the Crank sensor is affecting Cam sensor output?
So messing with the Crank sensor is affecting the Cam sensor?
Am I wrong about this?

Thank you for your fast responses!

I have only one crank sensor on the flex plate and one cam sensor is inside distributor. Changing the gap between flex plate and crank sensor only affects crank sensor output ( closer the better). As for the cam vs crank - nothing changes in crank waveform if I unplug cam sensor and nothing changes in cam waveform if I reduce crank sensor air gap.

ok, thanks for the correction. I must have been mixing up the two signals ( it has been a while since I looked at this thread ).

yeah, I'd definitely do those checks on the flex plate then.