Very odd idling issues

DISCLAMER- yes I have read eons and eons of threads on this and have done a crap ton of research. This is a special monster. Hello everyone. I’m new to here. I had a few questions maybe someone might be able to help me pin point. Basically as the post says, I’ve got some strange idling going on. For starts I’ll say, I own a 97 Cherokee sport with the 4.0l 199k on it. To better dive into this. Whenever I start it up it idles at about a grand or so and the drops down to around 7 and half after a min or so, and then it’s begins to slowly drop lower and lower till it’s around 500. It begins sporadically misfiring here and there, nothing significant you can only hear it from the back end of the car, or in the cab. So I did some work. Of course the usual. Spark, rotor, plugs, ignition coil, cleaned the throttle body, cleaned the IAC, tested the TPS, tested fuel pressure (read at 45-50 psi when turned on by jumping the pump. And then of course while it’s running as well. I did notice that the pressure went down to around to 40ish after 30 seconds, and then dropped to 30psi after about 2 mins, don’t think it’s the issue. But I wanted to note it) then I checked spark plug gape, read at 0.35, so it’s fine. After that I did a test on the injectors to see if they were spraying ok. Definitely not atomizing well, so I have some multi points on their way. And that’s when I noticed another really strange thing. After I did all that and tested the injectors. I started the car and and gave it a little gas and it would spike up to 1,500 then foot off gas and it would bring it down to 1000 and then after a second or two drop down to 750 and then start it’s normal rough idle. Then I did it another time. Same thing happend. And then another. This time it spiked up and began to climb. 1500, foot came off the pedal, then it spiked up to to 1700 then 1900 and then 2500 (all in 3-4 second steps. Not a continuous pull) so I shut it off and called it a night. I figured I pissed it off enough and got back to it the next day. I then checked it and it seemed to “work its self out” and didn’t climb anymore but continued to step down after reving to 1500 the foot off,brought its self to 1000 held for a second, then to 750 and then down to rough idle again and misfiring BUT WONT DIE. So what I’ve narrowed it down to is, map sensor, crankshaft position sensor (maybe? I feel like I’d have more extreme things going on if it’s that), cam shaft position sensor (dizzy does have a little play but I feel like that’s normal), o2 sensors, or a good ol tiny hidden vacuum leak, or perhaps engine coolent temp sensor or intake temp sensor I’m not sure. Oh it’s also worth mentioning I did a vacuum test and it read about 16 pounds of mercury, and was in the labeled area late timing. So I believe I’m resulting in timing issues because a sensor is trying to compensate. or a waky fuel, air, spark ratio. If you have any insight or ideas please let me know before I start firing the parts cannon at this.

I'd start by pulling the codes both the in-vehicle codes and with a code reader. Not always the answer, but sometimes an indicator. Don't get hung up on swapping sensors or believe every sensor code is the answer. Oil, corrosion, and even coolant in the connectors can cause issues. My experience tells me it is just as often a continuity issue as a sensor issue.
Sounds like the TPS but could be a number of other issues, even multiple issues. Often when the TPS screws up on later models the shift points on an auto trans go a little wonky.
Maybe check out the basics, like tighten intake and exhaust manifold bolts. I've had to retorque mine multiple times over the years. Check your spark and read your sparkplugs. Do you ever have to add coolant, I mark the side of my coolant reservoir and look periodically before start up (cold). A tiny head gasket leak can cause all sorts of grief. If you stay on top of it you can notice the level slowly falling over weeks or even months.
Simply separating a connector and replugging it in can slightly change the resistance values. I spray my connectors out with a high-quality contact spray. A leaky water pump or a tiny radiator leak can soak the connectors with coolant and mess with your low voltage sensor values and even the 12 volt systems.
Swapping out sensors is IMO a last resort. making sure all the peripherals are healthy often solves issues. Though doing the job right is time-intensive, there is sometimes no easy fix.

Fuel pressure must be 49 +/- 5 psi.

Start there.


Low battery voltage is always a likely suspect for the falling rpms's, clean and inspect the wires and wire connections, test the battery.

Oh wow. I honestly didn’t exspect any replies. Thank you. Ok so tonight I’m checking spark seeing if it’s white or yellow, but also i ment to but forgot to check and see if maybe the spark plugs are a little black or white or even wet with gas to help indicate if the engine is compensating. Considering I only get to work on her on the weekends and when I have a bunch of extra time. But I’ll start there. I do have a extra tps. I’ll go through and tighten all the grounds and clean them. If I’m not mistaken there’s 4? One by the battery on the fender, one ribbon on the firewall to suppress radio interference anddddddd two others I think? I’ll also take the battery up to autozone and let them test. I did just get the battery a few weeks ago but whenever I leave for the weekend I try and take the battery out and put it on the house. I live in Indiana and it got cold for a min (the vehicle is stored in a garage though). I’ll also go through and inspect all connections and clean and replug them. Maybe jiggle some points on the harness and see if anything changes. I did blow some smoke into the brake booster vaccum line that runs into the intake manifold to check for leaks considering I did a throttle body gasket. There was the smalls amount of smoke that escaped from a corner of the gasket it seemed so I’m going to redo it and torque to spec. Spec is only like 10pounds or something I can’t remember currently. But hey if I got it let’s do it. I’ll also do the iac one more time and the tunnel in the throttle body leading to it while I have the throttle body off. I’ll check radiator fluid as well as tighting the intake and exhaust to be sure. And finally I’m gonna check around the injectors and intake manifold for leaks with some carb cleaner and see if I can spot a vacuum leak, then when I get the compression tester in I’ll be doing a test for that as well. I’ll update tonight

Ever replace the plug wires? Check to make sure your intake manifold bolts are snug enough.

I accidentally let my battery go dead one time and didn't realize it. I tried starting my XJ, and got some clicks, but no crank. I realized the battery was dead, so I put the charger on it. After the battery was charged, it fired right up, but idled pretty much as you said. After it was warmed up, it would barely maintain idle at stop lights. Sometimes it would die. My XJ didn't do that before the dead battery. After checking a few things with no improvement, I reset the ECU by momentarily disconnecting the battery, and then it started idling normally again. Its probably not the solution to you issue, but OTOH, it only takes about 2 minutes to disconnect the battery and reconnect it.

cruiser54 said:

Ever replace the plug wires? Check to make sure your intake manifold bolts are snug enough.

Click to expand...

Yes I have replaced the plugs, wires, cap and rotor, and ignition coil. All as just a tune up. I did the ignition coil because the housing was cracked and thought it could of been the cause but wasn’t. But I’ll do those bolts like you said

maxbraketorque said:

I accidentally let my battery go dead one time and didn't realize it. I tried starting my XJ, and got some clicks, but no crank. I realized the battery was dead, so I put the charger on it. After the battery was charged, it fired right up, but idled pretty much as you said. After it was warmed up, it would barely maintain idle at stop lights. Sometimes it would die. My XJ didn't do that before the dead battery. After checking a few things with no improvement, I reset the ECU by momentarily disconnecting the battery, and then it started idling normally again. Its probably not the solution to you issue, but OTOH, it only takes about 2 minutes to disconnect the battery and reconnect it.

Click to expand...

Interesting. Now that you mention it. This all started around the time I put a new battery in it because the last one died. I never did reset the ecu. That’s so odd though. I would never think something like that could throw it off. But shoot it’s worth a shot. I’m guessing your talking about the process where you touch both the positive and the negative together for a allotted time? Also as a side note I’m getting no codes from the dash. No check engine light. And whenever I do the command for codes. I get the battery code and then end message code. Witch seems suiting for the last 50 start ups

Here is my check list for the weekend. In no specific order.
Find all grounds. Clean and tighten them
Spray carb cleaner around injectors and throttle body
Check vaccum below map sensor
Test spark for all wires (yellow bad, blue/white good)
Smell exhaust. Smells of sulfur or eggs. O2 sensor
Check cts (coolant temp sensor) and mat (manifold temp sensor)
Check spark plugs for wetness, blackening,(rich)or whitening(lean). Gray is ok
Tighten intake and exhaust bolts down to torque spec.
Reset ecu with positive and negative grounding
Send battery to autozone to be tested
New battery connectors.
Autozone test for stored codes.
TPS swap
Map swap
Injector swap
Throttle body cleaning and new gasket
Clean iac again

Any codes?
I'd start there. Next, get some data. Do you have a scanner that can get data? We need to see fuel trims and what your O2 is doing.

Sent from my Pixel 3 using Tapatalk

Ahvicha said:

Interesting. Now that you mention it. This all started around the time I put a new battery in it because the last one died. I never did reset the ecu. That’s so odd though. I would never think something like that could throw it off. But shoot it’s worth a shot. I’m guessing your talking about the process where you touch both the positive and the negative together for a allotted time? Also as a side note I’m getting no codes from the dash. No check engine light. And whenever I do the command for codes. I get the battery code and then end message code. Witch seems suiting for the last 50 start ups

Click to expand...

Its only necessary to disconnect pos or neg from the battery to reset the ECU. No need to disconnect both and touch cables together, but it doesn't hurt anything.

8Mud said:

I'd start by pulling the codes both the in-vehicle codes and with a code reader. Not always the answer, but sometimes an indicator. Don't get hung up on swapping sensors or believe every sensor code is the answer. Oil, corrosion, and even coolant in the connectors can cause issues. My experience tells me it is just as often a continuity issue as a sensor issue.
Sounds like the TPS but could be a number of other issues, even multiple issues. Often when the TPS screws up on later models the shift points on an auto trans go a little wonky.
Maybe check out the basics, like tighten intake and exhaust manifold bolts. I've had to retorque mine multiple times over the years. Check your spark and read your sparkplugs. Do you ever have to add coolant, I mark the side of my coolant reservoir and look periodically before start up (cold). A tiny head gasket leak can cause all sorts of grief. If you stay on top of it you can notice the level slowly falling over weeks or even months.
Simply separating a connector and replugging it in can slightly change the resistance values. I spray my connectors out with a high-quality contact spray. A leaky water pump or a tiny radiator leak can soak the connectors with coolant and mess with your low voltage sensor values and even the 12 volt systems.
Swapping out sensors is IMO a last resort. making sure all the peripherals are healthy often solves issues. Though doing the job right is time-intensive, there is sometimes no easy fix.

Click to expand...

Alright so here’s a little update. I sent the battery in tonight to get looked at. I cleaned all engine grounds, put new battery connectors on it and recreated the throttle body and iac. Tomorrow I’m going g to put a new gasket on it with rtv as well and tighten to speck. There was something I noticed when I took the throttle body off. There’s this residue inside the intake manifold and it smells like gas. And then I checked the spark plugs and they look a little wet and smell of gas and are a touch black. Then I looked at the exhaust and the tail pipe is is pretty black as well. I tired to get the down stream o2 sensor off but man is that stuck on there and I’m a little afraid to strip it. Any tricks or tips. I put a shit ton of pb blaster and heat on it.

I am going though the same issue with my 2000 XJ. Every time I replace a sensor the engines idles better.

I also had a stubborn rear o2 sensor. I sprayed PB blaster the night before. Worked on it by loosing it then tighten it. I used an extension for extra leverage and it finally came out.

puredrive said:

I am going though the same issue with my 2000 XJ. Every time I replace a sensor the engines idles better.

I also had a stubborn rear o2 sensor. I sprayed PB blaster the night before. Worked on it by loosing it then tighten it. I used an extension for extra leverage and it finally came out.

Click to expand...

Did you use a flare nut wrench or one of o2 sensor sockets on a ratchet?

I've had pretty good luck heating then spraying (quench), the expansion and contraction often breaks the bond. Then when you try to loosen the O2 sensor jarring it works better than constant pressure. I struggled with the O2 sensors for years and finally bought a quality O2 socket, wish I would have done it sooner.
A faulty MAP can make it run really rich, but if it were that bad you'd likely have a code.
A faulty O2 sensor in my experience can make it run lean. Usually most noticeable right after it heats up enough to switch from open loop to closed loop, maybe after a cold engine running for five minutes or less. O2 sensors usually have to be very bad (completely dead) to throw a code.
The fuel moisture inside the intake would make me suspicious of the purge system either the solenoid or possibly the charcoal canister overflowing with fuel. It has happened before, no idea exactly why best guess is too much fuel tank pressure.
A llittle oil film inside the intake is fairly normal.

Hmmmmmm ok I’ll take a better look at that as well. I’m not exactly familiar with that system but I can look up some things I think. I’ll do more updates tonight

Ahvicha said:

Did you use a flare nut wrench or one of o2 sensor sockets on a ratchet?

Click to expand...

o2 sensor sockets on a ratchet. The Husky ratchet that I use can extend pretty far out, I then used a pipe to get more leverage.

Any codes. Are the monitors complete? Can you get any data?

Sent from my Pixel 3 using Tapatalk

8Mud said:

I've had pretty good luck heating then spraying (quench), the expansion and contraction often breaks the bond. Then when you try to loosen the O2 sensor jarring it works better than constant pressure. I struggled with the O2 sensors for years and finally bought a quality O2 socket, wish I would have done it sooner.
A faulty MAP can make it run really rich, but if it were that bad you'd likely have a code.
A faulty O2 sensor in my experience can make it run lean. Usually most noticeable right after it heats up enough to switch from open loop to closed loop, maybe after a cold engine running for five minutes or less. O2 sensors usually have to be very bad (completely dead) to throw a code.
The fuel moisture inside the intake would make me suspicious of the purge system either the solenoid or possibly the charcoal canister overflowing with fuel. It has happened before, no idea exactly why best guess is too much fuel tank pressure.
A llittle oil film inside the intake is fairly normal.

Click to expand...

Hi guys sorry little late to it. So here’s a update for anyone following still. I did a compression test. All came back in great shape 140-145. Tried and tried and tried to get down on that 02 sensor and I got nothing slightly stripped now. But here’s another thing. I went to auto to get codes and it came back with a down stream o2 sensor high voltage reading. And then I got to thinking.....the o2 sensor I’ve been yanking on was the the upstream. This whole time. The upstream is fine. The down stream seems to be the problem. But....there is no down stream. It seems as though someone took it out along with the cat. Now I could be wrong. I’ll look one last time. Maybe I just wasn’t paying to much attention but I just didn’t see what looked like a cat or a sensor back there. But why would I be getting a reading saying high voltage if it’s not there. If I am right (considering I only see the Cherokee on the weekends) then I’m gonna need to put a new cat on it and a o2.

I know how does someone miss something like a cat on their exhaust. But if I’m honest. I haven’t been exactly paying attention to the rear for much time. I mean it’s a full exhaust but I don’t think I saw anything before the muffler besides a straight pipe. So would the fact that I’m missing the down stream sensor be a cause of the idle problem? I was under the impression that if you unplug them or one the car defaults to poticular settings to run. And that would “fix” the rough idle till you swapped them out. But again considering there’s something there maybe the other o2 is compensating? B then I would see some sort of code for that maybe? Maybe I’ll unplug the upstream too and see what happens. But besides that I put a new map sensor and multipoint injectors on it. I also redid the throttle body gasket and put rtv around it as well, and of course cleaned both the throttle body and the iac again. Could the CTS be a corporate? I’ve had the clutch fan off for sometime do to just being able to get into the bay better. I wouldn’t imagine it would rough idle from that. As a review the rough idle starts in after a min or till the rpms drop to 750. All other strange issues have been cured. The only thing to resolve now is the random miss. And to figure out why the engine is in “late timing” so says the vaccum gauge (reads 15-16 pounds of mercury, and floats in a section called late timing under it)