Starting to lose hope: rough idle and dying in park

[peoplethatcant]

I've been working on a way to bench test one with an Arduino (which obviously everyone has lying around) but while it worked in the sim, it did not work on the bench so I've tabled that for the time being.

Interesting. If you ever want a 2nd set of eyes on your code/circuit, message me. I do know my way around arduinos and H-Bridges. No pressure though, sometimes the best projects need a timeout on the healing shelf for a bit, while they hopefully fix themselves.

Anyway, you've replace this much, why not it, too?

Yeah... Just trying to limit the total replacements. Each sensor/part is not that expensive on it's own, but add them all up, on top of having to redo most of the suspension/steering already... Just hoping to stick with firm maybes now instead of soft ones. My wife has come up with this weird idea of 'figuring out how much that Jeep has cost you' which I find just absurd, but it's worrisome if that type of thinking spreads. haha.

Remember that all these sensors run on basically the same 5V & ground, so if one is problematic it can affect the whole system. There are a number of posts on this forum talking about a rough idle that ended up being resolved by replacing the clock spring. I've never traced the circuit, but I guess it shares that same 5V & ground as the CPS, CKS, , etc and there's some evidence that it can cause a rough idle.

I did read that a failing clock spring can add a bunch of noise to the circuit, and mess with a bunch of the sensors. My thought was to borrow an oscilloscope at some point and see how noisy that circuit is.

It's also very possible you've got one or more semi-shorted wires buried in the harness somewhere. In my experience I can all but guarantee you do; every jeep does. Whether or not it's the culprit I obviously can't say. Just one more thing to consider as you troubleshoot.

This is my worst fear. I have a bad habit of going all or nothing. So it'll be a nice easy find, or I'll end up with the body off making a new wiring loom from scratch during an engine rebuild. :S only half joking.

Good luck, these are always fun to track down.

Haha. Indeed. Cheers, and thanks for the input.

 

[hear]

Interesting. If you ever want a 2nd set of eyes on your code/circuit, message me. I do know my way around arduinos and H-Bridges. No pressure though, sometimes the best projects need a timeout on the healing shelf for a bit, while they hopefully fix themselves.

Part of it is that I don't know the sequence, and I don't know enough about these to know if there is a standard or of I need to find the docs sheet for this part. Or if I just need to open one up and figure it out based on the poles & windings. Also the Arduino gives 5V, and while I proved that 5V will make the IAC move, but Im not sure if the computer is sending 12V or the normal 5V reference signal.

Yeah... Just trying to limit the total replacements. Each sensor/part is not that expensive on it's own, but add them all up, on top of having to redo most of the suspension/steering already... Just hoping to stick with firm maybes now instead of soft ones. My wife has come up with this weird idea of 'figuring out how much that Jeep has cost you' which I find just absurd, but it's worrisome if that type of thinking spreads. haha.

I have a google sheet with all my jeep projects, parts bought, etc. I removed the SUM( ) formula. It wasn't helping. I still think this is cheaper than a car payment, but it seems to come in waves that make it feel way more expensive.

I did read that a failing clock spring can add a bunch of noise to the circuit, and mess with a bunch of the sensors. My thought was to borrow an oscilloscope at some point and see how noisy that circuit is.

You can unplug the clock spring which people who have this problem report that it makes the rough idle go away. Definitely worth a shot.

This is my worst fear. I have a bad habit of going all or nothing. So it'll be a nice easy find, or I'll end up with the body off making a new wiring loom from scratch during an engine rebuild. :S only half joking.

I'm a huge fan of removing the 25+ year old harness and peeling back the dirty brittle loom and thoroughly inspecting the entire thing. I've never opened one up and not found an issue. The under cab harness unplugs at the rear driver taillight and can be pulled up into the engine bay if you don't want to remove the whole thing. But oil + heat + old age + rough living = high potential for insulation to be missing, wires to be cracked, etc. Plus re-installing a freshly re-taped & loomed harness is very satisfying.

 

[peoplethatcant]

Part of it is that I don't know the sequence, and I don't know enough about these to know if there is a standard or of I need to find the docs sheet for this part. Or if I just need to open one up and figure it out based on the poles & windings. Also the Arduino gives 5V, and while I proved that 5V will make the IAC move, but Im not sure if the computer is sending 12V or the normal 5V reference signal.

According to the attached snippet from the '04 FSM, the IAC needs 12v to drive it on each winding, and they need to be able to reverse polarity. This is where a stepper motor driver comes in (Which is just two H-bridges that allow polarity reversal and logic voltage/current shifts from 3.3-5v from the arduino up to 12v in this case).

It sounds like a pretty standard stepper motor arrangement, so without delving too deep, I'd start with the pre-built stepper motor library for arduino and a prefabbed stepper controller board for it that is designed for 12v.

The resistance I measured across each winding was about ~50ohms, so the current capacity for the motor is nice and low (roughly 12v/50ohms=0.25 amps per winding). You do need to energize both windings together to keep it in position, but as long as the control board can handle 1 Amp say, I'm sure that would be more than enough since any testing will be with no load on the pintle.

The only downside here of course is that while many people have an arduino kicking around, not as many have a stepper control board on the shelf I'm guessing.

*Edit - Also I've had arduinos do weird things when I try to draw a bit too much current off their pins, so it's possible your existing setup would work perfectly with only a couple transistors bringing the logic from the pins up to 12v just to make it move. I also have no idea how much or little you already know about all of this, so I just want to point out that if I'm stating the obvious or misunderstanding your setup, I mean no offence.

 

[peoplethatcant]

I replaced the IAC... it died after running for about 1 minute, never been that quick before.

I went back and bought a TPS as it seemed like a good contender from the comments above. The Jeep didn't die for 15 minutes at idle, the TPS reading on my scanner was now dead flat (before fluctuating 0.5% at idle). Drove it around the block, the fuel trims seemed to stabilize at ~+7 and +5 at idle, and the idle speed is down lower. The idle was smoother, but still... I don't even know what to call it... halfway between a slight miss and an engine shake about half the time.

I did put a vacuum gauge on the intake, and converted from the Inches of HG on the gauge to the absolute PSI rating of the MAP. If my gauge and math are correct, the MAP is showing perfect while the engine is off (Based on looking up the absolute pressure for my location at the correct hour today) but the MAP looks like it's out by 1-2 PSI at idle.

Figured what the heck, might as well collect the set. Bought an IAT sensor, and ordered a MAP sensor for Monday. Once the IAT was installed it once again died quite quickly.

So I'll see on Monday. If the MAP doesn't solve it, then after a short bank account and mental health break (haha) I'll look into the CPS, clock spring, and recheck all the basics like spark and fuel and wiring then likely ship my PCM out for testing. Just figured I'd put everything here in case anyone following along in the future trying to solve their own issue wonders.

I'm going to see if my local auto parts store will consider a sensor stamp card... The 5th one free would be great.

 

[Peakbagger]

Hemi swap?

 

[peoplethatcant]

Hemi swap?

Haha. With only 240k kms on it, I feel like a hemi swap might be a tad rude unless the bottom end goes on strike and demands more money.

 

[hear]

unless the bottom end goes on strike and demands more money.

Are we still talking about engines?

 

[peoplethatcant]

Are we still talking about engines?

 

[peoplethatcant]

Already changed: TPS, IAC, MAP, IAT, CKP, TB Gasket/Shaft O-rings, Battery, Battery Leads/Grounds, Injectors, Coil Rail
Disconnected: Clockspring
Backprobed: Sensor ground, +5v, o2 sensor wire, PCM ground, B+ at obdII port
Measured resistance from PCM to: IAC, TPS, MAP, Fuel Injectors, o2 sensors
Wiggled the heck out of every wire connector I can find in the engine bay looking for RPM drops
Put new O-rings in TB Shaft bearings

Remaining symptoms are:
-High Fuel Trims at idle/low throttle (+14%, +10%)
-Vibration/Almost grinding type noise at idle and light throttle, goes away with higher RPMs
-Rough Idle after the jeep sits at idle for more than about 20 seconds
-Loss of power on accel unless am heavy on the pedal
-The air wooshing noise on accel is now more of a whistle

Stalling and PCM Resetting were both solved by new battery, leads, and clean grounds.

-I can blow smoke into the intake as hard as I can, it holds pressure (After the TB shaft seals and gasket)
-O2 sensors look really slow to me. Even though there is no code popping up (attached photo, look at time stamps). And downstream sensors flatline at idle.

I am going to remove the serpentine belt and rule out the pulley bearings for the vibe/grind

Then however, IF the o2 readings do indeed seem super slow? And since I've already checked resistance in the o2 sensor circuit, and K4 sensor ground, the only remaining thing in the FSM for o2 slow to respond is an exhaust leak between the o2s and the head. That, or all 4 o2 sensors have gone consistently lazy in the exact same way together (one of which is only 4 months old).

Even though I've hooked up a vacuum to the tailpipe many times, and made sure the bolts on the downpipe and exhaust manifold are tight, it looks like a bad exhaust gasket, or a cracked header, is the only viable culprit here? (I believe the 04s don't have the super susceptible to cracks header, so hoping it's the gasket)

It's impossible to get a good view of every part of the gasket for bubbles with soapy water, so I think I just have to assume that a mysterious exhaust leak that I can only hear under hard accel is a thing? And start pulling off the intake/exhaust manifolds.

 

[peoplethatcant]

A couple clarifications, I'm hoping someone can shed some light.

The FSM says fuel pressure should be 58PSI for 2004 TJ, but I've seen 49PSI listed in multiple spots on this site, including here and my pressure is 48PSI?

The FSM also says the purge solenoid should be cycling 5 or 10 times per second. Mine seems to cycle closer to 2 times per second at idle, and unplugging it (electrically) makes the fuel trims go quite a bit leaner immediately, and drop back down if I plug it back in. I have done the the tests on it as per the FSM, it will happily hold vacuum, and tests good. Just curious if the slower cycling and fuel trim swing is normal, as the rough idle does seem to follow the purge valve's timing a bit.

Lastly, cylinder 3 plug was showing lean the last time I pulled them. I've ordered a cheap oscilloscope to rule out intermittent electrical issues with the wires/injector driver (all normal tests show good), and a laser temp gun to see if it's maybe a heat soak issue (I have all the heat shielding installed). I'm just curious with good compression, should I be looking deeper here (Burnt valve, bad spring, etc)? And are there any definitive tests for these other than pulling the head?