Hello everyone, i am having a really hard time tracking down a rough idle on my 1998 4.0 cherokee classic 202,000 miles. I have a brand new head, head gasket, exhaust manifold, 02 sensors, cap and rotor, wires, plugs, and crank sensor. there does not seem to be a vacuum leak. it idles between 730 and 760 noticeably bouncing up and down on the tach. also it feels like there is an intermittent misfire but there are no CELs. When looking at the live data, the TPS sensor seems to be reading correctly (although i am not sure if it can read correctly but still cause the issue). Idle are control valve seems to be fairly new. throttle body is clean as a whistle. not really sure what else to check.... it runs fine while driving, its just got a slightly rough idle, but i am picky and want it to be fixed.

i did a compression test before swapping heads and the results were

1) 140
2) 140
3) 130
4) 145
5) 160
6) 125

have not tested since replacing the head, but 3 and 6 seemed low. so maybe that's the issue?

any input/suggestions are welcome

 

i’m having the same problem with mine. just made an account to try to figure it out. i’ve changed just about every part i can think of. starting to drive me crazy lol

 

Test and clean the injectors.

 

Its funny I came across this....

I've owned my '98 XJ for.... I think over 10 years now. It always ran so-so at idle. But I have driven it over 65,000 miles like that. It makes plenty of power, starts properly, hauls stuff, etc. It runs like new at anything over idle speed. I've driven it across the country like this and it has never let me down.

I finally got the bug to start looking into what is wrong. I can hear a low cylinder contribution at idle out the exhaust. I can unplug injectors #1 and #6 and it doesn't change idle quality. I replaced the #6 injector with a NEW one and it yielded zero change. Vacuum leaks are not common on these things, and I tried inspecting and spraying stuff around the #1 and #6 intake runners and nothing changes. Spark plugs look good and consistent and I switched them with other cylinders just to be sure. Cap and rotor changes doesn't make any difference. It will throw misfire codes for #1 and #6 sometimes. Compression is in that 145 psi range on all 6.

Vehicle history is unknown before I got it. Interestingly, aren't cylinders #1 and #6 partners? That means they are at top dead center at the same time and they are using the same spot on the flywheel to let the computer know where it is. Ideally, I would recommend that you hook up an oscilloscope to the crankshaft position sensor and take a look at the wave form which I have been too lazy to do and most people do not own one or know what to do with one. Aftermarket crankshaft position sensors *are* known to cause these things to not run right and with the air gap between it and the crankshaft reluctor ring needing to be set in a difficult spot to get to, it wouldn't surprise me if that is your problem. If you still have your old sensor and it ran fine with that, swap it back in.

For all I know, the engine has been out of mine and someone bent the crankshaft reluctor ring and that is what is causing the problem with cylinders #1 and #6 when all the others seem to be hitting fine.

There are countless factors that can affect compression readings, so I wouldn't let those numbers dictate your next actions. And those numbers aren't necessarily that bad, contrary to what some other people will say.

 

Ever hook up noid lights to see if they are getting power? I'd look at the injector drivers, although they tend to either work or not.

Actually they are polar opposites. One is at TDC compression while the other would be at TDC exhaust.

As the camshaft position sensor is responsible for syncing the fuel injectors, I'd be looking there first.

Specifically, with the aforementioned oscilloscope.

That said, seems after all this time a P1391 code would have been thrown if it was either CkPS or CPS (yet, they are notorious for NOT throwing codes regardless how bad they are).

 

Experience tells me that noid lights really don't do much of anything other than put on a light show. I use a stethoscope to listen for injectors clicking and a low amp current probe to watch the injector pulse waveform if I need to dig in farther. I can watch the current ramp up and then I can see where the injector pintle physically begins to move, then stops, and then the current ramps up further and then shuts off and I can see the inductive kickback. That way, you can see if there are any shorted windings within the injector and the health of the circuit to and from each injector. My next investment is going to be a pressure transducer to hook to fuel rails so that I can watch the fuel pressure within the rail drop as each injector opens, and it is a decent way to determine how much they are or are not flowing relative to each other. I do not like throwing $60 fuel injectors at stuff anymore hoping to see a change. Or spraying tons of starting fluid all over hoping to find a vacuum leak.

https://images.app.goo.gl/WTAEXKx1pVr4Fb7FA
This indicates that #1 and #6 are partners.... at least with the terminology that I was taught. That means both are in the same position in the bores. Maybe some would consider partners to be adjacent to each other in the firing order?

The computer does seem to be satisfied with the signal being sent from the CKP and CMP and it has never thrown a code for either unless I disconnected them for a compression test. Fuel injector timing on these old things isn't super-critical, as a person can swap the electrical connector on adjacent cylinders and observe minimal, if any difference, in how it runs. I have a feeling that ignition timing is getting retarded by a flaw in the crank reluctor, but at higher RPM the magnetic field is generating a different waveform which translates into a differently timed firing event.

*sniff* *sniff* Do you smell that? Is that bull****? I'm always too busy working on other peoples' crap to actually get around to my own.

 

Sounds interesting.

What specific equipment are you using to do this?

 

Using scopes is a highly advantageous method of diagnosing stuff that is overlooked by almost everyone I know. But the key is knowing what is a good waveform for the particular vehicle and getting access to that is tricky without paying a monthly fee. It is also very useful for checking data busses to find out when things are talking to each other and when they are not or if there is too much EMI interfering on it to get a clean signal.

Currently, I am using a P.O.S. OTC 2 channel oscilloscope which came "free" with my OTC Genisys scan tool. (By the way, I'll never buy another OTC scan tool as mine never failed to disappoint. Lets just say that many times, it made me feel as though Owatonna Tool Company was at risk due to me living relatively close by). This scope is inferior with its LCD screen which doesn't work well when it is cold but otherwise does the job. It is also older and electronics always gets nicer and cheaper as time goes by. There are MUCH better ones out on the market today and I would recommend getting one of those, such as the Autel Maxiscope or the Pico brand stuff. I have yet to make the investment due to the relatively low volume of work I do.

For a low amp current probe, These things are fairly universal and will interface with any shop oscilloscope as far as I know. On one of these old Jeeps, all you do is hook the probe around the +12v wire (and only that one) going to the injectors and then I would use my inductive pickup probe around a spark plug wire on the other channel. This way, you see the injectors getting fired sequentially on the screen, and then the inductive pickup probe around a spark plug wire (#1 for example) to show where #1 injector is in the row of waveforms produced by all the injectors. Then, you can follow the firing order to determine which waveform belongs to which injector and pick out the one that looks goofy.
You can also use it to somewhat indirectly determine the compression on an engine. With more compression, it takes more KV to jump the spark plug gap, and with that, less "burn" time which is the duration of the spark. So, if you suspect a weak cylinder and the spark plugs are a PITA to get to and get an actual compression gauge in there, you can see the less KV necessary to jump the gap, and typically a longer burn time to go with it. With a coil-on-plug setup, you can measure the current of the primary side of the coil and compare it to other ignition coils and tell if one is not up to snuff. Simply ohming out the ignition coils is NOT necessarily indicative of its health. A single shorted winding within it will not greatly affect its measured resistance.

There's all sorts of ways to skin a cat. But the key is knowing what to look for. But the older I get and the crappier the aftermarket gets, loading up the parts cannon and shooting it just seems less and less appealing every day. I am hoping that when I finally get off my bum and away from my other projects, I will find that when the CKP reluctor hole for cylinder #1/6 goes by, the waveform will be all screwy, possibly due to a bent flywheel reluctor ring. Who knows. If someone else has been in there before me, all bets are off.

 

OK thanks, I already have a Hantek PC-based oscilloscope that I can use my 23 inch monitor with. Here's my CkPS vs CPS waveforms:



Rather, I was interested in the setup for a

I was looking at Ditex First Look but TTBOMK cannot measure fluids.

PDS500x has potential but that looks to be a little pricy.

Thanks.

 

They're just starting to ship PDS500x. $415 for the transducer. $620 for a complete kit.

For that kind of money I'll just get bullets for the Parts Cannon!

 

What do you use for a garage computer? Does the Hantek software and hardware work pretty well?

It has been awhile since I've had a computer set up in the garage. I am pretty sure it is because I do not heat my garage all the time that stuff on the hard drive would go corrupt and I would have to start over again. But now I suspect that an SSD drive would solve that issue. I am going to need to perform a mild remodel of the garage this winter to make room for more equipment and such. Get air hose reels installed, plumb in some air lines going various directions, upgrade insulation, upgrade some shelving, convert the furnace to electronic ignition, etc. And I just ordered a shipload of LED fixtures to replace the fluorescents that are out there, so I will have my work cut out for me.

 

Shooting the parts cannon still makes a lot of sense. For example, I do not diagnose oxygen sensors aside from a quick visual inspection of the wiring to and from it. If the computer says it is bad, it gets replaced. The amount of time and labor that I would have to charge to verify the operation of a 10+ year old, 150,000 mile oxygen sensor that costs $60 and takes about 10 minutes to replace just doesn't make a whole lot of sense.

But I also work on other stuff, and a good example that comes to mind is a Suzuki Eiger 4 wheeler that came in with no spark. It was running and then it just suddenly died and never ran again. The service manual said to check for AC voltage from the stator to the CDI box and if that was good, replace the CDI box which was $450. The AC voltage while cranking read perfectly according to the DVOM. I verified that the kill switch circuit wasn't grounding out and a few other items, so rather than roll the $450 dice, I checked with the oscilloscope. The waveform wasn't congruent. The magnets inside the flywheel that rotated around the stator came loose and shifted, and the two crank triggers didn't align correctly and it wasn't happy and wasn't going to run like that.

 

The software's on a laptop so I'm portable. Even took it on the boat to check ignition waveforms.

The Hantek works well but requires some fiddlement. Although it's got 8 channels, with the stuff I've been looking at I've only been able to get reliably 3 channels at most and then it starts aliasing like a MF.

In this shot, simply putting in a 3rd channel made the CPS channels (I put in 2 just to see when the tipping point was) start to alias:



But for the money I think it's great. After all, how many times are you really going to need 2 or three channels simultaneously. Besides, I've only got one attenuator and one current clamp so I can't put up a pile of data even if I wanted to.

 

On that I have 4 oxygen sensors but can read the waveforms from the OBD port. I've been running with an error code for like forever, but since it's a post-cat (although it participates in the fuel mix) and it only drifts occasionally over the pre-cat (to trigger the code) and also I'm cheap I just let the code run.

 

It is nice to finally come across someone who finds the usefulness in a scope.
I always find myself running 2 channels simultaneously and a 3rd may have been helpful, but things never needed to progress that far. The refresh rate of most scan tools is sufficient for o2 sensor monitoring.

Have you attempted the trick of using a spark plug non-fouler to install a downstream oxygen sensor when the catalyst bites the dust? The theory is that it pulls the sensor out of the exhaust stream and into the equivalent to a river's backwater which slows down the sensor's response time so to speak. Success probably depends on the vehicle's diagnostic strategy. I tried it in my Suzuki which uses a wideband upstream sensor, and thus a completely different diagnostic strategy. But for whatever reason, I started getting downstream sensor voltage readings of well over 1 volt and I am not sure why that happens. I wouldn't care about the dead catalytic converter, except that when the engine light comes on, the cruise control quits working. So, I just use a cheap Amazon bluetooth code reader linked to my android head unit to clear the codes when I go on a road trip somewhere or otherwise want to use the cruise control. I suspect that the non-fouler looking to be some type of aluminum is causing some kind of chemical reaction and changing its ground potential or some crap like that. Otherwise, I am not fond of spending $350+ on an aftermarket converter that probably doesn't work or $1,200+ from Suzuki that does work.