Rich mixture at idle - SOLVED!!!

[Tally ZJ]

For the past few, several actually, months my 1996 ZJ 4.0L has been choking me to death with exhaust smell. It is really bad when I pull to a stop sign or traffic light. The fumes just roll into the cabin. Standing at the rear of the Jeep at the tail pipe will cause your eyes to burn with the noxious smell/fumes. I have had others check it and they too say it is a horrible rich smell.

I have no malfunction indicators lit, no stored codes, no pending codes. It matters not whether the climate is hot or cold, humid or dry. It matters not whether the engine is hot or cold. Frequently, when first cranked after sitting a while, the engine idle seems to chug for a couple seconds before coming up to proper idle. Many times it fires right up and comes to a slightly high idle then drops to proper idle RPM.

The problem seems to be only at idle and off-throttle situations. Runs great, has good power. Fuel mileage has dropped a little bit.

I have a scanner that reads live data but I am not sure what each parameter should actually be at operating temp.



I have done the following in efforts to correct this:

- New plugs, wires, cap, rotor, cam sensor and MSD coil.

- New coolant temp sensor, new TPS, new crank sensor, new MAP sensor, new IAT sensor, new IAC motor, new O2 sensors up and down stream. Clean air filter, several different brands and octane fuel, several additives were tried.

- Exhaust system replaced and all seams were sealed with flame and filler metal. I have no exhaust leaks.

- I do not have a cat converter but I have not had one for years and this issue is a recent development.

- I had over heating issues, particularly with the A/C on at idle, but a new severe duty Max Cool fan clutch has eliminated any cooling issues. I now have temps that never even reach a full 210*. Still, my rich idle issue was present prior to this fan clutch install.




I do not mind doing the footwork but I surely could use some help as to what I should be looking for and what the numbers should look like.


UPDATE: Problem is solved! Read entire thread for details

[BigDaveZJ]

You say you have no cat, but what are you doing to trick the downstream O2 sensor in to thinking you have one? You should get a CEL from that alone.

[Tally ZJ]

The downstream is mounted in an O2 sensor bung welded into the exhaust pipe where the cat used to be. Have never had a CEL due to this. It has been that way for quite a while now with no apparent issues.

I removed the cat a few years ago when it came apart. The shop that then installed the new pipe told me a double stacked O2 sensor bung will eliminate the CEL I might get from the downstream O2 sensor reading. All had been well from that point up to a few months ago.

I asked about this again when I had my exhaust replaced, this week, and the shop owner, knowledgeable but no Einstein, told me it should produce no ill effects on my model year.


I first noticed these exhaust fumes when my exhaust manifold developed a crack. I replaced that last winter. Then my exhaust system started having leaks so I chopped it short. That produced the fumes again so I had the entire system replaced from the down pipe back to the tail pipe.

[Tally ZJ]

I also thought I should mention this as well. We have the Snap-On Verus scanner system at work but it will not communicate with my PCM. I can get a Generic Direct OBD Scan connection but it does not give as much data as the other method and only provees slightly more data than my Matco Scan Tool. Could this point to an actual PCM issue?

If there are some data numbers I can provide that will help I will write them down and post them tomorrow.

[ATL ZJ]

How old are your injectors and have they been leak tested?

Do you have the FPR on the fuel rail or is it built into the pump (sorry not as familiar with 96 4.0s as much as 93-95). How old is your FPR and have you tested it?

Any blue smoke at startup or while idling?

If you hold the throttle wide open for a significant time does the smell improve? EDIT re-reading your post sounds like it does improve then.

[ATL ZJ]

I also thought I should mention this as well. We have the Snap-On Verus scanner system at work but it will not communicate with my PCM. I can get a Generic Direct OBD Scan connection but it does not give as much data as the other method and only provees slightly more data than my Matco Scan Tool. Could this point to an actual PCM issue?

If there are some data numbers I can provide that will help I will write them down and post them tomorrow.

Rule #1 it is never the PCM. I do not mean this 100% literally but there is a lot of truth to it and the times I have forgotten it I have wasted money.

[Tally ZJ]

How old are your injectors and have they been leak tested?

Do you have the FPR on the fuel rail or is it built into the pump (sorry not as familiar with 96 4.0s as much as 93-95). How old is your FPR and have you tested it?

Any blue smoke at startup or while idling?

If you hold the throttle wide open for a significant time does the smell improve?

I had been running the Neon injectors for a year or more and thought they could be contributors so I went back to my factory injectors which worked very well before. They have not been leak tested though.

My FPR is in the tank with the fuel pump assembly. It is a few years old now as I replaced it when I got the Jeep. I will run a fuel pressure test soon and check that.

If I hold the throttle open the smell does improve. It improves even at moderate throttle opening. No smoke at start up that I have noticed. It does puff some darker grey to black smoke if I just punch the throttle open and I have noticed the same (from the rear view mirror) if I lay into the throttle hard in low gear and accelerate hard. Just a puff though and then its gone. Nothing heavy or constant as to smoke.

[ATL ZJ]

If the problem is consistent with two different sets of injectors I think it is unlikely that both sets are equally leaky.

Although you have already replaced most of these sensors that send these type of inputs to the PCM, something could be causing your injectors to stay open too long and dump too much fuel. If this is already obvious to you pardon the oversimplification but I would then focus on inputs that affect the PCM's decisions about injector pulse width.

Is your TB clean and I assume you have no vac leaks?

EDIT: shame on my reading comprehension. you said you replaced your camshaft position sensor already.

[Tally ZJ]

If the problem is consistent with two different sets of injectors I think it is unlikely that both sets are equally leaky.

It sounds like something could be causing your injectors to stay open too long and dump too much fuel.

Is your TB clean and I assume you have no vac leaks?

Have you ever replaced your camshaft position sensor?

Yes, I have noticed no difference between the two sets of injectors. Throttle body is and has been cleaned completely (removed and cleaned). Have not checked for a vacuum leak but I have seen no usual indications that I have one. I will check that though.

Cam sensor was replaced two months ago when the cap and rotor was done.

[Tally ZJ]

Although you have already replaced most of these sensors that send these type of inputs to the PCM, something could be causing your injectors to stay open too long and dump too much fuel.

And this is my thought as well but I have limited knowledge and I am unsure exactly which one, or a combination of them, will cause the injectors to do this at idle. This is why I have refreshed every sensor on it in the past month except the O2 sensors which were replaced three months ago with NTK brand sensors (not Bosch).

[SB406]

- I had over heating issues, particularly with the A/C on at idle, but a new severe duty Max Cool fan clutch has eliminated any cooling issues. I now have temps that never even reach a full 210*. Still, my rich idle issue was present prior to this fan clutch install.


.

This struck me as odd, especially since your problem is at idle. If you can connect with the scanner, I'd watch the Coolant Temp sensor reading during warm up. I know you said you changed it. On a Jeep are there 2 temp sensors (PCM & gauge) or 1 sensor that pulls double duty?

The only other thing off the top of my head is to check that the fuel tank is not pressurizing. That will affect your fuel pressure.

[Tally ZJ]

This struck me as odd, especially since your problem is at idle. If you can connect with the scanner, I'd watch the Coolant Temp sensor reading during warm up. I know you said you changed it. On a Jeep are there 2 temp sensors (PCM & gauge) or 1 sensor that pulls double duty?

The only other thing off the top of my head is to check that the fuel tank is not pressurizing. That will affect your fuel pressure.

On the later models, as in mine, their is only one temp sensor. From that sensor input, the PCM controls the gauge reading.

[Tally ZJ]

Today I ordered a direct replacement cat converter. I want that peace of mind that at least my post cat O2 sensor will be sniffing what it should be sniffing. That will be in place next week.



I took live data readings with my scanner this morning at start up and again when I arrived at work with a hot engine.

Here is the results of the data I can get from that scanner:


Cold Engine Readings (key on - not running):

Engine Coolant Temp 81*
MAP (inHg) 9.4
Intake Air Temp 84*
Throttle Position (%) 14.5
O2s B1 S1 (v) 1.000
O2s B1 S2 (v) 1.000
Short Term Fuel Trim B1 S2 (%) 99.2

Cold Engine (Running):

Calc Load (%) 6.7
Short Term Fuel Trim B1 -7.0 to -3.9
Long Term Fuel Trim B1 -7.0
MAP (in Hg) 3.8
Engine RPM 760
Spark Advance 17.0





Hot Engine (Running):

Fuel System 1 CL (closed loop)
Calc Load (%) 5.1
STFT B1 (%) -0.8
LTFT B1 (%) -7.0
MAP (inHg) 3.5
Engine RPM 770
Spark Advance 15.5 to 19.0
Intake Air Temp 162*
Engine Coolant Temp 196*
O2s B1 S1 (v) 0.260 to 0.800
O2s B1 S2 (v) 0.180
Short Term Fuel Trim B1 S2 (%) 99.2

[Tally ZJ]

More testing tonight:


I disconnected the battery cables, held them together for half minute and cleared the PCM adaptive memory.

Cranked it up again with the scanner connected. Pretty much got the same readings except that the LTFT numbers seemed to clear to zero. I guess that is to be expected.

Coolant temp comes up as it should. As the temp reaches 150* the fuel system goes to closed loop. Normal I think.

The only number I have noticed maybe odd, even before clearing the PCM tonight, is the post cat O2 sensor percent reading never changes. It constantly reads 99%.

Tonight, as it warmed up, I checked the tail pipe exhaust and it smells normal. Does not burn my eyes as usual. I will check this again tomorrow and see if it passes the smell test.

I checked everywhere for a vacuum leak and found none.

I disconnected the CCV tube from its intake manifold port and the idle jumped up to about 1600 RPM. Normal. No CEL showed up from that vacuum leak. Not sure if a quick and short leak would set a malfunction light. From my research the MIL will only light up if the LTFT numbers change drastically one way or the other attempting to adjust rich/lean conditions.

The scanner also showed me that the TPS voltage changes smoothly through its sweep from closed to open throttle. Timing advance changes from 15 to about 20 depending on throttle opening. The MAP shows a drop in the numbers as I open the throttle and climbs back up as I close throttle.

The voltage number for the B1 S1 O2 sensor changes quickly as I throttle up and down. I cannot remember if it goes from negative to positive or vice versa.

[ATL ZJ]

I think you are approaching this the right way. Good job being so disciplined and level headed.

Glad to hear no vac leaks. After further thought that could cause more of a lean condition but still good.

Have you tried running with the downstream o2 unplugged?

Do the upstream o2s exclusively affect whether it runs open/closed loop?

Very interesting case so keep posting up. Hopefully someone with more jeep specific obd2 knowledge than me will give guidance on the parameters you're seeing.

[albersondh]

Well nice to see Im not the only one with uber high IAT's... I see around 170* on a hot day and thought that was due to my home brew intake/filter. Time for a re-work and true CAI... All the data you posted above I can pull with my phone (android app torque). Its hard to believe you wouldnt get CEL running that fat in CL. Im sure it is fat as you describe but it would be nice to know the extent and a WBO2 hooked up would help. I can pull a graph from my front OE NB02 and watch the voltage swing at idle. I can log it and record the highs and lows for reference but not sure this would help you as mine is a V8, not sure the 02 output would range the same.

I wonder what that 99.2% ST fuel trim is meant to indicate? No way thats duty cycle I wonder what that is referencing?

Can you log IDC (injector duty cycle) at idle in CL operation? This would at least give you and idea of what your ECU "thinks" is normal. I see from 2-4%IDC at idle.

Is your tool indicating corrected MAP? I would think its corrected and not a raw value (+ atmosphere) because you indicate it as inHg? 3.5 inHg is very low (vacuum leak). Normal idle should be around 10 inHg (gauge), but thats corrected. Your showing 9.4 inHg not running, assuming sea level (Florida) 14.7psi, this is not possible. I can select from "boost" (corrected) and it shows 0, key on, not running and around -10 at idle. When I select "vaccum" it displays inHg and shows around 1 inHg, key on, not running and 18-20 during warm up and hot running at idle. Regardless, the number should go from lower to higher not-running vs running assuming actual vacuum. If I was in your situation I would chuck a mechanical vacuum gauge on it. Look for around 18-20 inHg at idle much lower and I would suspect a leak.

[SB406]

Is your tool indicating corrected MAP? I would think its corrected and not a raw value (+ atmosphere) because you indicate it as inHg? 3.5 inHg is very low (vacuum leak). Normal idle should be around 10 inHg (gauge), but thats corrected. Your showing 9.4 inHg not running, assuming sea level (Florida) 14.7psi, this is not possible. I can select from "boost" (corrected) and it shows 0, key on, not running and around -10 at idle. When I select "vaccum" it displays inHg and shows around 1 inHg, key on, not running and 18-20 during warm up and hot running at idle. Regardless, the number should go from lower to higher not-running vs running assuming actual vacuum. If I was in your situation I would chuck a mechanical vacuum gauge on it. Look for around 18-20 inHg at idle much lower and I would suspect a leak.

The 9.4" Hg with engine off would make me think bad MAP sensor.
But....
I'm used to looking at MAP readings in KPa. You should see around 100 KPa not running (atmospheric pressure), which is 29.5"Hg. Perhaps a typo??
On a Chevy V8, I've seen anywhere from 30 to 50 KPa (8.9 - 14.8" Hg) at idle, depending on the cam. Which way it moves from off idle depends on the cam, but at WOT, it approaches 100 KPa (Throttle wide open; intake pressure = atmospheric pressure)

I agree and disagree with albersondh. 3.5" Hg is a very low manifold pressure (12 KPa). However, I see that as a heavy vacuum, not a vacuum leak, considering 29.5" Hg is atmosphere.

I'd maybe try a different MAP sensor, and check the port it's connected to.

[albersondh]

MAP readings look wonky. Could be inHg, HKPA, KPA, corrected/gauge, or not, whatever still looks wrong. I wouldnt assume a fault and throw a part at it. Grab a known value mechanical gauge, verify.

[Tally ZJ]

I wonder what that 99.2% ST fuel trim is meant to indicate? No way thats duty cycle I wonder what that is referencing?

I think the 99% is the catalytic efficiency?

Can you log IDC (injector duty cycle) at idle in CL operation? This would at least give you and idea of what your ECU "thinks" is normal. I see from 2-4%IDC at idle.

My little scanner will not do injector cycle. The Verus system at work will do it but for some reason it will not communicate with my PCM.

Is your tool indicating corrected MAP? I would think its corrected and not a raw value (+ atmosphere) because you indicate it as inHg? 3.5 inHg is very low (vacuum leak). Normal idle should be around 10 inHg (gauge), but thats corrected. Your showing 9.4 inHg not running, assuming sea level (Florida) 14.7psi, this is not possible. I can select from "boost" (corrected) and it shows 0, key on, not running and around -10 at idle. When I select "vaccum" it displays inHg and shows around 1 inHg, key on, not running and 18-20 during warm up and hot running at idle. Regardless, the number should go from lower to higher not-running vs running assuming actual vacuum. If I was in your situation I would chuck a mechanical vacuum gauge on it. Look for around 18-20 inHg at idle much lower and I would suspect a leak.

I can only post the information as my scanner shows me. It gives the value in inHg only. The number goes down as I accelerate. I will double check it again.

The 9.4" Hg with engine off would make me think bad MAP sensor.
But....
I'm used to looking at MAP readings in KPa. You should see around 100 KPa not running (atmospheric pressure), which is 29.5"Hg. Perhaps a typo??
On a Chevy V8, I've seen anywhere from 30 to 50 KPa (8.9 - 14.8" Hg) at idle, depending on the cam. Which way it moves from off idle depends on the cam, but at WOT, it approaches 100 KPa (Throttle wide open; intake pressure = atmospheric pressure)

I guess it is possible but my old MAP sensor gave the same readings as my brand new MAP sensor. I checked the rubber elbow that connects it to the throttle body and it is in good condition.

I agree and disagree with albersondh. 3.5" Hg is a very low manifold pressure (12 KPa). However, I see that as a heavy vacuum, not a vacuum leak, considering 29.5" Hg is atmosphere.

I'd maybe try a different MAP sensor, and check the port it's connected to.

Well, the MAP sensor is brand new but I still have my old one. Should I swap them out again? I cannot afford to buy another one. They are almost $100 at my cost.

MAP readings look wonky. Could be inHg, HKPA, KPA, corrected/gauge, or not, whatever still looks wrong. I wouldnt assume a fault and throw a part at it. Grab a known value mechanical gauge, verify.

Will try to get this done. Not sure if I know anyone with a gauge.

[Tally ZJ]

MAP sensor absolutely sits at 9.5 inHg at engine off/key on. When cranked, it immediately shows 3.5 inHg and when accelerated to WOT is drops to as low as 1.4 inHg.

[Tally ZJ]

Just some quick research on MAP sensors. They do not read manifold vacuum, they read manifold pressure, or a lack thereof.

Supposedly, to get vacuum numbers, you subtract your MAP reading at idle (in inHg) from the current Barometric pressure of your locale.

My local baro pressure is 29.94 in. 29.94 - 3.5 = 26.44in.

Does anyone make sense of this?

At key on/engine off, that means my baro pressure would be 10.44, which as stated already, is impossible.

[albersondh]

Just some quick research on MAP sensors. They do not read manifold vacuum, they read manifold pressure, or a lack thereof.

Supposedly, to get vacuum numbers, you subtract your MAP reading at idle (in inHg) from the current Barometric pressure of your locale.

My local baro pressure is 29.94 in. 29.94 - 3.5 = 26.44in.

Does anyone make sense of this?

At key on/engine off, that means my baro pressure would be 10.44, which as stated already, is impossible.

It would be very unlikley if not impossible that you are truley at 26.44inHg at idle, like SB406 said that is very high. If you have an aftermarket cam that would be even more unlikley as idle vacuum is lower due to overlap, SB406 points this out when refering to SBC stuff. Your numbers drop as you increase throttle which is correct. For your application a cheapo vacuum gauge will tell the story unless someone else can verify your numbers. Seems to be far enough away from logical regardless of how you calculate it, I would want a known reference. If you have to spring for a vacuum gauge its not a complete loss. Vacuum gauge and IR gun are IMO the most underutilized troubleshooting tools available, and they are cheap. Fun with vacuum gauge http://classicinlines.com/Vacuum.asp.


Knowing your actual IDC would be helpfull because without a CEL this is actualy what your ECU believes to be correct. If you know all your sensors are good but your ECU is still telling dirty lies, then I would begin to suspect an ECU problem. But like ATL pointed out its everyones knee jerk part and almost never the actual problem.

[Tally ZJ]

It would be very unlikley if not impossible that you are truley at 26.44inHg at idle, like SB406 said that is very high.

Idle?!? Heck, that is with the engine off and key on. At idle it would be even lower. Shouldn't it be at actual 29.94 as my local baro pressure reads?

Knowing your actual IDC would be helpfull because without a CEL this is actualy what your ECU believes to be correct.

Sorry but what is IDC?

[Tally ZJ]

I removed my MAP sensor from the throttle body and attached a hose to the port. Sucking a vacuum and reading the scanner as I do this will bring the number only as low as 2.4 inHg. Putting as much pressure as possible will only take the numbers to 9.7 inHg max. When I release pressure it goes back to 9.5 in Hg and sits steady at that number.

Of course, engine off and key on. Throttle body is removed at this time.

So, now I may suspect that my scanner gives numbers in inHG but is it possible that the MAP does not use the inHg increment? Could my scanner reading be giving me the wrong data?

I also plugged in my old MAP sensor and got exactly the same reading.

[albersondh]

IDC=Injector Duty Cycle.

MAP sensors are a bitch. I run 2 on my fun car. I had to upgrade to a 4-bar unit because I was boosting past the readable range of the OE 2-bar unit. Now the old stock 2-bar reads crank-case pressure (blow-by) for me and vacuum when my crank case pump kicks in. Scaling both of these sensors from their analog 0-5v range into something I could actualy understand (inHGG and PSIG) was a nightmare (differential equations). By far the hardest piece of tunning work I have had to do to the car. I cant claim to understand what your tool is actually telling you, KPA, HKPA, inHG, corrected, uncorrected..... When in doubt, vacuum gague.

Both sensors giving the same reading on your tool makes me suspecious. Dont get to invested in the MAP thing, verify, via gauge or someone who knows acceptable range for your application on the tool, and move on.

Your going to know way more than you ever wanted to by the time your done. Hope this is not your prime mover and you can have fun with the process. Keep at it man!