Smog readiness test showing pass then incomplete

[U8MYDZT]

Ok before I get blasted I have been searching forever and I am only asking because I have exhausted my search capabilities...

I have been trying to get my TJ to pass Cal state emissions for a while now. I have 3 drive cycle test that will not stay completed. My O2 test, O2 heater test and Cat test. Both O2 test keep showing complete and then a couple days later the show incomplete. I can not get them to stay complete long enough for the Cat test to pass. Any idea what could cause this?

History:
The TJ rolled a couple years back and sat for about 2 years. After finally getting it going again it failed smog. I replaced the head, cracked exhaust manifold, all vacuum lines, ECU, both O2 sensors (with walkers) and checked out the charging system.

I am also getting a strange symptom occasionally... When I turn my key to the run position I get rappid clicking from the fuse box area behind the glove box, the stereo won't come on and I don't get the 4 beeps right after you turn the key. This doesn't happen everytime and never causes issues with the Jeep running.

Any feedback would be greatly appreciated. This is driving me crazy.

 

[GoldenGorilla]

I am also getting a strange symptom occasionally... When I turn my key to the run position I get rappid clicking from the fuse box area behind the glove box, the stereo won't come on and I don't get the 4 beeps right after you turn the key. This doesn't happen everytime and never causes issues with the Jeep running.

This sounds like a battery dying of old age, which can cause all sorts of strange issues. What's the date code on it?

If the system voltage is dropping too low in those first x seconds when the O2 sensors are warming up, they might not register the "correct voltage drop" in the alloted time.

 

[U8MYDZT]

The battery isn't new, but it has been load tested and passed.

 

[macleanflood]

Got any way to monitor live ODB data? A Bluetooth ODB dongle and the free app Torque Lite would be great.

-Mac

 

[U8MYDZT]

I have a Bluetooth obd2 reader and the torque app. I can see the live data from most sensors.

 

[U8MYDZT]

Please anybody?

 

[GoldenGorilla]

History:
The TJ rolled a couple years back and sat for about 2 years. After finally getting it going again it failed smog. I replaced the head, cracked exhaust manifold, all vacuum lines, ECU, both O2 sensors (with walkers) and checked out the charging system.

I do not know who actually makes the Walker O2 sensors, but the TJs are VERY picky about O2 sensor. Chrysler are cheap bastards and decided to use 3 wires instead of 4 from the O2 sensor to the PCM. Instead of directly measuring the O2 sensor heater is active, they measure the voltage drop across the sensor signal and this has to happen in a specific time window. If it is too fast or too slow, it will set the O2 sensor in not-ready status. This is why Bosch sensors don't work reliably in a TJ, but NTK/NGK do.

Check your sensors against this thread

 

[U8MYDZT]

I got the recommended O2 sensors from that thread.

 

[macleanflood]

I got recommend O2s from the thread. On my 97 I tried a new NGK, a new Walker. None worked correctly. Both had voltages that trended above .9v.

A ~10 year old NGK from a rusted out 2000 Sahara I was parting out fixed my issue.

This is one of the many reasons why I'm pursuing replacing my PCM with a Haltech. I can use a quality wideband O2 sensor instead of these crappy three wires.

-Mac

 

[U8MYDZT]

My sensor are pretty much always in this range. And they physically have 4 wires.

 

[GoldenGorilla]

I got the recommended O2 sensors from that thread.

Sorry, forgot Walker was sourcing the same. As I understand it NTK are made in Japan as well, some were assembled in USA.

I'm trying to find a video that explains how the PCM tests the voltage drop as the O2 heater kicks on. Someone provided it to me when I was trying to figure out if I could feed the narrowband signal from my PLX wideband to the PCM. It's really the PCM being picky about this or a shorting fault.

This is an OK diagnostic step by step.


View: https://youtu.be/jlcTx1pIF1E?si=rEASVI9n2_cSNUQm

There are a couple things that could be going wrong, none of them especially easy to diagnose. The heater power may be intermittently shorting to ground. The 5V or signal may be intermittently shorting ground. The heater in your new sensor may not be working to spec.

 

[U8MYDZT]

Sorry, forgot Walker was sourcing the same. As I understand it NTK are made in Japan as well, some were assembled in USA.

I'm trying to find a video that explains how the PCM tests the voltage drop as the O2 heater kicks on. Someone provided it to me when I was trying to figure out if I could feed the narrowband signal from my PLX wideband to the PCM. It's really the PCM being picky about this or a shorting fault.

This is an OK diagnostic step by step.


View: %5BURL%5D%5BURL%5Dhttps://youtu.be/jlcTx1pIF1E?si=rEASVI9n2_cSNUQm[/URL][/URL]

There are a couple things that could be going wrong, none of them especially easy to diagnose. The heater power may be intermittently shorting to ground. The 5V or signal may be intermittently shorting ground. The heater in your new sensor may not be working to spec.

I saw a video about how the heater works. This is what got me to replace the sensors in the first place. The old ones had the same behavior. I am leaning towards an electrical gremlin. The 5V was dropping out before the ECU replacement (Wrangler Fix) causing the Jeep to randomly die. Now the 5V is pretty stable. I checked the harness all the way to both sensors. I replaced the front plug because it was melted from laying on the exhaust at some point. But again replacing the plug made no difference. It passed smog many times on the old melted plug.

 

[GoldenGorilla]

View attachment 695991
My sensor are pretty much always in this range. And they physically have 4 wires.

Understand the aftermarket sensor has 4 wires. The wiring harness to the PCM probably only has 3. I don't know the '97 specifics.

Error - this was from a file when my pre-cat died - correction below

The downstream looks off. At warmup the activity is lower.

Once the cat is fully warmed up it will mirror closely the upstream activity.

The expected voltage is .450 and with a narrowband higher than .450 indicates rich, lower indicates lean. On cat warmup the lower sensor will show rich in my experience.

 

[U8MYDZT]

Understand the aftermarket sensor has 4 wires. The wiring harness to the PCM probably only has 3. I don't know the '97 specifics.

The downstream looks off. At warmup the activity is lower.
View attachment 696001

Once the cat is fully warmed up it will mirror closely the upstream activity.
View attachment 696004

I thought it was the other way around. When the cat is cold the downstream will follow the upstream because the cat is not performing. Once warmed up the signal from the secondary O2 should be relatively stable and not swing with the primary O2 because the cat is doing its job and cleaning the exhaust gasses.

Also 97-98 had a different O2 setup than any other TJ I have a single upstream and a single downstream

 

[U8MYDZT]

View: https://youtu.be/KGpMzdEO5KY?si=mkLFCkK5vlWqWGpy

 

[GoldenGorilla]

Ooph, sorry. I was reading the wrong side on a file where one of my precats died

The warmed up activity can move out of that narrow range when you have a highly lean situation like a 0 throttle deceleration

But it is more steady when driving. Looking at the second chart group here. The STB1 element is defaulting to the wrong value here.

 

[U8MYDZT]

Thanks for the input! I will see if I can record the data from my sensors. I don't know how sophisticated my cheap reader is. My primary O2 doesn't always swing up and down that much. The secondary is pretty stable but will drop out on throttle lift.

 

[GoldenGorilla]

Thanks for the input! I will see if I can record the data from my sensors. I don't know how sophisticated my cheap reader is. My primary O2 doesn't always swing up and down that much. The secondary is pretty stable but will drop out when on throttle lift.

Then that sounds reasonable. If it's pegged high and never moves, then you have a bad O2 sensor.

When I use torque on my phone with a bluetooth ODBII scanner it looks like yours. I hope it's helpful to see the bigger picture.

I have the ability to scale in and out in HPTuners VCM Scanner. Unfortunately it doesn't display a time scale in the charts but upstream peak to peak or trough to trough are usually around 1 second interval. The sample intervals from various sensors are not consistent from the PCM either. The raw data export is chaotic with nulls, but the UI ignores them.

 

[GoldenGorilla]

Thought about your state a bit more and I think that's why I want that other video I'm having trouble finding. It had the procedure for backprobing the sensor on startup to monitor and time the voltage drop on start and confirm against the PCM expected time. I'm not sure if that test gets applied to downstream sensors (I think it is required). It may also narrow the potential sources of issue.

 

[U8MYDZT]

Hmm, that could be a good test. I thought I read somewhere that the meter itself can influence the reading and throw things off. I am sure there is a way to passively take a reading. If you do come across that video shoot it my way.