Narrowband O2 signal drop out - Chrome ECU v2

[Dougal]

Well my first thought was it was really going that lean, but according to your wideband it's going pig rich... If that's correct, it would indicate the narrowband is lying...

Yes my wideband shows rich with 12 and 11 AFRs during accel.

This is what has me a bit confused, I have a feeling perhaps the MAF has something to do with this?

I dont have a working stock MAF as it was damaged beyond repair.
I got the 482 Montero MAF to use for when i had time to install the 19T's but I since ended up using it because the stock one was broken.

[striker2]

Im wondering if there is an issue internally on some of the clone ECUs (like a bad solder joint or something) that is causing the rear O2 sensor to drop out occasionally. I say this only because I have never seen this happen with the multiple true 98/99 ECUs I have messed with and you both have Jester ECUs. I'm not saying this is definitely the case (I'm not even sure if the rear O2 sensor circuit is one that has to be modified) but if everything else checks out externally then I would be inclined to lean to an internal problem with the ECU.

The MAF probably doesnt have anything to do with this, but you can test it by unplugging the MAF and see if the O2 sensor still drops out (I'm betting it does).

[RealMcCoy]

Yes my wideband shows rich with 12 and 11 AFRs during accel.

This is what has me a bit confused, I have a feeling perhaps the MAF has something to do with this?

I dont have a working stock MAF as it was damaged beyond repair.
I got the 482 Montero MAF to use for when i had time to install the 19T's but I since ended up using it because the stock one was broken.

The narrowband says it's lean, and the wideband says it's rich... One of them is wrong. The ECU believes the narrowband and reacts accordingly. The primary mystery you need to solve before you worry about any other tuning, is why two sensors in the same exhaust stream are reporting completely opposite conditions... Do you have a front O2 boss you could switch to and see if both banks behave the same?

[Dougal]

The narrowband says it's lean, and the wideband says it's rich... One of them is wrong. The ECU believes the narrowband and reacts accordingly. The primary mystery you need to solve before you worry about any other tuning, is why two sensors in the same exhaust stream are reporting completely opposite conditions... Do you have a front O2 boss you could switch to and see if both banks behave the same?

I am pretty certain that the WB 02 is correct. Its an AEM UEGO.
However I can always plug the serial interface into my notebook and log to hyperterminal and see if the measurements are the same for what is through the wired ECU pin scaled in ECUFlash versus the Serial interface.
If its different then I know I have an issue with the WB 02 sensor or the scaling.

I do have provision for the front bank o2 as I am using a modified front exhaust pipe for an NA 3000gt.
The only issue is that I sits right next to the front engine mount leaving very very little room to actually install it.
Worth another look though as I could use the spare Bosch o2 sensor I have and run it to the ecu to see if there is any difference in signal quality.

A lot of sensors use the same ecu grounding. (MAF, TPS, Coolant temp, Narrowband o2)
Could the potential for readings being out of spec be related to this shared sensor ground back feeding into the signal wires of the sensors.

Is there a way around this where they can be divided so they are not sharing the same ground and potentially causing signal issues?

Thanks for everyone input.
This is really helping isolate the true cause of the issues. Hopefully helps other people with similar issues

[Dougal]

Im wondering if there is an issue internally on some of the clone ECUs (like a bad solder joint or something) that is causing the rear O2 sensor to drop out occasionally. I say this only because I have never seen this happen with the multiple true 98/99 ECUs I have messed with and you both have Jester ECUs. I'm not saying this is definitely the case (I'm not even sure if the rear O2 sensor circuit is one that has to be modified) but if everything else checks out externally then I would be inclined to lean to an internal problem with the ECU.

The MAF probably doesnt have anything to do with this, but you can test it by unplugging the MAF and see if the O2 sensor still drops out (I'm betting it does).

I will run the car today and unplug the MAF and report back what happens.
If it does not drop out the o2 signal and shows a very rich o2 signal then that helps tell us that the car is simply running super lean and I could tune more fuel into the airflow load range (when MAF plugged in normal conditions).

Looks like I have some things to try.

Will report back ASAP

[Dougal]

OK so i have been busy today at testing my o2 sensor stuff.

Started with a simple test of unplugging the MAF and seeing what happens to the o2 sensor signal.

From closed loop o2 cycling when the MAF is unplugged the o2 signal is steady and progressively get richer to a point (.60 volts)
Obviously STFT stops as the ecu knows the MAF is unplugged so its using a preset for MAF fault/Limp mode trim.

At idle its nice and rich but given any load then o2 sensor goes to zero.
obviously the absence of an airflow reading to the ecu.
O2 sensor drop out MAF unplugged.jpg

So based of the o2 sensor dropping out with the MAF unplugged and what i think is it behaving badly i went about adding a heap more fuel into my VE tables to see if i could raise the o2 sensor off zero when load applied.
I was a little aggressive in the addition of fuel requirements.
I guess i wanted to see how much fuel was needed to get my STFT trim to become a negative where the ecu would pull back fuel rather than always be adding fuel.

I am unable to drive this car to properly perform full load testing however for the purpose of the o2 test i figured stationary testing would suffice.

All up i think that either my setup with the Montero MAF requires me to modify the MAF Smoothing table quite a lot.
If its not that then i have a seriously MASSIVE intake leak (or injector issue?) that is completely screwing up my fueling.
Or a sensor ground wiring issue?

Here is what i did.
1. Increased addition of fuel via MAF smoothing VE table only the Idle airflow hz frequencies. (25,50 and 75 hz) and logged the results.
I did this three times for idle airflow range. (First with 104% then at 106% and then at 108%)
Noticed that by doing this my o2 sensor still cycled the same but my STFT got closer to zero
Eventually finding that it sat nearly on Zero STFT when idling.
See below extracts from Evoscan showing my o2 cycling and STFT graphed.


VE table comparison. Saved a new bin file each time i changed.
VE Table adjustments for o2 sensor testing.jpg

First log of Idle closed loop o2 signal at 104% (25-75hz)
VE 104 idle hz richer idle o2 signal.jpg
2nd log of Idle closed loop o2 signal at 106% (25-75hz)
VE 106 idle hz richer idle o2 signal.jpg
3rd log if Idle closed loop o2 signal at 108% (25-75hz)
VE 108 idle hz richer idle o2 signal.jpg

See next post for other screen shots

[Dougal]

2. Increased addition of fuel in MAF smoothing VE table for airflow load(100 hz to 200 hz)
I also did this twice for this airflow range. (First with 105% and 107% and then with 107% and 109%)

O2 sensor would still zero out at points when applied load.
No real clear reason that i can see based of the logs.
For no rhyme or reason it seems to drop out and WB shows lean at that point.
Sometimes it comes back and sometimes it does not.

Not nearly as much as it used to.
Also found that it would zero out more if it was a short burst of increased airflow where as if i progressively increased load it would show o2 signals still cycling
STFT trim was also not adding a heap of fuel into the mixture. Trims would stick around what they were whilst in closed loop.
Below shows all the screen shots of this also.

Were times on this that it still jumped to adding 12.5% fuel. Wideband was showing rich AFR at these points.

VE set at 105% and 107% for airflow range between 100hz and 200hz
VE 108 idle and VE 107 load under 200hz richer idle o2 signal.jpg

VE set at 107% and 109% for airflow range between 100hz and 200hz
VE 108 idle and VE 109 load under 200hz richer idle o2 signal.jpg

Datalog holding o2 signal
Datalog o2 sensor 100-200hz.jpg

Datalog dropping out o2 sensor signal.
Datalog o2 sensor drop out 100-200hz.jpg


Does it seem correct that i would have to add this much fuel to the VE table to get the o2 sensor behaving a little better or am i just covering up a different problem with more fuel?

My trims seem to have adjusted better and my WB o2 seems to be showing correct AFR closer to what is in the fuel tables.

Thanks. Input from you guys is appreciated.

[Dougal]

Been reading up on o2 sensor faults and diagnosis.

Going to pull the o2 sensor out of the exhaust and check to make sure its intact and that its not oil soaked.
Perhaps i have an oil leak from the turbos and its what is causing my o2 sensor drop out? (Dont have any odd colour exhaust fumes though)

Also going to check resistance of the sensor and then of the wiring back to the ECM.

Will check heater wire voltage also while im at it.

Im going to check logged voltage of the o2 versus voltage at the sensor wiring plug to also confirm what is coming from the sensor is what is read back at the ECM.

Might figure out whats happening by these tests. Hoping its something obvious but easy to fix.

[Dougal]

Happy new year everyone!

Ok so good news is that the sensor is not oil soaked and the pipes are clean also.

Resistance on the wiring is within normal spec. (at most one of the wires was .07 ohms with the others about .05 ohms)

Just in case the wiring and or the sensor were rooted i decided to swap it with the other sensor i have (same Bosch type)

Everything is now looking alot better.

I have o2 sensor signal between 0.86 volts to 0.11 volts for normal idle fluctuations

I still have my tune with significantly richer VE % in the MAF smoothing table.
This is think isnt so bad for now but as i really start to fine tune it i might find a leak or two thats part of the reason it was so lean.
The only other conclusion i have is that its because i have a Montero 482 maf.

Anyway here is some data extract showing the o2 signal staying strong and rich when load applied.

The real test of this will be going for a drive. Which im planning to do soon.

o2 sensor working at load.jpg
o2 sensor working at load excel.jpg

[Dougal]

Took car for drive.

o2 signal is fine now. Holds steady throughout driving.

Car has a really nasty knock (9 counts) but no knock voltage except for a small blip here and there (0.058 volts).

The only way i have been able to quell it was by dropping the target engine load to 100 so the car isnt trying to boost over wastegate spring pressure.

I ran out of time but was going to pull some ignition timing advance in the knock area.

Seems to knock at 3000-3500rpm at ecuload 100 or over. Boost spikes at that point also.

Knock voltage is zero and only randomly shows any signal all very low signal anyway.

The knock its recording is quite severe and car basically goes into limp mode.
Only way to run it is to shut it off and wait for the power to cease to the ecu.

If try and drive the car with it then the compressors surge and hoot like an owl.
This is the same issue ive had had from when i started with tuning the Chrome ecu.
Also same issue as posted here
http://www.3sgto.org/threads/16039-C...de-at-3000-rpm

I am going to go back to check knock sensor circuit and also try running my stock 92 GTO ecu.
Not sure that will work either considering Montero MAF, full vac delete done, egr delete and all the other peripheries are on and off to suit my setup with the Chrome ecu.



knock log.jpg

Knock logged graph area.jpg