Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Kevin Wood]

IIRC, Taxi Al had a very similar issue to this.. on a Manual 3.2.

Can't remember what fixed it in the end.

Two coil packs, two engine looms and alot of head scratching

Alot of bits were swapped in the process, only significant item was a year old pattern MAF was replaced with the original one (which is still fitted)...

It was a very similar set of symptoms, though, if memory serves. Rag it up to 4K+ in 2nd and the light starts flashing and it drops a couple of cylinders?

[05omegav6]

The threshold was lower, about 3k iirc, but same behaviour.

An auto ecu didn't cure it, so there might be some smoke/mirrors with the ecu (re)programming.

Also turning it off and on again won't work forever...

Running with the knock sensors unplugged helped, as the misfire echo caused it to lean out(iirc), making it misfire more

[Andy H]

The threshold was lower, about 3k iirc, but same behaviour.

An auto ecu didn't cure it, so there might be some smoke/mirrors with the ecu (re)programming.

Also turning it off and on again won't work forever...

Running with the knock sensors unplugged helped, as the misfire echo caused it to lean out(iirc), making it misfire more

The ECU takes a signal from the MAF to calculate how much fuel to inject. If the pattern mass flow sensor reached full scale when the engine was only partially loaded then the ECU wouldn't be aware of any additional load above that mass flow rate. More load (aka intake air) without any extra fuel is going to end in tears.

Jtypecav's MAF doesn't seem to be as sensitive as ex taxi al's but both are too sensitive for the V6.

Running without the knock sensors means that the ECU doesn't know about the pre-ignition 

I'm guessing that the engine ECU and auto box ECU work together to reduce load if knock is starting to occur but the ECU in a manual can't do that so has to resort to extreme measures (leaving you with just enough to get to the hard shoulder)

[Kevin Wood]

I think the ECU just retards timing if it detects knock. There's a knock retard variable for each cylinder visible by using a tech 2. I'm struggling to understand how the auto box would need to be involved, TBH.

[AndyStobbs]

I think the ECU just retards timing if it detects knock. There's a knock retard variable for each cylinder visible by using a tech 2. I'm struggling to understand how the auto box would need to be involved, TBH.

I would agree with that. The ECU constantly pushes the timing forward until it picks up knock, then slips back.

On the MAF thing, load is a function of vacuum, revs, and air flow. A static car could be redlined in neutral and the MAF would hit its max value - or very near to, but that isn't max load, no where near infact.

A MAF that peaks before the engine can't physically use air any faster isn't going to cause massive running issues, the power will just flatline, as a balance will exist between airflow, fuel quantity and load, the engine can't continue to build revs without more fuel.

[05omegav6]

My original MAF was replaced with a pattern one a long time prior to my issue. A MAF code cropped up a while after my issue started. Refitted original MAF and code went. Still fitted and working as required.

Changing the loom made the biggest difference, eventually solving the misfire. Only measurable detail which stood out was the voltage on one cylinder being 2/3rds of the expected value. Changing coil packs made no difference to this value.

[Andy H]

I think the ECU just retards timing if it detects knock. There's a knock retard variable for each cylinder visible by using a tech 2. I'm struggling to understand how the auto box would need to be involved, TBH.

If Bosch chose to they could have programmed the gearbox to drop a gear or unlock the lock up clutch in response to any signal from the engine to say that it was labouring.

I am not stating that they did, only that they might have done. Seeing as we are guessing as to why an ECU programmed for auto behaves differently to one programmed for manual it seems as good a guess as any so far.......

Back to my main point though, I strongly believe that a pattern MAF is screwing up the airflow reading causing potentially catastrophic lean running (and the ECU is throwing a hissy fit to protect the engine)

[AndyStobbs]

Exactly why the rest of the data is required, if its going lean the lambda's will show it.

ex taxi al - Which voltage are you referring to?

[05omegav6]

Exactly why the rest of the data is required, if its going lean the lambda's will show it.

ex taxi al - Which voltage are you referring to?

Coil pack to plug

One cylinder (6) was producing 40kv instead of the 60kv at the other 5...

[AndyStobbs]

And its still the same? Could be a cylinder issue, or a lead or the plug.

The tension will only increase to the point where the air gap is ionized, a smaller air gap will conduct at a lower voltage, a less dense charge will do the same. As long as your burn voltage and times all look similar and your cylinder contributions are more or less in line then its not really important.

[05omegav6]

For the umpteenth time changing the loom sorted my issue...

[Marks DTM Calib]

Just to re-state the evidence......

.....engine runs well when ecu thinks its in an auto box application

.....engine runs badly when ecu has possibly (as we dont know if it was done correctly) been configured as manual

For me that all but rules out most components associated with ignition/injection/measurement etc

[jtypecav]

Correct fuse. The factory manual ecu and a different auto ecu told to be manual produce exactly the same symptom s.

Car is manual from factory, running auto ecu at present with no performance issues.

[Marks DTM Calib]

Has the ABS ECU been changed?

[jtypecav]

Nope! How come?