The voltage should flick up and down, it's the lambda probes switching between rich and lean.

It looks like the system had learnt an overall correct of ~22%.

You cleared them, so learnt adaption was 1, at this point the system needed ~15% instantaneous correction.

Over time this short term correction is re-learnt into the long term, so you've ended up back at at overall correction of ~22%.

As too having 20% correction, it feels on the high side?

Mine reads around 4% long term......

A positive correction means the system is adding fuel, so 'reading' less air than there really is.

 

The voltage should flick up and down, it's the lambda probes switching between rich and lean.

It looks like the system had learnt an overall correct of ~22%.

You cleared them, so learnt adaption was 1, at this point the system needed ~15% instantaneous correction.

Over time this short term correction is re-learnt into the long term, so you've ended up back at at overall correction of ~22%.

As too having 20% correction, it feels on the high side?

Mine reads around 4% long term......

A positive correction means the system is adding fuel, so 'reading' less air than there really is.

Does this indicate the presence of a vacuum leak then?

 

Agree with sub8. The "adaption value multiplicative" readings are the long term trims so the pre-cat o2sensors are reading too much o2 and telling the dme to increase the injection time.

Those readings are high. If it gets much worse you will get the SES light and mixture control codes.

It could be an air leak but my first suspect would be MAFs as they under-read flow as they degrade. Another possibility is low fuel pressure.

 

Agree with sub8. The "adaption value multiplicative" readings are the long term trims so the pre-cat o2sensors are reading too much o2 and telling the dme to increase the injection time.

Those readings are high. If it gets much worse you will get the SES light and mixture control codes.

It could be an air leak but my first suspect would be MAFs as they under-read flow as they degrade. Another possibility is low fuel pressure.

MAFs are newish.

I want to better understand the low fuel delivery pressure angle better. Please let me know if I understood this correctly: a low fuel pressure would imply less fuel mass flow rate, implying that the injection cycle needs to be longer in order to deliver the right amount of fuel. What doesn't make sense to me is that if the DME can compensate for lower pressure by increasing the injection timing, the AFR should theoretically be ideal.

 

I think I understand now. Fuel trims and AFRs are two different animals. Just because one is fine doesn't mean the other is.

 

As mafs age they become fouled- a coating develops that causes the actual thermal film to become slightly 'insulated',,,this means that when say 100 cu feet of air is flowing, they only 'feel' 80cuft. The DME adds the proper amount of fuel for 80cfm. But then O2s say "wait a minute, that mixture is light on fuel, better correct" and they add more fuel. This 'correction' is the trim. A little is fine, too much and you get a code.

Your AFRs are typically correct at the tailpipe, it is the trims necessary to drive to that AFR tell you what is going on.

(Although, these motors do funny things at different RPMs - the combination of timing and valve position can impact AFR/trims at different rpms- sluggish vanos operation can cause the motor to run rich at some places on the rpm band, yet OK at others.)

How comfortable are you with the MAF quality? Stock paper filters, right?

A

 

How comfortable are you with the MAF quality? Stock paper filters, right?

The MAFs are Bosch 029s. Fouling is possible but not probable - the MAF test score is more or less the same as when they were brand spanking new. Affirmative on the stock air filters.

There's another possibility that is somewhat unique to the s62 because of the oil separator PCV design -a bad air leak into the crankcase, usually through old, compression-set valve cover gaskets. The higher the load on the engine, the higher the vacuum in the plenum and the more air pulled through the separators, air that bypasses the MAFs. So extra fuel has to be added to get back to the correct AFR. I'd be a bit surprised if the leak can be bad enough to get LTFTs over 20% but maybe.

Interesting possibility. I'm expecting valve cover gaskets in a few days' time from my local BMW dealer as preventive maintenance, so I should be able to discount this root cause relatively soon.

 

There's another possibility that is somewhat unique to the s62 because of the oil separator PCV design -a bad air leak into the crankcase, usually through old, compression-set valve cover gaskets. The higher the load on the engine, the higher the vacuum in the plenum and the more air pulled through the separators, air that bypasses the MAFs. So extra fuel has to be added to get back to the correct AFR. I'd be a bit surprised if the leak can be bad enough to get LTFTs over 20% but maybe.

 

Hot idle or low to medium load & rpm the system wants to run at lamdba = 1, with a perturbation of ~0.05 ish +\-. With everything perfect, all adaptions are 1, and the probes should switch back forth. Lambda 1 is around 0.5 - 0.6 volts. Any drift is measured by sensors requiring an integrated correction. This is then learnt to the long term adaption when certain entry conditions are met - stable speed and load for example. So the instantaneous correction gradually passes into long term and should return to 1ish.

Air leak is likely with high corrections values of > 10ish %. Assuming all other sensors / actuators are good!

 

Noticed yesterday that INPA mentions injection time as well.

How much should the injection time be if the fuel trims are healthy?

 

Injection time will vary wildly depending on the rpm / load / lambda requirment.

 

Wish I'd been clearer, "zero" load, engine at idle RPM.

 

Here's a screen print at similar conditions:

I think injection time is the injector opening time during each intake stroke so it's cylinder-filling dependent and basically independent of RPM. So your 6% higher idle time is not contributing to your injection time being 17% higher. So it must be your fuel trims that are driving it so high.

Bad O2 sensors may cause bad gas mileage. Has it got worse?

 

For your two examples, the top one is a higher engine speed with more load, as such is reading a higher air mass flow by ~13% (The HLM readings coming from the MAF sensor). Therefore inorder to achieve lambda =1 the fuel injector pulse width has to be ~13% greater, which it is approximately.

 

Air conditioning on?

 

Come to think of it, yes it was.

 

This weekend was a marathon of maintenance - took the plenum off and found a perished hose under Throttle Body #7 which was replaced with new vacuum line. No other vacuum leaks detected visually or with carb cleaner. I also replaced the valve cover gaskets as preventive maintenance at the same time.

In my eagerness to see the result of this maintenance I reset the adaptation values by mistake. After a few days and 50 miles of driving, here's what the fuel trims look like though:

It would seem that the vacuum leak was responsible for 5-7% of the initial 1.21-1.22. The car definitely feels more responsive than before, but with Adaptation Multiplicatives of 1.15-1.17, it looks like there's still room for improvement.

I'm not sure how to interpret the "Adaptation Value Additive" readings as they're higher than before on Bank 2. Does it indicate that the LTFT hasn't fully stabilized yet? I'm trying to make sense of the BMW Fuel Trim PDF attached in cab's post but it's hard to believe the Bank 2 O2 sensor is responsible for the sudden change in the additive value between pre- and post-maintenance.


<HR>

The other INPA readout that concerns me is in the Run Unrest tab, where it seems that Cylinder 4, 7 & 8 have higher readings when compared to the other cylinders. Does this indicate ageing spark plugs? The only thing I've found regarding this readout is Sailor24's reply to desmotesta which could do with further clarification.

 

And here are two "Analog 1" page screenshots at idle, with and without the AC running:

With A/C

Without A/C

 

Can anyone chime in regarding the "Adaptation Value Additive" field? 150+ miles into the adaptation reset, the Lamdba Integrator values have stabilised around 1.00, but Bank 1 is still reporting a relatively high -0.08 for "Adaptation Value Additive".

 

I wouldn't consider an additive adaption of -0.08 ms high. It's only a 3.8% reduction of the stock idle injection time. The DME doesn't set codes until the adaptions get up around 25%.

Have your multiplicative adaptions come down?

That cylinder unrest looks pretty normal too. I've seen mine up around 250 at idle. It improves as you increase rpm.

 

I wouldn't consider an additive adaption of -0.08 ms high. It's only a 3.8% reduction of the stock idle injection time. The DME doesn't set codes until the adaptions get up around 25%.

Have your multiplicative adaptions come down?

That cylinder unrest looks pretty normal too. I've seen mine up around 250 at idle. It improves as you increase rpm.

The -0.08 persisteth on Bank 1. The only thing that concerns me about it is that the additives were quasi-zero before. Integrator values have settled around 0.15 and 0.17 for Banks 1 & 2 respectively. Looks like the vacuum leak was not the sole contributor to the lean AFR. The next thing I want to check out is the pressure in the fuel line because the car feels like it runs happier when the fuel tank is full. Wondering about the health of my fuel injectors as well. Also wondering about the impact of ageing CPSs and TPSs on all of this.

Thanks for the insight as always, 68.

 

Where the error is learnt, with regards the speed & load you are driving at, will affect how the two adaptive values change. I think you posted a link to a comprehensive document outlining this.

In essense the additive adaptive is learnt at low engine speed and low to mid load while the multiplier is learnt at higher engine speeds and loads. So, if you`ve cleared the values but not driven much at higher speeds / loads the multiplier will not have learnt any adaption and most will be in the additive.

Don`t forget that these corrections are right at the end of the PCM calculation chain. Any erros in MAF, fuel pressure, injectors, air leaks, the lambda sensors themselves etc etc will all end up in these fuel trims......