FL S8 erratic idle

[MikkiJayne]

Hi guys,

So... this is not quite how I was expecting my engine swap to go. The new engine is in but it has a slightly erratic idle It starts fine, and idles at 800-830 rpm while warming up, but you can feel the shudder through the body of the car. Its not enough to be visible looking at the engine, but you can hear it, and with it on the lift there is more freedom of movement in the body than with it on the wheels I think.

This is a brand new zero-miles AQJ engine - mechanically identical to an AVP apart from the engine mount brackets and the water pipe on the back.

Original:
AVP harness and ECU
O2 sensors
MAF
Engine mounts
Gearbox

New:
Engine
Coil packs
Spark plugs
Injectors
Throttle body
Vacuum hoses (everything on the engine up to the brake servo hose)
Temp sensor
5w30 Longlife III

I've done a throttle body alignment which helped it start better but didn't affect the idle.

The ECU reports no errors or misfires. First O2 sensor readings are all plausible, second ones haven't had a chance to warm up yet but aren't showing any codes.

The MAF reads ~11g/s at idle which seems high based on the range in VCDS. Unplugging it has no effect other than a code and 0 reading. The idle is the same. Engine load is 25% at idle which seems very high (90hp just to idle?) but it is brand new so probably a bit tight. No AC or large electrical consumers on.

SAI is gone and blanking plates fitted to the heads. The solenoid is still connected so no code for that. It hasn't been in conditions where it would activate it yet so no codes for insufficient flow either. I'll get that deleted in the ECU when it moves again.

I can't drive it yet because it has no front suspension yet, but it revs freely enough and seems smooth above idle with no load. It sounds fine too.

Any suggestions please?

[Goran]

I am usually useless at diagnosing anything complex like this with no error codes,
so here is what ElsaWin suggests to check in exactly that situation, idle outside range with no error codes.
Let me know if you want PDFs for the checks it links to.

A silly question, did you loop round the vacuum hose to both outlets of the N112 SAI solenoid valve? Also does it continue clicking, Elsa says it should always be clicking.

[MikkiJayne]

Thanks for that - I hadn't thought to check ElsaWin. It just occurred to me that if one of the O2 sensors is bad it could be messing up the fuel trim for one bank. O2 sensors are first in the checklist in that doc I wonder if it can re-learn trim? I have another pair of sensors I could try actually. Hmm...

There is no vacuum connection the N112 valve any more. I've removed all of that plumbing completely, with just the manifold flaps remaining so there's just the valve sitting on the fuel rail with the plug on it. I'll check if its doing anything, and also the evap valve.

Thanks

[MikkiJayne]

Great success! Conclusive proof its the MAF giving bad readings, despite no codes.

Earlier today I unplugged the MAF to see if it made any difference. It didn't. I forgot to erase the fault code - that's important...

Reading up a bit more about the O2 sensors in ElsaWin I found the procedure to re-learn the adaptation values (basic settings 32 & 33), and also how to turn them on and off (basic settings 99). So, over to the workshop and do those, and immediately it starts running nicely All O2 values in range (despite being -25% earlier today), and a pound coin balanced on the inlet manifold for ten minutes without falling over. Rpm rock solid at 760.

Bonus. Oh, and clear the MAF error code while I'm here. Immediately rpm drops to 650, cough, cough, cough, runs like crap! Check the O2 values and they're pegged at 25% again. Turn the O2s off and it barely runs at all. Un-plug the MAF again, re-learn the O2s and it runs fine again. As soon as I clear the error code, back to crap.

Interestingly, engine load is just in the zone with the MAF unplugged too - 20%, with a range of 14-20. That seems plausible for a tight new engine and the rad fan running.

A call to TPS in the morning methinks

Oh, and Goran I owe you a for pointing me in the right direction!

[HPsauce]

Hmmmm, I'm not convinced by your logic. The MAF is a pretty dumb sensor (heated wire cooled by airflow), the codes are probably stored elsewhere.
And you said that unplugging it made no difference.

And how does erasing a stored fault code cause a problem? Which controller were you accessing? I'd be treating that as the primary suspect....

[MikkiJayne]

The logic is that with the MAF open circuit fault code tripped (or in fact the MAF unplugged altogether) the ECU is ignoring it and running from it's maps and the feedback from the O2 sensors. It can correctly learn the adaptation values for the O2 sensors, and it runs fine.

As soon as I clear the code it starts using the MAF again and the readings that gives are conflicting with the feedback it gets from the O2 sensors. Their adaptation values go out of range and it runs like crap. Elsawin says that a 25% value on the O2 sensors means a bad MAF, and given that it's readings are out of normal range (VCDS says 3.5-7.0 and it's reading 10-11) I conclude that while it is functioning, it is giving incorrect values.

Another telltale for me is that with the MAF unplugged (or the error code tripped) and the O2 sensors turned off it runs fine in open-loop mode (ie running from maps, rpm and throttle position). Reconnect the MAF and clear the code and it barely runs at all with the O2s turned off. Turn the O2s back on and it slowly comes back to idle speed but coughing a bit.

Finally, all of this tallies with what an Audi Master Tech told me many years ago when I had my first D2 - that they saw a lot of odd problems with these engines which were fixed by changing the MAF, even though there were no fault codes.

I am suitably convinced by it's behaviour that I'm going to order a new MAF in the morning, but I'm open to dissenting opinions

I neglected to mention it has a K&N air filter. This is the first case I've ever seen of a K&N and bad MAF happening together, and while I don't believe the tales of K&N badness told by the TDI crowd, I'll be ordering a stock paper filter to go with the new MAF

[Goran]

No worries, I'm glad it was an easy fix.
These MAFs go bad so often and produce weird symptoms, I'm thinking I should stock up on a spare one. The one in my car must be at least 5 years old.
Thanks for sharing your findings with the K&N, I never trusted it even though there isn't much proof around, so when I bought my car which came with a K&N I replaced it immediately.
Also some Japanese guy did a dust test and the K&N and some other foam filters and they let a lot through compared to the paper filter.

[HPsauce]

Originally Posted by MikkiJayne

The logic is that with the MAF open circuit fault code tripped (or in fact the MAF unplugged altogether) the ECU is ignoring it

OK, I see the logic of that, so it's rather more than a "stored" fault code, it's an ongoing effectively permanent "hard" fault.

Originally Posted by Goran

I'm thinking I should stock up on a spare one.

I swapped mine a couple of years ago (think I mentioned it recently) thinking it was the cause of a problem throwing no fault codes (it wasn't) so keep the old one as a spare.

[Adrian E]

Are there straightforward vcds checks for maf function? Mine's been on the car since I bought it 5 yrs ago and can't quite decide if the engine is slightly down on power/running a bit rich/ just a bit slow compared to the S5.....

[MikkiJayne]

I think there is something in Elsawin for it. I didn't get that far because the tests for the O2 sensors resulted in only one possible outcome

The best TPS could do was £140-odd plus vat with an £85 surcharge. £61 + vat from Euro Car Parts for the Bosch one is much more appealing so ordered one of those which will hopefully arrive tomorrow.