This issue has been haunting me for a while...
For a number of months my ECU data scan would drop into negative learned ignition timing off throttle... initially around -1.5. Then a while later it was showing -3 off throttle, then very occasionally -3 under light throttle.
Since replacing my fuel pump and turbo with a TD05-18g (maybe coincidence), I found a few times it dropped into -6 and later -11 degrees of timing at light throttle (at this point the car was not driveable). I had to reset the ECU and things would be OK for a while (not ideal I know).
I then replaced the ECU data scan and got a BtSsm - gives me a little bit more data to log.
I was noticing FKC at around -3psi, light accelleration, 0.97 load, 2000rpm. I could bring on the knock by changing gears a little early into 3rd and increasing the throttle to try and get it up to atmosphere/0psi.
So I started troubleshooting the knock:
- mechanic confirmed it's actual knock/pinging with knock ears on the block
- Matt (tuner) tried pulling heaps of timing out, made no difference
- vacuum leaks checked (turbo inlet pipe a little loose - fixed)
- found a cracked breather hose - replaced
- inspected PCV valve - functions ok
- boost leak tested system - all good
- pressure tested exhaust system (turbo up-pipe gasket leaking - fixed)
- cleaned MAF sensor
- checked and cleaned o2 sensors
- had rear o2 sensor disabled in test tune - made no difference
- had wastegate set to 0 in tune to rule out wastegate chatter
- had fuel pressure tested (DW65C fuel pump relatively new - all ok)
- had turbo fitment checked (Blouch TD05 18g fitted recently - all ok)
- had heaps of timing pulled from the tune - made no difference
After all this was done, not a bit of difference
However, Matt (tuner) noticed that pretty much every time the knock occurred – the "time (ms)" in the log file hung for sometimes up to 1.8 seconds. Matt suspects the ECU is hanging, maybe due to something shorting out, who knows. And this hang would cause timing and fuelling to lag, which would cause the knock.
Here's a log with knock in a few places - http://www.datazap.me/u/cynner/ecu-drop ... 5-21-23-24 (you'll need to view time(ms) separately as the values are so high it flattens the graph).
Has anyone seen this kind of thing before? Would the ECU be hanging because it is processing the knock and working to correct? Or would it be hanging because of an electrical issue, which is then causing the knock? (which comes first?)
My next step is to check the spark plugs...
Just posting here hoping someone might have some bright ideas?
Kim