addo wrote:The sootiness as shown on that earthing picture, seems common and I believe to be a feature of degrading plastic in a hostile environment (ie; solvent vapours, heat). There's negligible corrosion/oxidisation visible, which is a good sign.
I wonder whether the high electrostatic field around the spark plug leads helps attract soot or whether it contributes to the degradation in the earth wire insulation ? As there was no soot on the other terminal or any other wire for that matter.
I hope the replacement eye was well crimped; if you look at the cut cross-section of a professionally crimped connection it is usually squashed into a heartlike shape and there is SFA airspace visible. Many hand crimpers just mash the wire strands into a sort of dented oval.
Well the replacement terminals are not the two prong type like the originals that roll into a heart shape, its a butt joint crimp type eg just a round tube that gets squashed. I did use a double width ratchet crimper though, not one of the thin £5 manual ones with no ratchet, so its crimped as well as it could be. Seems to be a good tight fit with plenty of wire volume to fill the hole. (the wire only barely fitted into the hole before crimping)
Either way although it's good to have ruled it out the earth wires could just have been yet another in the long list of red herrings in the quest to find an intermittent fault! I certainly wasn't able to prove with measurements that they were faulty in any way. The Lexia ECU supply voltage reading after replacing them is identical to what it was before as well.
I've been thinking again about what was done yesterday apart from crimping the new terminals on. To measure the voltage on the ECU I had to remove the plug from the ECU and slip it out of its sheath to get at the back of the terminals - in the process the temporary memory (eg fuel trim and knock retard tables) will have been cleared, after crimping the new terminals I went for a test drive with the plastic top cover still off the engine and the back of the ECU plug still exposed - perfect performance, tons of zip, no hesitation at all.
I then put the top cover back on the engine, reassembled the ECU plug (once again disconnecting the ECU for a few minutes) and then went for a second test drive around the same "circuit" - again perfect performance with tons of zip and no mid journey sag or fall off in performance or responsiveness. I did these two separate tests just in case it was something to do with the top cover pressing on the spark plug leads etc.
I then let the car sit for several hours and went to pick someone up from Glasgow - it started off very zippy but within only a couple of minutes I started to notice the typical flattening out of performance and hesitation creeping back in, and by the time I'd done 10 miles it was feeling quite lethargic and lacking in oomph with "rubber accelerator pedal syndrome".
What's the connection if the terminals were placebo ? Possibly two related things - one is that before both trips where the performance was excellent the ECU had just been reset by being unplugged...
The second is that on both occasions the engine had been started and left to idle for over 5 minutes without being driven or even being revved up - I was busy measuring voltage drop between different terminals, observing the idle to see if it was smooth or missing, checking live data from the Lexia against my voltmeter etc.
On the third occasion where the performance rapidly deteriorated the car had just sat for several hours, was started, and then driven away about 10 seconds after starting. I have noticed this apparent connection many times before - if the car has sat for a long time and I start it then drive away quickly chances are good that performance will be poor, immediately or relatively quickly and remain that way for the rest of the journey.
If the car has been left to idle for a good few minutes before being driven performance is usually much better. It occurs to me that any time I'm doing work on this car chasing the problem of course the car will sit idling for a few minutes while I'm checking things. Therefore its a self fulfilling prophecy - the car is left to idle for a while while/after doing work on it, it performs better, I think its fixed, when it goes back to regular use where it is started and then driven off quickly it misbehaves.
In an attempt to confirm that its driving/revving the engine while cold that causes the problem today when going out I let it idle for about 2-3 minutes before driving off, then for the first mile of the journey I tried to drive slowly and keep the revs below 2000 by manually upshifting (snow mode 1-2-3) until it had started to warm up, after dropping off the passengers I gave it a bit of a thrash on its usual circuit and for the first 5 minutes it was going like a complete rocket again just like it was after replacing the terminals. Bingo! It's not the work I'm doing to the engine its the changing conditions under which it's being started, warmed up, revved, and driven.
Long story short I think we're back to the knock sensor activity retard theory again, where either non-knock engine noise is triggering a properly working knock sensor (noisy tappets or other mechanical noise besides knocking) or a problem with the knock sensor itself or wiring. (I've even read of cases where a knock sensor bolt has come loose causing the knock sensor to rattle on its mounting)
A couple of the tappets are still quite noisy despite the engine oil change, is there any possibility that the oil pressure is low, or is low for the first few minutes then comes right ? Could it be that if the car is driven quickly after starting and revved the oil pressure is a bit low making the engine a bit nosier setting off the knock sensor ?
Resetting the ECU clears the knock table giving it a fresh start, however perhaps if the engine is revved under load when cold the knock sensor picks up noise that the ECU thinks is knock, the ECU then starts filling in retard values in the appropriate RPM/Load cell in its table. With real knock the knock will go away when some retard is applied but if its just mechanical noise the "'knock" will not respond to the retard and not go away causing the ECU to progressively retard the timing of that cell to the maximum allowable amount, which from my reading is typically between 6-10 degrees on most ECU's but can be as high as 20 degrees on some systems. (There's no information that I can find for the ES9J4 unfortunately)
Depending on the throttle/load combinations that you drive the engine through when its cold and "noisy" you end up with seemingly random cells in the knock retard table filled in with large negative values and other cells that are untouched hence flat spots, hesitation and severe lack of power, but only at particular rpm/load combinations. It certainly feels like retarded timing, and it does feel like there are certain rpm/load combinations where its severely retarded and bogging, and others where it feels quite normal.
So what do I do guys ? I guess the first step is I have to prove conclusively that knock sensor activity is indeed retarding the timing when it shouldn't. The problem is I don't believe the knock retard is separated out on the Lexia display - I though it was included in the "torque reduction" figure but now I'm pretty sure its not. I believe that figure is reserved for "deliberate" torque reductions, eg the gearbox requesting torque reduction during a gear change, and I also think its used for timing controlled (fine) idle speed trimming, with the knock table related retard being built into the base advance figure.
That means there's no real way for me to see it on the Lexia because the base advance will change under different conditions and I don't know what it should be in the first place. So monitoring it with the Lexia is out.
I think the only way to prove this is to turn down the sensitivity of the knock sensor a bit with some resistors and see whether the power loss, flat spots and hesitation go away. I'm probably safe to do this while running on 99 octane and I would only reduce it a small amount anyway.
Assuming this does identify the knock sensor signal as the culprit, finding what's actually causing it could be difficulty. Low oil pressure ? Faulty hydraulic tappets ? Other engine wear noises ? Loose or over tightened knock sensor ? Broken knock sensor ? Broken wire from the connector to the knock sensor ? (I've already confirmed the harness side wiring is ok) There are a dozen different things mechanical and electrical that could cause false knock triggering...
I WILL beat this thing!
Or at least find what the problem is...
PS I think the new coil pack did genuinely die on me in the last month, the symptoms of that were very different than those I'm still chasing which have been there intermittently for months, I think the coil pack was just an annoying distraction that sidetracked me from the real problem.
Performance at higher rpm above about 3500 seemed normal. 
hand to heart, its not the cat, and in no way could it match the symptoms. For starters there is no loss of performance above 4000 rpm. Although I had performance loss above 4000 with the coil pack playing up I'm not seeing any now. If I put my foot down in sport mode above 4000 rpm it pulls like a train, which wouldn't be possible with a restricted exhaust.