RomRaider

A friend has the above mentioned Impreza and he asked me to help him find out the problem - he was at various Subaru dealerships and other then a PAK update and a bit of head scratching, they couldn't help. The car is pretty much stock, only exception is the rear muffler which is much 'louder' then stock.

His car often loses power, especially at low revs and mid-throttle. The car bogs down, doesn't accelerate, it 'shakes' or 'vibrates' a little - but if he goes WOT, it pretty much picks-up and goes. Since his girlfriend has an Impreza with the same engine (that works OK), he tried swapping over sensors like MAP, Knock sensor, he also swapped EGR, coil pack, HT leads - pretty much everything he could.


So we did a little logging run and what I noticed immediately was that at times, there was -9° or -10.5° 'knock correction advance'. Later, as I was reviewing the log, I also noticed that 'atmospheric pressure' is making wild jumps as well - from 0.76 to 0.98 bar (11 - 14 PSI), which I'm sure is not right.

There a quite a few times in the attached log that his car bogged down, but the one I remember most is at 617th second, where we started climbing a hill that my car Legacy (similar NA engine) takes in 2nd gear with ease, but his engine bogged down and struggled a bit even in 1st gear. Just as we turned into the hill, the 'atmospheric pressure' dropped to 0.82 bar, then jumped back to 0.95 bar. 'Knock correction advance' went from 0 to -10° to -9.5° while lambda was about 0.87 - 1.02.


The service manual for this MY says that the atmospheric pressure sensor is built into the ECU, but looking at the PCB board itself, I couldn't find anything looking like a pressure sensor:



Link to a larger photo - http://shrani.si/f/3U/f6/DNhdCvv/dsc00359.jpg

So if it's not in the ECU, where is it and how does it look like?

Danilo

This is a common problem of the JECS\Hitachi NA engine ecu.
False knock ignition retard and incorrect atmo pressure calculations.

You may try to correct knock sensor sensitivity by a C866 2200 pf ceramic capacitor that is not installed currenty.

I306 atmo pressure sensor is not installed and is disabled inside the ROM.
Atmo pressure value is "calculated" from the manifold pressure.
Open the throttle wide for a short time and atmo pressure value becomes correct for some period.

This ecu needs those C866, I306 and reflash badly.
There is no DIY solution for JECS\Hitachi ecu reflash currently.

It is probably 22611-AJ730 ecu and as far as I know there was no official update for this ROM.

If I understand correctly, the C866 capacitor might help even without a re-flash?

I also understand that you can do a proper re-flash - do you offer that as service? How much do you charge?

D.

C866 installed postpones upcoming hesitation due to false knock ignition retard.
I do reflash those ecu's that is why I am familar with the mentioned above drawbacks.
I suppose this forum is not a correct place to promote reflash services.

We added the C866 today, tomorrow we'll see if it helped.


However - today we drove another Impreza with the same 2.0 engine and the same ECU, it's the same MY and has approximately the same mileage - but this one drives normally. We logged the same parameters and did the same test route.

I'm seeing similar patterns in the log - the atmospheric pressure is not stable, knock correction advance is there (albeit a little less - just -6.5°) - generally, these logs don't look that different... Yet this one drives normally - it tackled the same incline in 2nd gear with no problem (starts at 558th second in the log).

The bad Impreza must have some other issue - I was thinking that maybe the timing belt is off by a tooth, but then it would probably run like that all the time (which, according to the owner, it doesn't - an ECU reset makes it run 'nice' for a few kilometers).

An electrical issue, perhaps? A bad ground? But then we'd probably see some 'weird' data in logs, right?

D.

This one is almost as bad as previous.
Knock processing algo should be corrected and EGR is to be disabled if those cars are planned to be revive.

Based on logs one could assume that these two cars drive similarly, with similar problems and symptoms, but this one (the 'good' one) actually drives much better than the other one - in fact, as a passenger I couldn't notice any problems, especially compared to the 'bad' one where I could hear the engine tone change and not accelerate at all.

I was hoping that there would be some substantial difference in logs of the two cars, which could explain why one drives fine and the other doesn't.


I'll PM you for details...

Danilo

so what was the solution to the problem? Probably uploading a patched ROM. I have uploaded a patched ROM, there is a slight improvement in engine power, I do not have a SAP error but I still have noticeable jerking under load between 2000-3000RPM.

That ECU was sent to Sasha_A80 and it came back working well (the car changed owners and I don't have recent updates).