I have had a wee sabaitcal from Subaru's, so my noggin needs some gentle cohersion into the way of all things "stars of palaedes".
The car in question is a phase 2 Legacy, 2.5 SOHC MAP based ECU - JECS. UK spec car - 1998 (99MY).
This problem existed 3years ago and has not changed.
2.5 years ago this was my daily driver, lovely car sweet as anything. I put the car in for its annual MOT (ministry safety) test and it failed for emissions. I picked the car up and could barely drive the thing home. One of the exhuast valves was burnt out. I am to this day dubious that it should happen on the emissions test as the car is tested at idle and fast idle only. Live and let live though.
Car is now rebuilt - 2.5years later
. It has been standing for a long time without a battery. Whilst the heads were off I ported them and raised the comp a little. Fitted a new lambda sensor and obviously - a new battery. The car passed its emissions test. These last two paragraphs may well be irrelevant.....
I have done about 6 journeys subsequent to its rebuild. Varying from 2 miles to 150miles. About 4 times during these journeys the CEL has lit - for no more than 30seconds - and knocked a lot of torque out of the engine (feels like max ign retard). The only common occurance is that it tends to trigger at lower vehicle speeds - say below 30mph.
Because I have a JECS ECU the OBD is of little use to me as it is not OBD2 compliant (EU didn't have to be compliant until 2003 for petrol cars). I seem to remember that I can short the connectors between the steering column to read the flashes off the CEL. However, I don't have a perm CEL, only intermittant, will I still be shown the true fault path without a perm fault code?
Also, any ideas spring to mind? Faulty knock sensor maybe? Cruddy 2.5 year old fuel......?