legacycentral bbs

My legacy turbo was originally an automatic and years ago it was switched to manual. Occasionally the ECU would trip a code 33. I know the ECU was changed to a manual and the automatic one was tossed. I thought it was a bad MAF so I swapped it out but it didn't seem to make a big difference.

The car was horrible on gas. I did the standard things, plugs, wires cleaned the air filter etc. It also had a brutal hesitation around 0 PSI.

I was going to change the O2 sensor but instead I threw in my wideband. Seems that at full boost it runs about 10.2:1 which is not uncommon for a factory subaru engine. It's got the stock turbo, an intake, FMIC and exhaust.

Around 0 PSI where the engine is hesitating the car is going so retardedly rich that I wouldn't be surprised if fuel is dripping from the tailpipe. Try 6:1 to 7:1. Surprising the mixture even ignites.

I don't have time to do a standalone yet but I'm concerned about fuel contamination in the oil. Do you guys think the difference between an auto MAF and a manual one would be that large?

It starts fine so I don't suspect the CLT. Anything I may have missed?

-Michael

All the turbo models use the same MAF sensors; there is no difference between manual and automatic.

10.2:1 is a lot richer than my engine runs under boost... In my experience it gets up to around 11:1 but only at high RPMs.

What kind of wideband setup are you using?

Code 33 is for the vehicle speed sensor. Can you run my scantool to see what the ECU's vehicle speed reading is like?

That's your problem I bet. When my VSS wasn't working at all, I had huge hesitation problems that only occured as I came on boost or ran at very lite, to no vacuum. I don't have a wide band O2, so I have no idea what my mixtures were like though.

Had that and it included horrible miss- firing under light throttle going up a long steady hill at cruising speed. The colder the weather-
the worse the miss-fire. Nothing showed up on diagnosis.
Changed the stock ECU, with no improvement. New wires,
plugs and grounding wires--no change.

Installed SDS ECU and no problem at any throttle setting
up to 12# boost. Now have one TPS for the ECU and one
for the 4eat. No problem with tranny dollars

Aha! I must have mis-read the code 33 to be a MAF problem! Of course that makes sense to me now as I just replaced the speedo cable and was scratching my head going gee, that CEL hasn't come on in a while.

I'm having the exact hesitation problems you describe. As for the mixture, all the WRXes and Legacies I've ever had my wideband in ran in the 10.4-10.6 range under boost - if you were running leaner then there must be a reason. Unfortunately right now it's costing me a fortune in fuel and I don't have time to solder up another ECU for myself. I'd just like to fix this for now.

-Michael

Have you reset the ECU since you changed the speedo cable?

Yes. I guess the previous owner got a CEL so many times he installed a reset switch on the dash! There is currently no CEL showing on the dash.

-Michael

Did you try cleaning it with rubbing alcohol or something? I have a really crappy filter on mine and when the sensor gets dirty it makes the car buck under load. I ordered a K&N to replace the cheap ebay filter so hopefully i can stop cleaning it soon.

Just wanted to provide an update on this issue for anyone with the same problem. I've solved the issue. It was an intermitant CLT failure. The sensor would read properly but if you shook it the resistance would jump to infinite. Obviously the sensor element has worked its way loose inside. So when I tested it on-car it was fine but when I removed it the intermitant problem became apparent.

The interesting thing with this failure is that it wouldn't go to infinite resistance long enough to trip a CEL but it would jump around enough to put the car into like -50C warmup enrichment and that explains the plug fouling problem and insane richness.

-Michael

Earlier in this thread I posted that my twgn also ran excedingly rich.

This ended with the upgrade to a stand alone ECU.

However, to satisy curiosity--what's a CLT

coolant temp sensor I believe is what he's referring to by CLT.