Supraquick

New car to me and it didn't run well when I bought it due to bad seals (and coils, and plugs, and VSVs, and...). I have since installed a Mazda reman and the car runs great with a couple of band-aids.

I found an issue with a lazy MAF sensor and MAF sensor with bad scaling and have that worked out for now, but I also found a suspected EVAP leak. I tested the EVAP solenoid and it was sticking open some of the time, so I swapped that and it seems to be working fine now. The current problem is when the engine call for EVAP purge and the valve to open, the car goes lean. STFT +15% or more at idle and +7% or more at cruise rpm. This did set off the LEAN code but no EVAP Leak codes.

To verify I have a leak, I have installed a variable flow control valve between the upper intake manifold and the EVAP solenoid (in place of the nice little u-shape hose on top). As I restrict down the EVAP air flow, the fuel trims go back toward normal.

I replaced the gas cap, checked the connector near the oil dip stick, and the following lines under the car based on searches I have done so far. All looks OK, but all the EVAP equipment in the back looks like a PITA to get to and will need the rear diff removed in order to test.

Without a code, it seems that would limit the number of places the leak can occur.

Any easy common leak places to check first, or will I need to bite the bullet and tackle the rusty bolts on the rear diff to test all the EVAP parts and connectors back there? Maybe try starting fluid at those parts and connectors while the car is running in the air?

Loki

iTrader: (1)

What are the fuel trims when the EVAP system is not energized? I've never gone into this much detail on the EVAP system, but wondering if there is an underlying issue and the EVAP activation just pushes you over the edge.

Supraquick

Here is the average and standard deviation with the EVAP at zero during the hour plus commute this morning.

LTFT - Average 1.8%, STDEV 1.8

STFT - Average -0.5%, STDEV 1.9

Combined - Average = 1.3%, STDEV 2.8

So if we think about 97% falls within +/- 3 standard deviations, we are looking at +/- 8% fuel trim.

If I take out the data points for full decel where the accelerator is zero and the wide band is at 1.38 Lambda, the combined goes to 0.3 average and 2.5 STDEV.