Hondata

2012 Civic SI
Bone Stock
105k Miles

Possibly worth noting that this engine is under a Honda TSB for excessive oil consumption from the factory, quart per 5k miles in my case, cylinder blow-by. It does not have an apparent effect on proper functionality.

Purchased a FlashPro mainly to fix revhang but also for when I get a downpipe in the future. I've been pouring over the documentation and playing with a few things trying to stay safe for the engine but something isn't working correctly...

I started with the Hondata Reflash map, WOT Comp tables copied from stock EQ for the sake of AFM tuning (only saw the minor change near VTEC cross-over anyway). I don't have a photo from the first change but performed the same actions as I detailed below.

Warm-up, start logging, 20min blend of stop/go and open highway. Check XY Plot resulting in Tuning1 Adjusted as noted.
2 hours later
Flash, warm-Up, start logging, 15 Min blend of stop go, 40MPH Cruising. Resulting in Tuning2. ok... so... now we need +5 across the board but fairly linear anyway with +/-3 variance on the mean line. Modify accordingly.

Flash, warm-up, start logging, 12 min blend of stop go, 40MPH Cruising, and a short highway jaunt. Resulting in Tuning3... WTF. S.trim points are all over the place, mean line is back to jagged and got further from 0 in some places...

I've attached the latest datalog, a current map, and a stock EQ log as well. I'm wondering if I'm looking at a faulty O2 sensor or something.

Note that at 2:00 on AFMTuning3 log I have relatively steady throttle ~%10 but AF hits 17 before S.Trim ramps up to 24%... Similar actually happens under stock EQ at 9:33. 16% throttle, 17 AF then ramp to 15% S.Trim. For Hondata Tuned map Lambda tables show 10%+ lean in some lower RPM places and 20% high on the high speed tables but I haven't touched them. 24 knocks on this log too.

Needless to say I will be running stock EQ for the time being.

I tried the "Stock Tuned 6000RPM Launch" basemap because it was a MAP tune and had similar inconsistencies, again lambda tables were all over the place.

Suggestions on where to go from here?

Thanks,
-B&C

Hoping to get some input before I drop $100 on an O2 sensor.

Any input is appreciated.

Thanks
-Nick

Was the issue repeatable?

I had to leave town for work before I could truly verify that thought but yes, from what I have seen the issue is "consistently inconsistent".

I was considering cleaning the MAF sensor but it seems if that was the case I would be seeing consistently biased values... Yet I feel like the same would be true (for the most part) for an O2 sensor.