Inconsistent tune on track
 

I ran NASA HPDE @ LRP over the holiday weekend. The car is running a MSLABS built Basic MS3 with a fairly rough (and rich) tune on it.

I expected to be a few points rich in boost due to my "incomplete" tuning.(target is 11.8, tune is closer to 11.4 currently)

My first session out I was right where I expected. 11.6-11.2 down the main straight @ 150-160 KPA

My second session out I was almost a full point richer. 11-10.2 at the same boost, same temps, same tune, same everything.

I was only using SD logging during the sessions so I don't have all the info. I do see my Fuel Total Correction was 103% on the last session for the most part. I have no idea where that 3% comes from. The error I am seeing is larger than 3% lambda anyway.

Can anyone help explain the variance. I know I need to get my car tuned A LOT better, but I would expect more consistent results under long pulls in the same conditions.

My logs don't want to attach, but my tune is attached as well as 2 photos showing the average AFR ERROR

EDIT:
More details:
GT2560R with shitty 6psi actuator
EBC by MS3- very rough tuned but stable
MSLabs MS3 Basic
Water Meth running 3gal/hr nozzle, 50/50 mix, 150psi pump, solenoid. Should be about 50% dc @ this boost
GM IAT after meth. Looked stable on logs.

Heatsoak IAT correction table?

At least I remember a table like that...

mat air density table

...but he did say same temps, unless he's just guessing

Mat density table is still the original one based on ideal gas

IATS were logged and consistent between runs. Only thing that stands out is if the meth was spraying differently. Less atomized or something. Maybe a failing pump . The IAT sensor would still get wet and therefore read lower, but the actual affect on combustion is different?

These 2 sessions shown are actually session 1 and 3. Session 2 the tank was down to 1/4. Low tank light is not setup. It is possible the pump ran dry a few times in session 2. I doubt that's enough for failure thought.

Inconsistent tune on track
 

Can't pull the logs up because I'm on mobile, but are you sure the final delivered pulse width was consistent between runs?

No not sure. Ill check my logs again but because it was SD card logging I may not have pw logging

Inconsistent tune on track
 

Yeah check that out because the final pulse width is the sum of all the various fuel calculations.

If the pulse width is changing, then some input is changing (temperatures etc). If the pulse width is about the same and the conditions are the same then it would point more toward something like a meth injection issue.

I didn't log pw. Only total fuel correction , which is 100 % in the first run and 102-103% in the later run. That doesn't account for all error tho.

As far as a potential meth issue, based on the scatter plots above it looks like the error is across the board (and consistent ) thru all loads. This makes me think there is a correction on top of the fuel calc.