Struggling with WOT MAF tuning

[HappySalesman]

So I've been trying to get my MAF curve dialed in, and I'm struggling to get accurate readings.

One problem is the MAF heatsoaking in the afternoon, especially if I get stuck in traffic, which seems to be throwing my readings way out of wack.

The other problem is having enough counts to get accurate information. As it is at redline I can only get around 5 counts or so which I know really isn't enough, unless I sit there bouncing off the rev limiter (which I'm sure the car cutting fuel/spark will throw off my wideband anyway).

I've attached a copy of my current tune, as well as a couple of logs.

Log 1 was a drive in the morning with it being relatively cool outside and minimal traffic. Log 2 was in the evening when it jumped to around 85, and I got stuck in a fair amount of traffic before I could find a place to go WOT.

Any/all advice is greatly appreciated.

Log 2.hpl
Log 1.hpl
MAF 080222.hpt

[aaronc7]

Is the MAF surrounded by metal piping? Trying to isolate it from direct radiant heat can help avoid some of these iat reading falsely high issues

I try to get multiple WOT runs in a single long log, and then use filters to remove transient data. Hopefully it's over 10 cells at higher rpm/maf flow areas, but I just take whatever I get at that point.

Ultimately some variation is normal and why LTFTs and closed loop fueling is used. I try to get a long 30 minute log if possible and as much good and varied data and just use the averaged out result. Once you to go +/- 5 percent you're starting to split hairs and could end up chasing your tail, IMO.

I can't open the logs right now to see how bad or inconsistent your data is though.

[HappySalesman]

Is the MAF surrounded by metal piping? Trying to isolate it from direct radiant heat can help avoid some of these iat reading falsely high issues

I try to get multiple WOT runs in a single long log, and then use filters to remove transient data. Hopefully it's over 10 cells at higher rpm/maf flow areas, but I just take whatever I get at that point.

Ultimately some variation is normal and why LTFTs and closed loop fueling is used. I try to get a long 30 minute log if possible and as much good and varied data and just use the averaged out result. Once you to go +/- 5 percent you're starting to split hairs and could end up chasing your tail, IMO.

I can't open the logs right now to see how bad or inconsistent your data is though.

Yes it is. I'm fairly certain that's part of the issue as my IATs on the second drive never dropped below 100, despite ambient air being around 85.

If I up the cell counts to 25 it filters out the highest RPM data, but gets within a few cells (stops at about 8125Hz, where normally my airflow pegs out at 9000Hz) but it's close.

Looking at the data from the graphs, most everything except for a couple of cells are within 5%. So maybe I'm worrying over nothing. This is my first time doing this, and i want to make sure I do it right and I'm not about to horribly damage my motor.

[aaronc7]

Here's what I came up with for O2 error for both logs, with some basic filtering. Looks pretty decent to me.

I like to start with a stock MAF table (or scaled MAF stable as required), and make broad changes based on data from a good log. For example I won't multiply one cell by 1.05, then one 0.98, next one 1.04. I try to look at the big picture and make broad / manually averaged changes so I don't get this weird spikey MAF curve or two cells next to each other with the same/close value etc. I've had good results doing it this way. At the very end I might make some finer level tweaks if I keep seeing the same data/results.

Then make sure closed loop is back on and move on. Maybe do a close loop, fuel trim datalog to make sure it 'agrees' with your wideband/open loop results. Also helps verify your wideband and or narrowband sensors aren't out to lunch

https://i.imgur.com/zQ5N3MI.png