Trim Heatmap

Graphs a heatmap of short term fuel trims (STFTs) from a ME7Logger log for use in tuning the ME7.1 KFKHFM MAF corrections map.

How it works

Given a log with nmot rl and fr/frm data, it runs through each data point.

For every datapoint that is preceded by "similar" RPM/load samples (to throw out acceleration/deceleration points), sort it into a KFKFHFM cell.

For each cell that has enough samples, calculate the average and convert the lambda correction to a trim percentage.

Each of the above filters can be tuned individually - see below.

Download

Check the releases section of this repo. Download the zip file for windows or one of the source archives for MacOS/Linux.

Installation

Unzip the zip/tar file into an install directory.

Linux/MacOS

pip3 install argparse matplotlib numpy pandas seaborn

See also requirements.txt

Running

This is a command line utility. It does not have a GUI. If you don't know what cmd.exe or /bin/sh is, this tool is not for you.

Windows cmd.exe

C:\PATH\TO\HEATMAP\INSTALL\DIR\heatmap.exe log.csv

heatmap.exe C:\PATH\TO_\SV\log.csv

C:\PATH\TO\HEATMAP\INSTALL\DIR\heatmap.exe log.csv log2.csv log3.csv

You can also use the built in filename globbing to do this for you:

C:\PATH\TO\HEATMAP\INSTALL\DIR\heatmap.exe *.csv

Linux/MacOS shell

chmod +x /path/to/heatmap/dir/src/heatmap.py

/path/to/heatmap/dir/src/heatmap.py log.csv

./heatmap.py /path/to/csv/log.csv

/path/to/heatmap/dir/src/heatmap.py log.csv log2.csv log3.csv

You can also use shell globbing to do this for you:

/path/to/heatmap_dir/src/heatmap.py *.csv

As usual, you can always add heatmap.py or heatmap.exe to your PATH!

Headless

For pure text output, use --csv or --text

Adjusting the filters

• Adjust the "previous samples" amount with the -w WINDOW option
• Adjust the "similar" load filter with the -l LOAD_FILTER option
• Adjust the "similar" RPM filter with the -r RPM_FILTER option
• Adjust the "minimum number of samples" filter with the -s MIN_SAMPLES option
• Disable the filter entirely with -n

Adjusting the heatmap

• Show a continuous heatmap (instead of bucketed for KFKHFM) with -c

• Choose frm instead of fr with the -fr option
• Choose an unweighted median using the -u option. The default is to average the samples weighted by their distance from the "center" of their cell (does not apply to -c). Use -v to see the difference between the weighted average and the mean.