I have tested previously the AFR command and it works very well. Sets custom AFR and also when its set it goes to something like semi open loop.
The problem is blm are still active where they are and still triming.
The solution is to add clear blm + froze blm command and then set AFR target.
good to know! my car actually doesn't run closed loop, so I wouldn't have noticed that until others tested. I'll throw that in. so i guess it stops updating BLM (obviously) but the open loop AFR is trimmed by the existing cell.

If you want to clear all the effects just send full command with zeroes.
I just saw the panic button which is what is doing I guess.
yeah, and sends several times, just to be safe. when closing the program it hits the panic button in the background too.

one giant reliability problem is there's no sane reply to the mode4 command; you are never 100% sure it succeeded. i've thought about sending it twice just to be safe, or re-sending at interval as a keep-alive.

The programm is great.
i'm only partially happy with it right now, but couldn't have done it without ya! thanks for your help

Actually i have some ideas for things that it needs to do.
I hope you accept customs requests.
requests are welcome for sure! see latest version 0.5, i am adding/fixing stuff daily right now. my site now has a changelog.

this is just a hobby project so i'll add things slowly. right now the code is a bit shoddy in terms of how adaptable the event loop is, im going to have to clean it up before adding any substantial new features. this program is non-blocking single-threaded just because QT's serial driver seems to barf its guts out if you even look at it sideways.

i want to keep this fairly noob-friendly, it's designed for someone reasonably experienced with tuning but not experienced with hacking or datastreams.

im currently experimenting with Mode 3 requests, and making a new tab for extended parameters that are pulled right from memory. these will be fetched interleaved with regular requests at a controlled rate, so regular datalogging still happens (but slower as you enable more extended parameters).

also plan to integrate my prototype flash tool and test it (once i get some sockets in my spare ECM...). the analyzer will be really nice too, when it's done, as I already have a very useful analyzer written, i just have to port it.