E-throttle testing with 4G69 throttle.

[Bethanyc]

Hay Link team, 

having some fun in lockdown to see if I can get a 4G69 e-throttle working correctly, its slightly different to the Evo X throttle body, but its same dimensions to Evo 6 throttle body so wanted to see if I can get this working first before going down the Evo X throttle path. They are also much cheaper. I was able to successfully calibrate using the calibrate button.  

Initially I set to Evo X settings in the manual, but having a few odd issues.

First one is the TPS sub is not linear compared to TPS main and stops at 54% and is occasionally throwing a sync error. Is there a way around this with a custom table somehow ? 

Second is when fast off the throttle pedal it bounces around a bit, ill play around with PID, but higher PMW frequency seemed to help i dont know if 10000 is too high ? . any suggestions welcome. 

logs and ecu file attached 

 

Thanks Beth. 

 

 

 

[Bethanyc]

I have had a better look at the logs after some sleep, 

For the TPS sub, i guess ill leave it at TPS sub (100%) to 54 % 

it looks like the error was just due to difference between target and TPS main, so ill have a play with PID and frequency today to see if i can get rid of the bounce and differences

 

[Adamw]

TPS sub and main look linear enough to me, there is no reason that shouldnt work as is.  

The rest of it just needs PID tuning.  I would set the frequency back to about 1000Hz, 10K, makes it quieter but generally generates more heat and electrical noise.   Try around 7P, 0.15I and 65D as a starting point.  

To tune the PID set up a time plot with TP target and TP main overlaying each other, have it running live.  I generally start with big step changes by constantly changing the target in the throttle table between 20 & 80% (just manually type in, or switch between two tables) and watch the response speed and amount of undershoot/overshoot, adjust 1 PID gain at a time and you will see the effect each has fairly quickly.  That will usually get you something close enough.  

[Bethanyc]

thanks @Adamw your setting were pretty much spot on. 

How does this look ?  0.152I seemed to be the sweet spot   

 

[Adamw]

Yeah it looks close enough, that is about the best you will get with PC Logging which is a bit slow to see the small changes.  You really need to use ECU logging at ~200Hz to get it better than that but I would only do that for special cases for example sequential shift with blip.

[Bethanyc]

sorry i had to have a look at 200hz logging and the oscillations at either end of the 20/80% testing were horrible !!!

especially worse when coming down to the 20% with big overshoots 

so i played around a little more, and ended up at a P = 5 and I = 0.95 

i still get small overshoot on the 20/80% testing coming down to the 20%, but up and and the way down is now more stable.  

when set back to linear it looks good, just that small overshoot when off the throttle very fast ? do you even feel this ?