E throttle P I D settings

[ALPerf]

Hi all 

Having a bit of a mare with e throttle, only on initial setup an run in tune, done around 120 miles running in the motor all seems fine with hardware and the calibration goes through no issue, the problem is randomly occurring drive 10 mile fine then the data log shows the tps oscillating rapidly then faulting with target error and another, it's baffling me as it seems ok for a period of time then goes berserk!!

P 7    I 0.125   D 25 

Tried swinging them around every which way only results in the body clicking rapidly and bouncing around please help it's a ridiculously expensive throttle body from RCM and a subaru pedal, are there supposed to be specific PID numbers for the body and if so how do I get them

As always thanks in advance for any help greatly appreciated 

 

[Vaughan]

Attach a copy of your tune and a PC Log showing the issue please

[ALPerf]

Will do, I got that fed up I posted in desperation!! left laptop in the car at work will share tomorrow 🙃 thanks for reply 👍 

[ALPerf]

Some info on throttle body which may be useful?

[Adamw]

On 5/31/2024 at 10:11 AM, Vaughan said:

Attach a copy of your tune and a PC Log showing the issue please

 

[ALPerf]

Hi Adam/Vaughan

Thanks for reply guys sorry for delay could not get to the car until now, heres the logs and trigger scope

subaru sti v6 trigger scope.llg

how do you insert large file keeps rejecting the file size sorry

 

[Adamw]

What is the triggerscope for, you only mentioned a throttle issue above?

How to share a file: 

 

[ALPerf]

Included it as when you see the log there's a massive random spike in rpm and thought it may help with issue, I'll post log when I'm back thanks Adam

 GUNMETAL GOBSMACKER 10.pclr

Subaru v6 sti 2.5 gtx3582r 1300cc E Throttle.llg

Edited June 3, 2024 by ALPerf

[Adamw]

Did you attach the correct log?  I dont see any throttle oscillation or the "massive RPM spike" in that one?

[ALPerf]

please see attached sorry this one has it at 11.20-11.40, also original log does show the rapid pulse of the tps,pedal is wide open and showing 100%-90% variation with a spike in rpm after releasing the pedal

thanks Adam

Subaru v6 sti 2.5 gtx3582r 1300cc E Throttle.llg

subaru sti v6 E Throttle issue.llg

[Adamw]

Ok, a quick tip for PID tuning - when your control output (motor DC) oscillation is exactly 180° out of phase with the process variable (TPS), this means the oscillation is coming from excessive proportional gain.  

Example below from your log, the logging rate is a little low so the oscillation is a bit distorted but you can see the troughs in motor DC align with the peaks in TPS and vice versa.  If the oscillation was due to derivative or Integral then the peaks and troughs get offset from each other, and you can usually differentiate between integral and derivate as an integral-caused oscillation will be a slower rolling type effect.  

 

So, I would try dropping your proportional to about 6.0 and increase your derivative to about 40. The reason for increasing the derivative is because this throttle appears to have a lot of friction or stiction - the motor DC needs to change a lot to get the throttle to move, increasing the derivative often helps in these cases as it introduces some noise.   

[ALPerf]

Hi Adam 

heres a log of the results with engine not running seems a lot better, hopefully when i drive the car it will prove as successful!!! just to add as the log goes on i increased the deadband from 0.1 to 0.5 and this totally stopped the fluctuation that can be seen on 15%pedal  thank you so much for spending the time looking at this for me, you are a credit to the link brand and the aftercare is second to none!! go Team LINK!!!

Modified throttle P I D.llg

[Adamw]

3 hours ago, ALPerf said:

i increased the deadband from 0.1 to 0.5 and this totally stopped the fluctuation that can be seen on 15%pedal 

I suspect that will be too much deadband in real life, 0.5% is quite a large change in air flow at idle so I think you will find idle quite variable.  The "fluctuation" at 15% is the friction/stiction effect I was talking about, I would try the deadband at zero and try increasing derivative even further to see if that gets rid of it. You will often need to increase proportional when you increase derivative so you can still reach target quickly.   Also when testing with the engine off it is best to have a battery charger connected as you will find the lower voltage makes quite a difference to throttle control.