View Single Post
Old 05-27-2011, 06:48 AM   #4802 (permalink)
mora
EcoModding Apprentice
 
Join Date: Oct 2009
Location: Jyvaskyla, Finland
Posts: 143

Golfwagen - '89 Volkswagen Golf mk2
90 day: 107.14 mpg (US)
Thanks: 44
Thanked 35 Times in 28 Posts
I believe it is all about how and how fast controller can match its output with your input (throttle). In ideal situation there is no overshoot and time to match the goal is close to zero.

See CTM: PID Tutorial

Because every system is different there is no way to make one setting that is "good" for all. Motor and voltage are the variables.

When rise time is high (controller reacts very quickly to throttle changes) the output will shoot over the desired result. This overshoot is reduced with other parameter (Ki?). Faster correction might result in undershoot and slower correction long stabilizing time. Output will bounce over and under its desired value before settlng. Timeframe for this whole operation might be something from milliseconds to a second. If output value is bouncing too much you might experience rough starts or motor revving up a bit when letting go of throttle pedal quickly.

Slower rise time results in less bouncing but lack of response.

Is lower Kp higher or lower rise time? What about Ki? I'm not sure. I experimented with mine until I found somewhat right settings I felt happy with. Playing with these settings made good difference in response. Sometimes during testing it was really awful though, hehe.

Now somebody please correct if this is total bull. I've understood the whole Kp/Ki thing this way.

Oh, I have almost 1500 miles on my controller. Doing great.
  Reply With Quote
The Following 2 Users Say Thank You to mora For This Useful Post:
dneale (06-10-2011), mrbigh (05-27-2011)