History of MembersPage/AttilaSzucs/BoostControl
Older Newer
2005-08-02 22:50:24 . . . . MembersPage/MarcellGal [pid hints]
2005-08-02 20:54:27 . . . . AttilaSzucs [question about boost PID]
2005-08-02 20:53:18 . . . . AttilaSzucs


Changes by last author:

Added:
Look for "pid" on SiteIndex. IAC might be a better playground to learn pid basics first (with less risks).
Changed:
boost_pid_kp=20

boost_pid_ki=80

boost_pid_kd=01

boost_pid_ilimit=FF

boost_pid_kp=20 (I'd try 40)

boost_pid_ki=80 (I'd leave it at 80 first)

boost_pid_kd=01 (I'd try 10)

boost_pid_ilimit=FF (I'd try 30)

Changed:
I've got VERY big and long overshoot. The boost rising to 1.3-1.4 bar, and slowly (about 1-1.5 sec) fall back to 1.0 bar, and stays there. I'm not asking you to give me the correct values (it's impossible, I know), I'd like to know what parameter should I change to reduce overshoot. The boost-PID tuning is a dangerous sport :)
I've got VERY big and long overshoot. The boost rising to 1.3-1.4 bar, and slowly (about 1-1.5 sec) fall back to 1.0 bar, and stays there. I'm not asking you to give me the correct values (it's impossible, I know), I'd like to know what parameter should I change to reduce overshoot.

* decrease I-term, which is the multiple (boost_pid_ki * boost_pid_ilimit). Preferrably decrease the boost_pid_ilimit first, unless you think the integral change is already too fast.

* increase P-term

* increase D-term (it was crazy-low, almost 0)

The boost-PID tuning is a dangerous sport :)