Pid Tuning - Practical Tips; Slave Axis; An Axis With Toothed Racks (Straight Toothed) - CS-Lab CSMIO/IP-A Manual

• Next we tune the kVff. Again chose
constant motion mode
moves are the best because a posi-
tion error is the largest on u-turns.
On the graph watch a position error
(red line) and increase (by +1) the
kVff parameter. The error ampli-
tude will decrees significantly till
some moment ant then and then
increase again. Find the kVff value
with which the error amplitude is
the lowest.
During tuning remember to click „Apply" after any value change only when these values are approved and
send to CSMIO/IP-A.
AT this stage the PID controller (regulator) is already tuned. Now you can decrease maximal permissible
position error („Max following error"). You can set 4 x value of „Max Error" from the Monitor group.
Close the configuration windows, all the settings will be saved on a disk. Close Mach3 and a system. Con-
nect a motor of another axis and start tuning like you did before.
11.7 PID tuning – practical tips
With a little practice the manual PID tuning takes 3 minutes and gives better results than 95% of cases
with autotuning use. However sometimes there are some not typical problems which can make your work
harder. Below some examples from "our experience".

11.7.1 Slave axis

In case of a slave axis we should tune both sides „Master" and „Slave" alternately going with the gain Usu-
ally, the best is to have equal gain on „Master" and „Slave" side unless we have different motors or if
there are significant differences in the inertia or friction. If we tune an axis which is a gantry e.g. of a plot-
ter or gantry mill we should set a spindle in the middle of the gantry to align the inertia during tuning.

11.7.2 An axis with toothed racks (straight toothed).

Toothed racks are not very "liked" by servo drives. Torque fluctuations that come from the toothed racks
cause resonance at some speeds and it is very hard to handle with it. The sharper the PID is the larger is
the resonance, the gentler PID (smaller gain), the smallest resonance but positioning is worse. We can
only help ourselves with kD parameter, which has some damping properties. Depending on a servo drive
we use we can try to use digital filters built in, although very often the filters can also make the positioning
worse. If we are going to use the filters then we strongly recommend very small damping values. It is very
important with this drive train system to use proper pulley. We tuned machines where a constructor used a
C S - L a b s . c . – C S M I O / I P - A - C N C c o n t r o l l e r
Short
Page 85