Automatic gain scheduling, A.10.2 automatic gain scheduling, A.10.2.1 tuning a speed controller – ElmoMC SimplIQ Digital Servo Drives-Bell Getting Started User Manual

Page 80

Advertising
background image

The SimplIQ for Steppers Getting Started & Tuning and Commissioning Guide

MAN-BELGS (Ver. 1.1)

80

For manual gain scheduling, tune a series of fixed controllers and log their
parameters. Keep in mind that the High-order filter must be similar for all the
parameter sets. Then program them to the controllers’ array using the KG[N]
command, and set GS[2] for the selection of the appropriate controller. Refer to
the chapter on the Speed and the Position Controller in the Application Manual
for more details.

A.10.2 Automatic Gain Scheduling

In many applications there is a good reason to schedule the controller gains by
the speed command. The two main reasons are:

At low speeds the information comes from the encoder at a reduced rate. The

resulting delay destabilizes the controller. The motion controller starts to
develop oscillations until the instantaneous speed is enough to stabilize the
motor. The only way to render the information delay insignificant is to reduce
the controller’s bandwidth, sacrificing high-speed performance. The automatic
gain-scheduling process enables switching to slower controllers when the speed
is slow, and recovering full performance when the speed demand or the
position error12 are high again.

At low speeds the plant behavior may change significantly, due to frictions and

plays. Frictions appear at higher speeds as constant disturbance, easily taken by
the integrator of the controller. Plays may be less disturbing at high speeds,
where the sign of the torque is fixed.

A.10.2.1 Tuning a Speed Controller

The highest speed for tuning is about two encoder pulses per controller sampling
time, since for higher speeds the information delay is insignificant.

The tuning speeds are calculated to generate equally distributed information
delays. For this reason, at high speeds there is a larger gap between the speed
options, whereas at slower speeds the gap is smaller.

You can select any subset of the option speeds and tune the controller for them.

The following algorithm calculates the maximum recommended closed loop
bandwidth for a given speed:

Delay = Speed controller sampling time + 1 / (2 Speed reference). The first term
in the formula includes the computation delay, the zero-order-hold delay,

the

speed calculation delay, and other minor delay contributors. The second term is
the information delay.

Minimal recommended speed settling time = (20 to 25) times the delay.

For example, for a speed of 500 count/sec, the information delay is 1 msec. With
TS=75, the speed controller sampling time is 300 usec, totaling 1.3 msec of delay.
The best speed settling time for the speed of 500 count/sec is about 30 msec.
Speed settling times down to 20 msec can be achieved for the speed of 500
count/sec with reduced phase margins.

12

In position control, a position error translates into a speed demand.

Advertising