Closing The Loop; Proportional Gains; Integral And Differential Gains - Delta Computer Systems TMC 188/40 Series Manual

How It Works
as possible. Servo valves usually need a current output so the output drive will not be affected by
changes in coil resistance.
NOTE: Some servo valves have two 40 milliamp coils which will draw a total of 80 milliamps when wired in
parallel (MCM configured for ±100mA current mode). This means the valves will be fully open at
only 80% of the drive. If so, the OVERDRIVE status bit will not be set since the output drive never
exceeds 100%, which can interfere with the automatic feed forward adjust command. See
Appendix D for an application note showing how to scale the output current to avoid this problem.

Closing the Loop

The TMC 188/40 uses the information from the position sensor to correct for differences between the
ACTUAL POSITION and the TARGET POSITION. A positioner will tend to drift away from the
TARGET POSITION and the change in location is sensed by the position transducer. The Motion
Control Module finds the difference between where it is (ACTUAL POSITION) and where it should be
(TARGET POSITION). It then changes the drive output so the ACTUAL POSITION will move back to
where it should be. This is called "closing the loop."
When an axis is moving, the MCM's target generator changes the TARGET POSITION so it follows
the move profile defined by REQUESTED POSITION, SPEED, ACCEL, DECEL, and MODE. If the
ACTUAL POSITION is behind (lagging) the TARGET POSITION, the MCM will increase the drive to
help the ACTUAL POSITION catch up with the TARGET POSITION. If the ACTUAL POSITION is
ahead of the TARGET POSITION (leading), the computer will reduce the drive output to slow the axis
down. The amount of drive change for a given position error is determined by three different error
gains: PROPORTIONAL STATIC GAIN, PROPORTIONAL EXTEND GAIN and PROPORTIONAL
RETRACT GAIN.

Proportional Gains

The ability to independently adjust STATIC, EXTEND, and RETRACT GAINS lets you compensate
for differences in system dynamics such as the difference in force and velocity constants of a
hydraulic cylinder when extending and retracting. Thus the difference between the ACTUAL
POSITION and TARGET POSITION can be reduced by increasing one of the three error gains.
However, if a little bit of error gain is good, a lot is not always better. The point where an axis will
oscillate depends on the response time of the system. The slower the response time, the lower the
error gains required to cause the system to oscillate. Generally, use the largest error gain that
doesn't cause oscillations.

Integral and Differential Gains

In addition to the three proportional error gains (STATIC, EXTEND and RETRACT GAINS), there are
two others: INTEGRAL and DIFFERENTIAL GAIN. The majority of applications do not require the
use of either the integrator or differentiator. If you need to use them, start with a value of 2 for the
DIFFERENTIAL GAIN and 50 for the INTEGRAL GAIN.
The integrator is only active when the axis is in motion. When a 'G' command is issued, the integrator
starts adding the position error (ACTUAL POSITION - TARGET POSITION) to an accumulator every
2 milliseconds. As long as the error is not approaching zero or has not changed sign, it is added to
the accumulator. When the axis starts ramping down, the accumulator is decremented at a rate
calculated to bring the value to zero when the axis stops. The drive provided by the integrator is
given by:
44
TMC 188/40 Motion Control Module
Delta Computer Systems, Inc. 360/254-8688