Table of Contents
Advertisement
XPS-Q8 Controller
XPSDocumentation V1.4.x (EDH0301En1060 — 10/17)
use of a small Integration Time value will limit the integration range to the latter parts
of the move, avoiding the need of a large overshoot at the end of the move to clear the
integrated following error value. The drawback is that the static error will be less
compensated.
Variable Gains
14.1.2.5
In addition to the classical Kp, Ki, and Kd gain parameters, the XPS PID Corrector
Loop also includes variable gain factors GKp, GKd, and GKi. These can be used to
reduce settling time on systems that have nonlinear behavior or to tighten the control
loop during the final segment of a move. For example, a positioner or stage with a high
level of friction will have a response which is dependent on the size of the move:
friction is negligible for a large move but becomes a predominant factor for small
moves. For this reason, the required response of the system to reach the commanded
position is not the same for small and large moves. The optimum value of PID
parameters for small moves is very often higher than the optimum value for large
moves. It is advantageous to modify PID settings depending on the move size. For users
that do not need to make PID corrector adjustments (or prefer not to) benefit from the
compensations provided by the variable gain correctors. This compensation is made
automatically by the XPS variable gain corrector by applying a gain that is driven by the
distance between the Target Position (position that must be reached at the end of the
motion) and the Encoder Position. As shown in the figure below, when the distance to
move completion is large, the total output gain from these parameters is fractional (the
"Kform term" is fractional), but as the move size or distance to final position is small
the Kform term approaches 1 and full GKx output is provided.
The parameter GKx is used to adjust the amplitude of the total output and the parameter
Kform is used set how soon this Gkx is applied. As seen in the figure below, if a Kform
of 1 is implemented, the GKx is not applied until the positioner is very close to its target
position, in this case 0. But a Kform of 10 will implement the GKx much sooner and
tighten the control of the loop further from the target position. This can be very
effective when positioning high inertial loads or when very short settling times are
critical. The default setting for the Kform parameter is 0 for all standard Newport
stages.
Figure 53: Variable Gains.
176
Motion Tutorial
Advertisement
Table of Contents
