Tuning with Feedforwards

What is Feedforward Tuning?

Feedforward coefficients tell the controller what forces will be required to move the stage in an arbitrary trajectory. Feedforward tuning is the process of determining the optimal feedforward coefficients for the system. This will allow the controller to accurately predict the forces required to move the system.

Feedforwards result in more accurate motion than if feedforwards were not used.

Feedforwards result in virtually instantaneous response of the system.
Feedback control loops (using PID loop or similar) take a finite amount of time before reacting.
Feedforwards rely on an imperfect model of the system. This means that the feedforwards need help from the feedback control loop in order to get accurate motion.
Feedforwards make the bulk of the move very quickly, while the feedback control loops corrects the small errors that remain. As a result, a faster settling time is achieved than if feedforwards were not used.

There is a common misconception that feedforwards are similar to control loops and result in instability.

Feedforwards are open loop, so they cannot suffer from closed loop instability.
Feedforwards are typically less sensitive to being misadjusted than closed loop parameters.
Feedback control systems can be excited into instability by grossly misadjusted feedforwards. However, the amount of misadjustment in feedforwards that is necessary to cause such instability is very rare.

What Does Feedforward Tuning Not do?

Feedforward tuning has no impact on a system's response to external disturbances.
Feedforward tuning does not go unstable by itself.

What Can I Expect from Feedforward Tuning?

A lower following error during motion.
Quicker settling after motion and changes in motion trajectory.

What Can I Not Expect from Feedforward Tuning?

Better performance while the servo is at rest.
Better resistance to external disturbances.
Smoother motion during a constant velocity move.

Previous | Back to Table of Contents