Steve,

I don't know what the term AFAIK means.

Haven't read Astrom, but I think I heard him speak in the early eighties about a simulation tool he'd developed. He's from Sweden isn't he?

About the square wave thing. Might he have been using a square to excite every mode in the loop? You know, looking for resonances?

In the early eighties it was required of me to measure the frequency response of a digital position loop. The only place to measure with the equipment available at the time was at a potentiometer used for position feedback. The solution was to break the wire between the wiper and the controller and sum in (series) a variable frequency, variable amplitude signal. Output was measured in the normal place, the potentiometer wiper. Since the test signal was applied at the feedback input upstream of a compensation filter in the feedback loop, the closed loop equations had to be manipulated a little bit to compute the expected open loop transfer characteristic. After that experience, the hardware and software guys took pains to provide a test signal input channel which could be mux'd to any point of the system for which test excitation was necessary - nearly everywhere!

Cheers,

Bruce