Re: [SI-LIST] : Overshoot/Undershoot

D. C. Sessions ([email protected])
Mon, 23 Feb 1998 16:52:26 -0700

Mike Mayer wrote:
> =Alas, 'tis true. For too many years the digital crowd, having
> =heard the term 'undershoot' but never really worked with it,
> =assumed that it was what is known more precisely, negative-going
> =overshoot. Books, articles, and instrumentation all made
> =this usage the only one known to many young and impressionable
> =engineers.
> So for those of us who have only heard "undershoot" applied to
> negative-going overshoot, what is the correct definition of
> undershooot?

You'd call it "bounce back." Once the signal crosses its
steady-state value the overshoot and undershoot form the
extrema of its envelope around that value. It's usually
presented in terms of the behavior of a two-pole response
to a step function, where the error forms a damped sinusoid.
More complex responses (such as delay lines...) aren't so
susceptible to closed-form analysis.

D. C. Sessions
[email protected]