[SI-LIST] Re: 3dB or Knee Frequency

  • From: "Ingraham, Andrew" <Andrew.Ingraham@xxxxxx>
  • To: <si-list@xxxxxxxxxxxxx>
  • Date: Thu, 13 Jun 2002 07:55:25 -0400

The concept of "-3dB frequency", when equated to 0.35/Tr, implies a
single pole (or single RC) roll-off acting on an ideal (Tr=0) stimulus.

Or, a rounded pulse edge with exponentially rounded edges, such as you'd
get by passing an ideal step function through such a single pole
low-pass filter.

If that is the situation you have, then I think significant harmonics
extend well past that point.  At twice that frequency, the roll-off is
only -7dB, or little more than 2:1 versus an ideal square wave.
Depending on application, I'd go higher still.  But it all depends.
Even the ideal square wave (Tr=0) has harmonics with decreasing
amplitude vs. frequency, despite an infinite "-3dB frequency" by this
definition; i.e., the harmonics are more than 3dB lower than the
fundamental, WELL below 0.35/Tr.

I have thought of "knee frequency" (or "corner frequency") as being the
same as the "-3dB frequency", when it is a simple roll-off like this.
If you plot the log-log response of a single pole, it's the point where
the two asymptotic lines intersect.  Don't know about 0.5/Tr.  Maybe
someone found that it's better for modeling real world signals, which
often are not exponentially shaped?

But they are all back-of-the-envelope bandwidth estimates, not
necessarily meant to replace a detailed understanding of what your
requirements are and what your signals do.


To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List archives are viewable at:     
or at our remote archives:
Old (prior to June 6, 2001) list archives are viewable at:

Other related posts: