I had a course in linear systems which used Bracewell's "The Fourier Transform and its Applications". Courtesy of the dictionary of transform pairs in the back I can work in both time and frequency just drawing cartoons with pencil and paper. And Octave will take care of the detailed calculations.
My biggest handicap at present is I spent my career working with digital systems in recorded time. So I've familiar with things that can't be done in real time or in the analog domain. Lots of rabbit holes to fall into.
Consider a triangular pulse with symmetric 22 ps slopes. Can that be converted into a sawtooth pulse without significant loss of BW? If so, how? Mathematically it just requires making the pulse minimum phase to get a vertical leading edge. But TANSTAFL. If it were easy, 1 ps rise time pulses would be common.
I keep hoping that a member of the 118xx design team will wander by with schematics from a box in their garage.
Have Fun!
Reg
--------------------------------------------
toggle quoted message
Show quoted text
On Tue, 3/19/19, Jim Potter <jpotter@...> wrote:
Subject: Re: [TekScopes] Test input for an SD-32 sampling head
To:
[email protected] Date: Tuesday, March 19, 2019, 9:36 PM
Reg,
This sounds interesting. I have HFSS so I can
analyze the kinds of circuits you describe in the frequency
domain. It sounds like a high frequency peaking circuit. It
may be that HFSS or some part of the suite can do time
domain calculations. I don't know. I design rf linear
accelerators and waveguide systems for them in the frequency
domain.
Jim
