Input Matching Network for modulated square signal

[pYrana13]

Hi all,

What type of matching should I use for input matching network at transistor gate? The signal that drive the gate is a modulated square wave, which mean looking from spectrum is a few harmonic around 1 GHz frequency(between 0.925GHz-1.075GHz). If I match using stub technique, the reflection I got from S11 only effectively suppress 1GHz signal(-15dB suppression), whereas the sideband of 0.975GHz and 1.025GHz got only -5dB suppression which also lead to reflection problem.(see attachment)

How could I widen the reflection dip with good suppression? Any particular technique anyone?

Thank you!

 

[FvM]

Generally speaking, multi LC networks or transmission line transformers are good for wide band matching. Polyfet application notes can give some good hints for the latter.

I don't even try to understand the nature of the "modulated square wave", because the spectrum is far from looking plausible at first sight.

 

[mtwieg]

I don't even try to understand the nature of the "modulated square wave", because the spectrum is far from looking plausible at first sight.

Looks like an ASK modulated carrier, with an unsuppressed carrier.

 

[biff44]

I am not sure at all what u are trying to say. Do you have a 1 Ghz square wave hitting the gate? Exactly HOW is it modulated?

In general, stub matching will not work for square waves, due to the very large bandwidth needed in the matching network (at least up to the 5th harmonic)