Smith chart? To the load or to generator question.

[mr ed]

I am taking it upon myself to learn something on microwaves so this is a Smith chart question by a newby..
I intend to design a frequency multiplier using just diodes. As I understand it two back to back diodes can work as a freq tripler (the 2nd harmonic is somewhat suppressed).

I can work the matching net from the diodes to the generator input using the 1/ 4 wavelength trick and a short z matching line in series. What to do for the output side? Do I consider the diodes as a new generator at 3x freq and work to the load?  Or from the load impendance to the generators?  There should be a 180 degree error if I get this wrong, I think.

The circuit would be a series of microstrips...

50 ohm in, 1/4  match, short z match to 2 diodes , z match at 3x freq , 1/4 match at 3x and 50 ohm out

[szoftveres]

Diode has varying characteristics based on the operating point (i.e. how hard it's being driven, ambient temps, etc..).
In practice, the driving signal is plenty, there's no concern about snr or too low signal levels - you just need good filtering in order to get rid of all unwanted components.
it's quite unusual to make a perfect matching network for a simple diode multiplier, as it doesn't provide any practical benefit, and severely restricts the operating conditions of the multiplier to a very certain input signal level.

[geggi1]

Google Charles Wenzel for his  publications on 3x and 5x multipliers.

[mr ed]

The diode operating point chosen is 0 to 10dbm where the dv/di curves are quite straight so the pure resistance is known.  Typical capacitance and package inductance are given but not locked down per diode so yes, it's a bit of an unknown. Some diodes have smith charts that wander around by frequency. I assume the output diode impendance is figured at 3x freq but my  question is still how to do a z match on the output?