开云体育

Something that I have never fully understood is that S21 is supposed to be
measured using a "matched load".

Many VNA receiver ports do NOT precisely match 50 Ohms;
inserting a precision attenuator before CH1 improves matching.

Isn't the "matched load" of a VNA equal to 50 ohms?

nanoVNA bridge and ICs are nominally designed for 50 Ohms;
I don't find a nanoVNA specification for CH1 return.

Are we to assume that 50 ohms is the only possible "matched load"?
Or can a VNA be calibrated for any value of "matched load"?

Software can "calibrate" for other loads,
but e.g. S11 sensitivity is not improved.

For two-port measurements without precision attenuator at CH1,
nanoVNA firmware does NOT compensate S11 for CH1 mismatch (see attached).
Consequently, S11 measurement for 2-port DUT
is ideally terminated by precision load or equivalent, rather than CH1, and
use only S21 measurements for DUT terminated by CH1
while appreciating that S21 are somewhat impacted by CH1 mismatch.

Warren Allgyer's 1-3000 MHz RF bridge method compensates CH1 return..
/g/nanovna-users/message/2484