#21 incorrect s-para simulation when Zo Not=50

[Anonymous]

I found that there are some errors in the simulated s-
parameters when Zo Not= 50. E.g., port 1 is 50, port 2
is 100, and they are connected with a quarter-wave tx
line with Zo=70.7 ohm. The simulated S21 in this case
at the center frequency should be 0 dB (no loss). But
the program gives the db(s21) as -3 dB.

lyokecho@dso.org.sg

[Nobody/Anonymous]

Logged In: NO

Even worse, sometimes a circuit with no active elements can
show s21 of greater than 0 dB!

[Wojciech Kazubski]

Logged In: YES
user_id=286510

The transmited wave at the circuit ports have to be
normalized by a square root of the port impedance ratio (Zin
to Zout). This is to make power levels agree at both ports.
If it is not done the circuit may show fake gain or loss
depending which impedance is greater. S12 should be +3dB for
the example circuit.