On 2 mayo, 19:38, Cecil Moore wrote:
On May 2, 9:57*am, Wimpie wrote:

Assuming that your black box is a 100 Ohms quarter-wave line with 200
Ohms termination (all inside the black box), S11 (50 Ohms based) = 0,....

This is what happens when one changes math models in mid-stream. s11
is NOT zero at the mouth of a stub. The impedance looking into a stub
is IEEE definition (1)(B) and cannot have an s11 of zero unless the Z0
of the stub is infinite, which it is not.

From your text I didn't conlude it is a stub (but just a series line
section that transforms 50 OHms into 200 Ohms).

Thanks for proving my point. The single-port s11 is completely
different from the dual-port s11 and that is most likely what is
happening with your attempts to measure the source impedance of an RF
amplifier. How can you possibly promote an experimental approach where
s11 changes by an infinite percentage depending on whether one
measures it as a single-port parameter vs a dual-port parameter???

Please explain (or someone else), as I don't understand anything of
the above with regards to a PA. You described a single-port device and
now starts talking about a two-port device.

If you want to prove that for a two port device the impedance
corresponding to S11 may not be equal to the input or output impedance
of port 1, you are right. When S12*S21 isn't zero, the input
impedance depends on the termination of port 2. This is just S-
parameter math, nothing magic.

For the PA case, the active device is the termination for the two-port
matching network, so in that case S11 measurement equals the input/
output impedance of the PA (just a single port measurement). See it as
the output impedance of the active device is connected to port 2 (of
the matching network), and the VNA (or load) is connected to port 1
(that is the output side of the matching network).

The PA reduces to just a single-port network with a source in it
(compare it with the antenna example where a transmitter is in the
vicinity). As long as the injected signal (or slight mismatch
connected to transmission line with increasing length) is small, you
can apply the small signal approach.

As suggested earlier, you may dive into active load pulling:
http://www.focus-microwaves.com/template.php?unique=232

It doesn't matter whether you show some mismatch to a PA, or a perfect
matched load with a small source in it representing the signal that is
reflected towards the amplifier.

--
73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

Kind regards,


Wim
PA3DJS
www.tetech.nl