On Apr 12, 3:35 am, "Keith Dysart" wrote:


Just for clarity, an example problem that has been previously
analysed is the following:
A generator with a 450 Ohm source impedance drives a 450 Ohm
ideal transmission line terminated in 75 Ohms. What is the
magnitude of the re-reflected wave at the generator?

Interesting to me that CM railed at me for complaining that his
trivial trumped-up non-real-world example wasn't worth considering,
and now he's unwilling to accept an example that IS quite realizable.
I can EASILY do it on my bench, though I'd prefer to use a 50 ohm
generator and 50 ohm line, and a 300 ohm load at the end of the line.
It's certainly no more difficult to get a known output source
impedance than it is to get a lossless transmission line, but we talk
about lossless lines all the time, to remove an unnecessary
complication from our discussions. Since we have to test our
generators to insure that they really are 50 ohm sources, I feel quite
confident that their source impedance doesn't depend on the load you
put on them.

With respect to the output impedance of a transmitter, it may indeed
depend on the tuning/loading of the transmitter's PA, likely even the
power level it's running, and _maybe_ even on the load impedance you
put on the transmitter's output terminals. But what I do doubt is
that it _changes_ for a given setup including a given load impedance,
and in steady state, for sure the load impedance isn't changing. So
in steady state, can we determine if there are or are not reflections
on a transmission line connected to the transmitter, at that
interface? If you can't, how much do you have to disturb steady state
to make that measurement?

If battles have raged on for years about the output source impedance
of a transmitter, I submit that the people making the measurements
either (a) don't understand what they are doing, or (b) have not fully
specified the conditions under which they made the measurement, or
both. I assume they would't battle if they agreed they measured
different values, but that the conditions were also different, but
that may not be a valid assumption--some are known to battle
regardless.

As I mentioned, we have to be very careful about input and output port
impedances in measurement equipment. It's not necessarily an easy
thing to get "right," but I'm confident that we've been doing it
right, and probably for longer than the "debate" over transmitter
output source impedance has been going on.

If you don't believe there's a solution to the example Keith posted,
you have no right to believe in the results of a measurement with a
vector network analyzer, and you should certainly not trust the
indicated output level of any signal generator.

Cheers,
Tom