> OK so you have a reel of cable and two connectors to make up
> the jumper between transmitter and duplexer.
>
> The duplexer is tuned using 50 ohm test gear and the
> transmitter has been optimised into a 50 ohm load.
>
> Unfortunatley the output impedance of the transmitter is not
> 50 ohms and a length of cable to the duplexer will transform
> this impedance,detuning the first can in the transmit chain.
No, it won't, you're putting the cart before the horse. The duplexer and
antenna system is the load. You can't change the Z of the load by changing
the Z of the source.
Think about it this way. Get rid of your duplexer and antenna and replace
them with a fixed resistor, pick a random value, say 1000 ohms, and use a
"perfect" half-wave cable between the resistor and your PA. Now use three
different amplifiers to transmit into this 1000 ohm load. Does each PA have
the same source Z? No. Has your load Z changed? No. Has the Z at the
input to the half-wave cable changed? No. No matter what you do at the
source end, the load Z is 1000 ohms and will remain that way for ever and
ever Amen.
Let's look at it another way using a practical example. Say you have a
single bandpass cavity filter that is properly tuned to present 50+j0 at its
input port when its output port is terminated with a 50+j0 load. Now you
hook it up to a PA that has a strange source impedence using a half-wave
cable with a Bird wattmeter somewhere in the middle (its location in terms
of distance from either end is immaterial). Will the Bird show any
reflected power? Of course not; the Z on the transmission line is still 50
ohms, the E and I are still in-phase at a 50:1 amplitude ratio, the cavity
hasn't been "detuned" at all. The Z on the line is 50+j0, the Z at the
input to the cavity is 50+j0, the Z of the load Z is 50+j0.
Now, if you look in the REVERSE direction, FROM the duplexer INTO the
transmitter, as you change PA's the Z looking the other way is going to
vary, but that's not what we're trying to match here. If we want to go down
that road, I would argue just how BAD a Z-matcher could potentially be when
used on the transmit side of a duplexer in terms of how it can throw off the
termination Z and mess up the RECEIVE side performance. Ponder that for a
bit (hint: what should be an "open" at the antenna tee no longer is).
A Z-matcher is no substitute for an isolator.
--- Jeff
