> The length of coax doesn't effect impedance.
That statement is misleading, if not totally wrong. If there is anything
other than a perfect match at the load (in other words, if the VSWR on the
line is not a perfect 1:1), the coax behaves as a transformer. The
resulting Z, as measured at the source end of the line (i.e. the transmitter
end) WILL vary depending on the length of the coax.
Don't confuse Z with VSWR. The two are related, but you can't use them
interchangably.
> Now if you put a voltage null at your transmitter, what would
> happen? Normally with high SWR your transmitter will get hot
> because its dissipating the reflected power into its
> heatsink.
No, that's not right. Reflected power doesn't get dissipated into the
heatsink, unless the transmitter happens to have an isolator on its output,
in which case most of the reflected power ends up in the reject load which
sometimes is mounted to the same heatsink as the devices, but that's really
stretching...
A mismatch between the transmitter and the load can result in lots of
things. It may cause the amplifier's efficiency to degrade as the devices
are no longer able to transfer power to the load efficiently, but that's
only one possibility. In some cases, it might actually cause the
transmitter to make more power, sometimes at higher or lower efficiency,
depending on many factors related to the design of the amplifier itself. If
the matching network (if one xists) was adjusted to properly match the PA to
the mis-matched load, all of the power (minus transmission line losses) will
ultimately get to the antenna.
> The only time coax length makes a difference to power out is
> if your using it in a matching stub, or a matching section
> ie. if you take 1/4 wave of 75 ohm cable put it on the end of
> 50 ohm cable you will get a match with a 112.5 ohm load.
Coax itself doesn't affect the power output. The LOAD IMPEDANCE that
terminates the transmitter, and how the transmitter is (or isn't) matched to
that load is what affects the power output in real-world transmitters.
Varying the length of the coax MAY result in a change in the transmitter's
power output when the VSWR is not 1:1 because the coax acts as a
transformer. It transforms the antenna's mis-matched feedpoint Z to some
other Z at the far (transmitter) end.
Since a perfect 50+j0 match is practically impossible to achieve, coax (or
any other form of transmission line) will virtually ALWAYS act as a
transformer except at exact half-wavelength multiples.
> You make an interesting point though, why does the cabling of
> duplexer's need to be a certain length.
If you're talking about the cables within the duplexer harness, the reason
is to properly repeat and transform impedances. The simplest example I can
give you is to consider what happens at the antenna tee of a duplexer. At
the transmit frequency, one side of the tee feeds the transmit side of the
duplexer which presents a good match. Looking toward the opposite side of
the tee, the receive side of the duplexer typically presents a SHORT at the
transmit frequency (typical of a pass/reject cavity). By offsetting that
SHORT by a 1/4 wavelength section of cable going to the tee, the SHORT gets
transformed to an OPEN, therefore no power can flow in that undesired
direction.
Maybe it's time for somebody to do a "Smith Chart 101" article. No, I'm not
volunteering myself :-)
--- Jeff
