Tim,
You stated the problem quite well, even for a Saturday morning!
Since power transfer between RF circuits is maximum when the source and
load impedances are matched, that happy state is what we seek to
achieve. If an RF power amplifier's output impedance was a stable 50
ohms at all drive levels, we would not be having a problem with
impedance matching and cable lengths.
Solid-state PAs, especially those at the low end of the price scale, are
notorious for unstable output impedances. They seem to work great when
feeding a nice 50 ohm resistive load, but they sometimes go bonkers when
presented with a highly reactive load- such as a duplexer TX cavity.
Tube-type PAs, such as many of those made by GE, are far more docile in
this regard because they have tuned output stages.
The tuning instructions for most commercial duplexers invariably call
for the use of a network analyzer or a high-end spectrum analyzer with a
return-loss bridge. These instruments present a precise 50 ohm source
to the cavity under test, so that the bandpass peaks and notch depths
are adjusted to present a 50 ohm load to the transmitter at the precise
frequencies of interest. Remember that cavities are resonant circuits
by definition, so they should NEVER be detuned once they have been
properly adjusted.
Okay, we use high-quality 50 ohm coaxial cable, such as RG-400 or
RG-214, to connect our 50 ohm cavity to the power amplifier- and we find
that the power loss through the duplexer is far more than it should be.
The correct action is to adjust the PA so that it 50 ohms. If that is
not possible, perhaps because the PA is a broadband solid-state unit,
then the next best action is to transform its output impedance to 50
ohms. One way to do this is to create an impedance transformer of sorts
by adjusting the cable length of the TX-to-duplexer jumper, or by
installing an impedance matching device (Z-matcher) at the PA output
connector. In any case, the duplexer cavity should NOT be detuned to
fix this mismatch.
Some solid-state PAs settle down if a ferrite circulator is installed at
the output. While a circulator is often provided as a means to present
the PA with a constant load impedance, it also serves to shunt incoming
RF to a dummy load, thereby greatly minimizing the possibility of
intermodulation. When a circulator is paired with a dummy load, many
manufacturers call it an isolator. Most high-tier transmitters being
made today have at least one circulator built-in to the PA, especially
UHF stations. A ferrite circulator will generate a second harmonic,
since it is a nonlinear device, so there must be a low-pass or notch
filter following the circulator.
Finally, be aware that changing the drive level on a solid-state PA will
usually change its output impedance, thereby upsetting the impedance
balance achieved earlier. This must be considered on repeater systems
that switch to a lower power level when operating on backup batteries.
73, Eric Lemmon WB6FLY
"Tim S." wrote:
>
> So from what I am getting here.
>
> Only some PA's are affected by the cable length. Mine happens to be a
> Master II Tband. It's only trial and error to discover it the problem. It's
> trial and error to make the cable length right so the PA is happy.
>
> If I understand this right, what cause the problem is the PA's output is not
> a true 50 ohms by failure or design. This causes the 50 ohm coax to not be
> matched to transfer the power from the PA to the duplexer.
>
> So you alter the cable length to make it match the output of the PA
> impedance. Which in turn helps it to transfer the power to the duplexer.
>
> Why does this work? Isn't the input to the duplexer also 50 ohms?
>
> Too much thinking for a Saturday morning!
>
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/Repeater-Builder/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
