> I never suggested that the cable between the TX and the
> duplexer had to be an exact length and if the TX output is a
> true 50 ohms then any cable length produces no consequence at
> the cavity input. But if the TX impedance is not 50 ohms, I
> think any cable length other than half wave will make the
> cable a line transformer affecting the impedance presented
> to the cavity with a quarter wave producing the greatest
> impedance shift. That's not necessarily a bad thing because
> the impedance match of a random length cable is statistically
> as likely to improve the TX/cavity match as it is to degrade
> it.
Exactly!!!
> Diddling around with this proved to be so frustrating
> that I eventually caved in and installed a Z-matcher at the
> TX output and tuned it for maximum cavity output at the pass
> frequency. It seemed (and seems) to me that this is a cleaner
> way to deal with the matching issue than trying to adjust
> BP/BR cavities with an input other than 50 ohms.
If you want to use a Z-matcher to make your transmitter happy, you really
need to monitor current draw when adjusting it in addition to power
measurements. Just because you can install a Z-matcher and diddle with its
caps to get a fraction of a dB more apparent power out doesn't indicate it's
properly tuned. It could just be that the transmitter makes a little more
power when looking into a load that's not 50 ohms, but at the possible
expense of higher DC current draw, lower efficiency, and increased RF
currents, all of which can be detrimental to the longevity of the PA. Also,
apparent increases in TPO may be due to the generation of spurious products
which are more likely when the transmitter is looking into a badly-matched
load, which a Z-matcher is quite capable of creating. A spectrum analyzer
would help tell the whole story (cleanliness is next to godliness when it
comes to a mountaintop transmitter!).
I'd rather see an isolator between the PA and duplexer, and if the PA is
still not happy with that, fix/replace the PA.
Also, when you're using a Z matcher at the output of the PA, the load Z
(duplexer/feedline/antenna/etc.) is the other half of the equation. The Z
of the antenna system is anything but constant. Electrical cable lengths
(phase) vary with changes in temperature, thereby rotating the antenna
feedpoint Z around the Smith chart. The antenna feedpoint Z is being
transformed to some other Z at the end of the line, so the Z you are
matching into is going to change as the electrical length of the line
changes. As a practical example, a 200 foot length of Heliax on UHF has
about a 20 degree phase change between 0 and 100 degrees F, and Heliax is
much more phase-stable than most other cable types (especially solid
polyethylene dielectric cables). The antenna feedpoint Z itself will also
change with humidity/rain/icing, etc. Bottom line - matching to an antenna
system is always unpredictable when done at the far end of the feedline.
What looks good on the day you do the tuning may look ugly six months from
now.
It is for these reasons that I have always been against makeshift matching
techinques, including "cable length games", to make a system perform
properly, for the performance improvements you think you're getting are
likely to change due to factors you have little or no control over. I'd
rather have everything on the ground kept as close to 50+j0 as possible,
leaving only the antenna feedpoint Z as the variable you have no control
over. And if the antenna Z is poor under "normal" weather conditions, it
needs to be replaced.
> You mentioned that the power adjustment levels would not
> affect TX output significantly on an M II or Micor. I'm not
> familiar with the power control network they employ but does
> the collector voltage on the output devices remain constant
> during power changes? If not, how do you avoid significant
> output impedance shifts?
On both M2 and Micor, the collector voltage of one of the early PA stages is
varied to adjust the net gain (output) of the amplifier. The finals always
have full 13.8V on the collectors (or emitters, in the case of PNP highband
Micor transistors).
--- Jeff
