> But why? If all of the power (or, let's hope, at least
> 99.999999% of it)
> > is
> > on-channel, *should* a properly-designed and properly-functioning
> > transmitter misbehave due to the poor match a duplexer presents at
> > frequencies far removed from the channel center?
>
> Well yes, properly designed transmitter. But how much do you
> want to pay for
> it?
Me personally? I'll pay for a transmitter that works, and works right.
The way I see it, repeaters are like cars. You have to get your car
inspected for safety. Your car doesn't pass safety inspection? You can't
drive it on the public roads, lest you'd be putting other people at risk.
Same with a repeater transmitter. If it's unstable and has the potential
for causing interference other systems (ham repeaters, public safety,
aviation, etc.), it shouldn't be on the air. Either fix it, or if you can't
afford to fix it, take it down. I don't want some scmuck driving a beat-up
1972 pickup down the interstate in front of me and having his rear bumper
fall off any more than I want somebody putting some clunker up on a
mountantop and having it go spurious and interfering with EMS or ATC. That's
just the way I see it, sorry if that rubs some people the wrong way.
> A built in isolator will solve all of those problems as
> an example.
Maybe. An isolator will help flatten the load on, and around, the carrier
frequency, but isolators, too, have a finite VSWR bandwidth, they won't
provide a perfect load across the entire spectrum. And if you can afford an
isolator, you can probably afford a better PA.
> It is almost impossible for a high Q cavity to not present
> some reactance
> away from the tuned frequency.
It's not almost impossible, it's definately impossible.
> If it didn't then it would not have any
> selectivity.
Right.
> The random length cable of course transforms
> that reactance to
> something that the transmitter may or may not be comfortable with as
> discussed above.
Just to clarify, the complex Z is being transformed (both R and jX), not
just the reactive component.
The thing with random-length cables is just that - they're random. How do
we know what cable length is going to make the transmitter happy? Does the
transmitter like more XL or more XC, or bigger R's or smaller R's, and at
what frequency, because as I'm sure you know, the complex Z is going vary
wildly at different frequencies, due to the duplexer's Z, its behavior as a
transformer with respect to the load Z at the antenna port, the antenna
feedline acting as a transformer with respect to the antenna feedpoint Z,
and the cable between the PA and the duplexer also acting as a transformer,
so you end up with this complex system of cascaded transformers. Chances
are if the PA is that picky, its behavior may also change with temperature,
voltage, who knows what else.
Antenna feedpoint Z's change with environmental conditions (precipitation,
icing, etc.). Feedline electrical lengths (phase) change with temperature,
so the resulting Z at the duplexer antenna port is also going to change.
There are *so many variables* that will constantly be changing over time
that what may seem to work when you walk off the site may fail miserably
months, days, maybe even hours later after you think you've found that magic
cable length. At least with an isolator we've taken the bulk of those
external variables out of the equation - I can agree with that. But, call
me a fundamentalist, I still believe that a PA should work, and work right,
when it sees 50 ohms on-channel no matter what's happening off-channel.
--- Jeff WN3A
