I'm going to take a stab at this, at the risk of possibly stepping on Gary's
toes.
1. RF amplifiers in general (not only solid state) don't *have* a 50 ohm
source impedance, they're (nominally) designed to work *into* a 50 ohm load.
The difference is subtle, but significant. Transmitters aren't classic
generators.
2. GE offered the matching network on station PA's for a number of reasons,
among them:
a) Amplifier circuit designs (solid-state or otherwise) have a finite
bandwidth; a tuning network allows for some output matching adjustment
b) Ideally the transmitter will be looking into a nice 50+j0 load (assuming
that's what it was designed for), but the world isn't perfect, hence the
adjustable output matching network to correct for *minor* load mis-match
(strong emphasis on minor)
c) Although not explictly described in GE's tuning procedures, significant
improvement in efficiency can be obtained with proper tuning of the
Z-matcher. Tuning for 50+j0 at the input to the Z-matcher is NOT
necessarily the RIGHT match!
d) To charge more. I'm half-joking on this; I can't say I've statistically
seen more or less failures on M2 PA's with or without the Z-matcher, so I'll
give this answer half a smiley: .-,
3. As far as Gary's comment about off-channel Z and its effect on
transmitters, some sub-par (or damaged) PA's will go spurious when looking
into a load that presents a bad match off-channel, even if it presents a
nice flat load on-channel. Some manufacturers suggest playing with cable
lengths to "tame" misbehaving PA's. Again, this is a shortcoming in the
PA, and I, for one, am not into band-aid fixes for design flaws or defective
equipment; I fix (or replace) the PA. When I walk off the site, I want to
KNOW the PA is going to be stable in the future as the load changes, because
it WILL change...
As far as "optimium power transfer", anyone that has passed their tech test
probably already knows the textbook answer to that question (the maximum
power theorem). But that's not really the issue here, is it OM? Again, we
have to accept the fact that amplifiers aren't classic generators; we can't
just look at the problem from the perspective of power transfer into a 50
ohm load. We have to look at the devices being used in the PA, the networks
doing the impedance transformations, the behavior of the amplifier as a
whole (including all cascaded gain stages), its behavior as voltages and
temperature are varied, and, one of the most important parameters,
efficiency.
Just to back up a step, let's revisit the textbook answer of "optimum power
transfer", which again, is based on a classis generator. In such a case,
the optimum power transfer is the *maximum* power that is received by the
load. Well, in our little RF corner of the power transfer world, it's not
that simple. We're not out eek the last watt out of our amplifier - that's
not the goal (or at least it shouldn't be). We all know we can sometimes
squeeze a fraction of a dB more out of an amplifier by purposefully
mis-loading it, but is that "a good thing"? Does that make it an "optimum"
match? Hell no. Among other things, we need to look at *efficiency*, and
plotting that against power output if we want to find the sweet spot.
Efficiency is a primary performance metric for RFPA matching, especially
when it comes to continuous-duty solid-state RFPA's where heat is your worst
enemy.
As far as SS VHF/UHF amplifiers go, good RFPA design should dictate that you
have adequate hardware headroom such that you're not stressing the devices
or any support components to make rated output, so "maximum power transfer"
should be the least of the worries for the tech tuning the equipment.
Stability and spectral purity should be a given in a properly-designed RFPA.
So the only parameter that should need to be monitored during
fine-adjustment at the output is really efficiency/current draw if
everything else was done right from the get-go.
--- Jeff WN3A
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of allan crites
> Sent: Friday, August 13, 2010 5:41 PM
> To: [email protected]
> Subject: Re: [Repeater-Builder] Re: Coax length, etc.
>
>
>
> Gary,
> Perhaps you can give us some examples to illustrate your thoughts.
> Perhaps you can also explain why GE chose to include a pi
> network on the output of the HB M-2 base xmtr to match the
> xmtr output to 50 Ohms, the shunt capacitor values and the
> series inductor value used.
> I'm interested to hear your explaination on how you would
> determine the length of cable needed.
> AC
>
> ________________________________
>
> From: Gary Schafer <[email protected]>
> To: [email protected]
> Sent: Fri, August 13, 2010 2:36:23 PM
> Subject: RE: [Repeater-Builder] Re: Coax length, etc.
>
>
>
> Hi Allan,
>
>
>
> Do we really care what the output impedance of the
> transmitter is? Most transmitters do not present a pure 50
> ohm output but are tuned to transfer maximum power into a 50
> ohm load. This often comes out to something way different
> than a 50 ohm source impedance.
>
> As the source impedance does not affect SWR the system
> doesn't care what it is as long as the transmitter can
> transfer maximum power into 50 ohms.
>
>
>
> What the transmitter does sometimes care about is the
> reflected impedance from the first cavity (being hi Q) that
> is not on frequency. This presents a highly reactive load to
> the transmitter that can sometimes cause the transmitter to
> overheat or reduce output. Placing a cable of a different
> length between output and the first cavity can sometimes
> change the unwanted off frequency load seen by the transmitter.
>
>
>
> 73
>
> Gary K4FMX
>
>
>
> ________________________________
>
> From: [email protected]
> [mailto:[email protected]] On Behalf Of allan crites
> Sent: Friday, August 13, 2010 12:56 PM
> To: [email protected]
> Subject: Re: [Repeater-Builder] Re: Coax length, etc.
>
>
>
>
>
>
>
>
> Nate,
>
> I have both the 12th and 14th edition of the ARRL Antenna
> books, the 12th I acquired in 1974 and have read and re-read
> the section on transmission lines and impedance matching
> probabily more than anyone else has. I sometimes learn new
> things with each re-reading, as there is much to be learned.
>
> In my discussions with Kevin Custer about the length of the
> transmission line connecting the xmtr output and the input to
> the duplexer, he suggested and I accepted, to colaborate on
> an article explaining the problems associated with matching
> the output impedance of a solid state transmitter of somewhat
> different than the normal 50 Ohms, and the attempts made by a
> manufacturer of duplexers to adapt (read match) the xmtr
> output via certain lengths of transmission line and
> readjustment of the tuning of the cavity closest to the xmtr
> output to effect this matching, ignoring the possible
> degradation resulting to the pass and notch characteristics.
>
> The transmitter in our discussions was the HB GE Mastr 2
> which, in the information available to me, appears to be
> having an output source impedance of 35+ or - (some unknown)
> reactance Ohms.
>
> Kevin commented that it appears that many hams are unaware
> of, or understand the methods needed, to do an appropriate
> job of impedance matching.
> Therefor we will be making this article for the benefit of
> those who don't understand the impedance matching necessary
> for optimum power transfer with a simple to understand way of
> impedance matching without the use of the infamous Smith
> Chart (which I have utilized for the past 50 yrs in all my
> impedance matching solutions and cannot be without).
>
> I agree that much information for impedance matching is
> contained in the ARRL Antenna Book but in my experience, real
> life adaptation of this information is and can be difficult
> to many hams.
>
> There is also another book I rely on and recommend, which is
> "Electronic Applications of the Smith Chart" by Philip Smith.
>
> Now, if you would like to contribute to our efforts I would
> gladly accept your contributions.
>
> Thanks for your input.
>
> Allan Crites wa9zzu
>
>
>
> ________________________________
>
> From: Sid <[email protected]>
> To: [email protected]
> Sent: Fri, August 13, 2010 10:38:25 AM
> Subject: [Repeater-Builder] Re: Coax length, etc.
>
>
>
> I have a note in my file that I do not recall where it came
> from relative to cable length between the duplexer and the TX
> or between the duplexer and additional filter. Length =
> (30)(32.785)(vf/freq).
> 30 is for 30 degrees, vf is velocity factor, freq is the
> average of the pass and reject frequencies. If too short add
> 180 degrees. Don't know if this is good info or not. The
> article would be appreciated. Sid.
>
>
> --- In [email protected]
> <mailto:Repeater-Builder%40yahoogroups.com> , Nate Duehr
> <n...@...> wrote:
> >
> >
> > On Aug 5, 2010, at 11:20 AM, Kevin Custer wrote:
> >
> > > Allan Crites and I are currently in discussion which will
> be used as the basis of a RB web article that will explain
> exactly what is happening, why it happens, and why an
> 'optimized' cable length can be used to transfer power ending
> up with the stated loss of the duplexer and have little
> reflected power toward the transmitter - so long as the
> duplexer is tuned properly and exhibits good return loss on
> the frequency it's designed to pass.
> >
> > There's already a great book on that topic, it's called the
> ARRL Antenna Handbook, and the chapter on transmission lines
> covers it in more detail than anyone will ever need to know
> in the real-world, who's not a practicing RF Engineer.
> >
> > That book if read cover-to-cover, is also damn good for
> insomnia. Or at least it'll keep you distracted while you
> can't sleep! :-)
> >
> > --
> > Nate Duehr
> > n...@...
> >
> > facebook.com/denverpilot
> > twitter.com/denverpilot
> >
>
>
>
>
>
>
