In article ,
Jim Lux wrote:

An ordinary pot hooked to the shaft probably won't do as a feedback
device. a) resolution issues, b) life... Typical panel pots have lives
in the million moves range, which seems like a lot, until you do
scanning 24/7 with a move every couple seconds.. say 20,000 moves/day
will burn through a million moves in less than 2 months. Their
resistance vs position will also change over life (you don't notice this
on a volume control). There are feedback pots designed for this kind of
duty that have much longer lives and have high quality resistive
elements and wipers, but they aren't cheap. You'd probably want to use
a shaft encoder (and incremental shaft encoders with 1000 pulses/rev are
easy to come by).

Any way you look at it, there's a fair amount of electromechanical
engineering that goes into making a reliable system like this (which is
why the commercial units are fairly expensive..It's not just because
they're connected to multi $100K plasma etchers.). In the classic
antenna tuner case, it's a bit easier, because you don't need high
quality initial position, just get close enough, and home in on the
optimum setting for the match.

Good points!

A couple of years ago, I was at a local hamfest/fleamarket, digging
through a big pile of relatively-new-looking industrial surplus. I
came across a nice air-variable capacitor, and asked the seller for a
price... got it for around $5, I think. He then said "Oh, if you like
those, I think there's one of them inside one of these things... that
one there." Metal case, N connector on one side, high-voltage
connector on the other, and a DB-25 for some sort of control connection.

I offered him $15 and he accepted.

After looking at it for a while, and doing some research, I've
concluded that it's just what you indicate - a matching unit for a
single-frequency RF plasma-etching system. It has what appears to be
a fairly spiffy SWR detector, and then an L-match network designed to
match a 50-ohm (I assume) RF feed to a lower-impedance load having
some amount of reactance. The shunt element is an air-variable cap.
The series element is another air-variable cap, in series with a hefty
silver-plated coil which feeds the high-voltage connector. These are
big cheese-cutter caps with widely-spaced plates. I get the feeling
that this matcher was handling quite a lot of power.

Both air-variable caps are turned via geared-down DC motors, and have
position-sensing pots. The motors, pots, and SWR detector are all
hooked to the DB-25, and I assume that in normal operation there was a
control console of some sort at the other end of the cable which had
the necessary control logic.

I haven't been able out a way to make real use of it yet, but it's a
glorious beast to look at, and I don't think I want to know what it
probably cost when it was new.

As to the original query - I suspect that an important part of the
practical engineering of an ATU of this sort, is enabling the digital
logic to co-exist well with the RF. Not only do you need to keep
birdies from the micro out of the RF path (by e.g. idling the
processor when not tuning), you also have to keep induced RF on the
control and sensing lines from sneaking into the digital logic and
causing it to either malfunction, crash entirely, or spew Magic Blue
Smoke out of its poor little fried-CMOS brains :-)

--
Dave Platt AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!