On Wed, 03 Nov 2010 21:04:04 GMT, Owen Duffy wrote:

John Ferrell wrote in
:

...
I am slowly working towards a motor driven remote T-Match roller
tuner. I intend to use the same algorithm that I use to operate a
manual tuner.

Then you have a considerable challenge.

That is to start with the capacitors at center, the inductor at
minimum. Run the inductor down until the minimum SWR and center it
there. Then do the same with the transmitter side capacitor. Finally,
do the same with the antenna side capacitor. The controlling SWR
signal is from the transmitter point.

If you want to automate this, you really need to know more than just the
SWR (or magnitude of Gamma, the complex reflection coefficient), or a
related signal (such as 'relected power').

Have a look at some practical implementations, and see how they work.

You have a three variable problem to solve, and one input value (using
your proposed method).


I am concerned about making the transmitter happy as quick as

Ah, 'happy' transmitters, as in the ATU's purpose in life is to make the
transmitter happy.

possible. I expect that this is not be the optimum match but it will
serve the purpose of making the transmitter happy.

One of the design decisions you should make is how close to perfection is
the target 'match', and what do you do if it changes during operation.

Can you determine an optimum match (meaning VSWR=1, efficiency=max) using
SWR?


I start with the inductor because the losses are greater there so less
inductance is likely to have lower losses. I had overlooked that I do
try shunting the tuner first.

In the real world, it gets a little more complicated that the classic
Rules of Thumb (RoT). Inductor Q is not constant with varying L,
capacitor Q (or D) is significant if ceramic caps are used (as in many
autotuners).


After having explained that I realize there is probably a better way.
I will try starting with Max transmitter side C, Min Inductor and
Minimum antenna side C.

I think the ideal procedure would be to calculate the needed match,
build it to minimum loss (in the tuner) and switch it in. I will have
to think about that...

If it were me, I would create a software simulation of the algorithm, and
include real physical times in the simulation. See if your algorithm
converges on a stable and correct solution, and then, how quickly it does
it.

Owen
It looks like a "not-to-complicated" problem to me. I hit a wall
somewhere along the way, I will have to give it a try to find out.

I am not seeking a perfect solution, just a homebrew project to solve
a problem. I never thought I would own an HF amplifier but now I do. I
am accustomed to the convenience of an SGC-237 tuner in the shed next
to the vertical antenna. Too much power for the tuner so I have slowly
acquired the parts to construct the High power tuner project.

Not a Bread board but more like a cabinet door prototype. After all,
it will be on a shelf in the shed. I am thinking a PIC micro
controller (I have tools & Experience) and electric screwdriver
actuators.

Ugly Construction is beautiful IF IT WORKS!

BTW, it is one of many low priority projects I have...

John Ferrell W8CCW