Remote tuner
 

On Nov 2, 8:17*pm, Jim Lux wrote:
Art Unwin wrote:
On Nov 1, 7:59 pm, Jim Lux wrote:
Art Unwin wrote:
On Nov 1, 12:34 pm, you wrote:
In article
,
*Art Unwin wrote:
Browsing thru E bay I came across a 1kw
antenna tuner for sale from Cyprus.Now in my younger days I got hold
of one of these and intended to salvage a Icom AH2 to somehow make it
remote since I use Icom radios. After all there are only three motors
that have to be driven! Now I am retired I pulled my old one from
storage after seeing the same on E bay. Question is are there any
kits, salvage ideas out there that would get me going on this long
lost project?
I have never used a external tuner but this seems like a realistic
project to play with.
Of course if there is a unit out there that I can salvage a controller
that is all the better. There are very good photos of this particular
tuner on E bay but like mine it has no controller! I suppose I could
feed a 10 watt signal to it and control the motors by hand to tune but
I do like the idea of automation.
Regards
Art
It isn't the tuner hardware that is going to be a problem, as much as
designing the tuning algorithm that drives the hardware. Mark Johnson
did a very good job, reverse engineering the stuff that Bill Shield
brought west from Motorola Military back in the 70's for the SEA
Autotuners. SGC copied the SEA Firmware & Design enbank, for their
stuff. The best stuff, was the stuff Collins did for their Military
Autotuners. Starting from scratch is going to be a BIG design project.

Remote tuner
 

On Tue, 02 Nov 2010 18:17:10 -0700, Jim Lux
wrote:

Art Unwin wrote:
[snip]

An L network match either RloadRgen or RloadRgen, depending on which
side the shunt element is on.

In autotuners using an L network (e.g. the LDG boxes), they have a relay
that switches the shunt C to either end of the series L.

In autotuners using a Pi network (e.g. the SGC boxes), there's shunt
elements on both ends, so you can match anything (subject to the
component ranges).

So if you have a motorized C and a motorized L, you'd need to add a
relay to switch the configuration to be able to "match anything"

By the way, a L or Pi network cannot match Rgen=Rload, but you can do it
trivially setting both C's to zero and the L to zero. It's not really an
L or Pi (or T) network then, which is why some automated network
calculators choke on it. A real autotuner doesn't care, because it
almost always starts in the "bypass" position anyway.

That is a great post!

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.
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.

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

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.

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...
John Ferrell W8CCW

Remote tuner
 

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

Remote tuner
 

On Nov 3, 4:04*pm, Owen Duffy wrote:
John Ferrell wrote :

...

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

Owen
Most radios have a tuner inbuilt or as an add on.
These types are ripe to uprate the components for high power for
placing at the antenna whilst keeping all the controls at the
operating position. This way you keep the lower power timing and other
tuner controls in position as well as the memories and of course band
changes so you don't have to reinvent the wheel or use skills that you
don.t have

Remote tuner
 

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

Remote tuner
 

An L network match either RloadRgen or RloadRgen, depending on which
side the shunt element is on.


It's not quite that simple. For example, a load of 5+j30 requires the
same L-match configuration to match to 50 Ohms as does 500+j30: shunt
C at the Load end in both cases.

Steve G3TXQ

Remote tuner
 

On Nov 4, 10:09*pm, steveeh131047 wrote:
An L network match either RloadRgen or RloadRgen, depending on which
side the shunt element is on.

It's not quite that simple. For example, a load of 5+j30 requires the
same L-match configuration to match to 50 Ohms as does 500+j30: shunt
C at the Load end in both cases.

Steve G3TXQ

Conversely, a load of 10+j22 needs the shunt C across the load,
whereas 10+j18 needs shunt C across the generator.

Steve G3TXQ