G8WRB asked:
>I'm contemplating making the tuning automatic. Does anyone have any
>ideas of how it can be done ? I faintly remember seeing something about
>adjusting the tuning until the anode impedance is completely real (i.e.
>at resonance)...
>One possibility is to store previous locations of the stepper motors in
>an EPROM/EEPROM, but I don't like that idea...
>Are there any commercail VHF (or even HF) amp that do automatic tuning ?
>My guess is there are. Any thoughts appreciated...
Geurts, Gerard (PA3DQW/AA3ES) wrote:
> In theory it is easy. If the triode is in grounded grid, the tube has
> a purely resistive load when the cathode and plate voltages are exactly
> 180 degrees out of phase. This takes care of output capacitance etc.
> as well. You need a way to detect the phase difference between the
> cathode and the plate voltages and use this to drive a servo controlling
> 'tune'.
>I used to have a copy of Pappenfuss' 'SSB principles and circuits' (I
>think it was called). He described circuits using phase detectors and
>feedback loops controlling servos. This book was written in the 50ties,
>so computer control was not an option yet. Unfortunately my copy was
>never seen again after I lent it to somebody. You might be able to find
>a copy in your university's library.
G8WRB conculded:
>Thanks for your comments on automatic tuning. I assume a computer could
>be programed to do it from just the RF output, grid current and anode
>current - after all, thats all we use.
FOR AMPLIFIER AUTOTUNING (FEEDBACK CIRCUITS):
Yes, the text you want to take a look at is "SSB Circuits and Systems",
1964, by Pappenfuss, Bruene, and Schoenike, (all of Collins Radio). This
classic book has several chapters covering linear amplifiers, with special
emphasis on tetrodes with distortion reduction techniques. The autotuning
technique of using samples of the grid tank current and plate voltage to
drive a servomotor is discussed on page 166.
It is stated there that you could also look at peaking screen current, to
indicate maximum load impedance on the tube(s). This would indicate also
maximum plate voltage swing. But low Q output circuits may exhibit errors
of 10 - 20 degrees of phase shift off resonance. Also, tubes that have
secondary emission on the screen can run negative current (we have all seen
this!!), and would make such circuitry useless.
The preferred method to adjust a resonance tuning capacitor (or inductor)
according to the authors is to use phase detectors made up of diodes in a
bridge. To also autotune the loading of the amplifer, one must assume
linear tube operation, and set another bridge with diodes to detect the
ratio of ep to eg. In this case, the fixed relationship is set (proper
output impedance value = correct loading) by a servomotor driven from this
bridge, which controls the loading control (output capacitor in a Pi
network).
On page 217 is one such autotuned amplifer, a 2 - 30 Mhz two stage tetrode
feedback amplifier with (3) 4CX250B family tubes in parallel, driven by (2)
6CL6's. Collins Radio sampled the grid RF voltage and the plate voltage. In
a properly tuned amplifier, these are 180 degrees out of phase. On page 219
is an even more complex beast, a three stage 45 KW amplifier, again
autotuned throughout using the same principles.
Some other references on autotuning:
V. DeLong, "Automatic Tuning Techniques for SSB Equipment, Proc. IRE, pages
1766-1774,Dec. 1956.
V. DeLong, "Automatic Tuning for High Power Transmitter, Electronics, pages
134-137, July 1956.
You do NOT want to look for plate current dip, with a computer control. For
one thing, it is not level independent. If the drive level changes, then it
looks like a dip might be occuring. Plus, you don't want your servos to be
starting from off-resonance as you power up, and having to track across and
find the dip. (quick way to make smoke and noise) I have done the converse
of this using an HP Basic series 9000 workstation once, driving a sweep
oscillator which fed power into a resonant cavity. I wanted to know the Q
automatically for a dielectric measuring automatic tester. The algorithms
must look at the slope of the peak (or dip) and determine when you are
parked on the top of the curve (or bottom). Nowadays, Hewlett Packard has
built-in firmware to read Q and bandwidth automatically in their network
analyzers. These computer controlled algorithms are good when there is only
one distinct resonance in the frequency span being measured, and when no
noise or stray modes (resonances) show up on the side.
Grid current peaking is a simple way to adjust an input circuit to a power
amplifier, again we must be careful of noise, and allowing the amplifier to
run across a range of input tuning to find this.
FOR ANTENNA MATCHER AUTOTUNING (FEEDBACK):
The later edition of "SSB Systems and Circuits" by Sabin and Schoenike
(Collins Radio) is probably still available from McGraw Hill Inc. Mine is
dated 1987. Since most transmitters are solid state by that time, the
emphasis is not so much on autotuning in power amplifiers, but in antenna
couplers. Chapter 16 discusses various automatic techniques, with HF
discriminators for measuring the phase and loading. All HF practicioners
planning to develop autotuners own themselves a look at this chapter.
Reference:
H. Landt, "Tracking Impedance Measuring System", US Patent 4,506,209. 1985.
PRESET AUTOTUNING METHODS:
Then there are the techniques using look up tables, as you mentioned. Of
course, my ICOM AT500 does this with little preset pots. And a number of
other amateur and pro manufacturers use variants on this idea.
It's cheap and easy, as long as the load never changes.
Delta Electronics made an HF tuner (some of you may have them - the Delta
HFAS-4) for 10 KW PEP, to load a long whip or wire. It used preset pots
also, on little plug in cards. When a station changed freq., the card was
swapped out, or switched out, to a different set of pots, with
corresponding different servomotor positions on a large inductor and vacuum
variable capacitor. These types of systems, as you would guess, are only
useful for an antenna or TX with a constant load. No bird nests or icing
problems please. And would require adjustment for different frequencies
also, although a straight line interpolation is usually possible, and can
be automated, to track the frequency of the exciter. Using a look up table
in EEPROM is the modern day analogy of the analog memories of 20 years ago
(potentiometers).
CONCLUSION:
I personally would use the phase feedback methods if I had to autotune.
BUT, I like to manually fiddle with the tuning and loading on power
amplifiers, as long as it's not too early in the AM, and I have had my
coffee. It sort of gives me a feel for the thing, and for any abnormal
conditions, such as the quick surge as you tune across some parasitic
oscillation, or the crackle of a melting fingerstock followed by a loud
bang. The first time I tuned the plate tuner (sliding short) in our 3 MW
200 MHz triode amplifiers, while operating near plate dissipation limits,
was an unnerving experience. Took some getting used to: Tap the motor drive
switch, watch three meters and scopes with only two eyeballs
simulataneously, and then realize it went the wrong direction! I'm sure the
folks at VOA and other high power TX sites had similar worries before
autotuners. When your bottle costs over $100K, you don't want to see it
burst in smoke and noise!
John
K5PRO
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|