Yep. All understood mate. Thanks.
I'll forward this immediately to John and I've just arranged to be at
his place at 9am tomorrow (Tuesday 01:00UTC)
Yes, he did the VNA looking back towards the resistor before he ever
applied power. Also did the input tune like that. It got everything in
the ball park.
I'll stay in touch. Your ideas and assistance is invaluable.
Cheers, Alek.
On 3/10/2022 3:46 pm, Steve wrote:
Cc and C1 is what the two variable caps are labelled in the original
circuit, there's no special meaning beyond that. Cc is the series
coupling from the output, C1 the shunt capacitor.
I'd say it's misleading to call them 'tune' and 'load'. In a typical
hf pi network the load cap predominantly sets the wanted load
resistance and the tune cap adjusts the load to be purely resistive
and (to a very loose first approximation) each does their job
independently of the other. In the amp you're working on both caps
affect both functions.
Think of the wanted load at the anode as a point on graph paper. In a
pi network 'load' cap moves your position left and right, 'tune' cap
moves you up and down. In your amp, one cap moves you on a slant / and
the other on a slant \ so it's a much more iterative process to find
the best tuning spot.
The original amp gives figures for running at 2.5kV and 1.2A - that's
a much lower load at the GS35 than you would target for 3.3kV and 0.8A.
Don't forget that connections around the blocking caps contribute to
the total series L in the circuit. Can you can patch in a short leaded
resistor of the desired load from anode to ground and look into the
output with a VNA to help preset the tuning on the amp when cold?
Good luck with the experiments, play safely.
73, Steve G8GSQ
OK Steve. Thanks for your valuable input.
Firstly, the loading adjust is very simple. Both the tune and load
capacitors are simply shhets of copper. The tune plate is wound in
and out by a threaded rod. The load plate is fixed at a 1/2" spacing.
John added another threaded rod to enable that plate to be to be
pushed in, to give some adjustment capability. As it happens, we
actually ended up bending that plate further away, to increase the
gap and loading and then brought it back in to the best position with
the threaded rod. So, we didn't actually change anything. The loading
plate is now probably effectively 1/8" further away.
I am hoping to get to John's place today or tomorrow, so I'll bring
my laser dot temperature thingy and we'll see what gives.
He did remove the plate bocking cap yesterday and showed me the
picture. It has sustained damage and I suspect it is toast. It could
be our problem. It was 2200pF at 10kV. RF rating unknown but
obviously not good.
I'm bringing 2 x 68pF 10kV 7kVAr doorknob caps and we'll put them in
and hope for the best.
I'm sorry but I'm not familiar with the meanings or C1, Cc etc but I
guess they may be Tune and Loading in my language. John is an
engineer. He'll probably slap me for not knowing. Hi hi.
Anyway, we have some more investigating and experimentation to do.
I'll report back with any successes oor continuing failures.
73, Alek, VK6APK
On 2/10/2022 9:12 pm, Steve wrote:
You mention adding 'the ability to adjust output loading'. What
was added and where? The original circuit does allow the loading to
be altered, just not in the way that you might be used to with 'tune'
and 'load' controls on a pi network.
Different combinations of Cc and C1 allow the load presented to
the GS35 to be made resistive and varied over a range but the tuning
is very sensitive to either of them being varied. I looked at it on a
Smith chart to get a feel of how it operates; I'm sorry but I lack
the skill to explain it simply in any other terms. Perhaps others can
help.
Getting the wanted load at the anode is also quite sensitive to
the inductor and total capacitance to ground at the GS35 (which can
be very much more than the Cag figure in the data). Copying the
circuit without knowing the original layout in detail could easily
throw the end result out.
From what you describe, it seems like a thermal thing; maybe
valve capacitance, maybe the inductor or capacitor(s) are heating up.
All of them could result in what you describe. Can you get an
infrared camera to peer in there?
Both anode and grid current high suggests to me that the load
impedance is too low. Less Cc and more C1 is the combination to push
it higher.
I took a guess at component values to get a realistic end result
and the loaded Q at the L/C1/Cc junction can head up into the 30+
region. Increasing Cc and C1, decreasing L results in lower loaded Q
for the same load resistance at the GS35, which might help, but
you'll have to cut and try until it works.
73, Steve G8GSQ
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|