[Amps] Phase and amplitude of 2 tube amplifiers

[John Lyles]

Over the past decade I designed a system with three high power 
transmitters that each use a pair of tetrodes in combined operation. 
These are TH628L, which run at 23 kV DC each, and produce about 1.5MW 
pulsed per tube, average power is approx 180 kW per tube, at ~200 MHz. 
Combiners and Splitters had limited possibilities and this is a fixed 
frequency operation for driving high Q cavities for a particle 
accelerator. As such, reflected power can also be horrendous at times. I 
chose quadrature combiner, with a 4 port hybrid custom made by a company 
in 12 inch diameter hard coax. The combiner is called a branchline 
coupler. The splitter is also a hybrid, fed with 3 1/8 inch hard coax. 
It uses a pair of coupled lines 90 deg long inside a box. Both of these 
are 90 deg devices. If the amplifiers are running at about the same 
power, with the same voltages, the phase shift through them is similar 
enough. We built a custom vector phase meter that measures the real time 
phase difference between the PAs. When I ramp the power up from zero, 
there is approx a 10 deg phase variation in each amplifier but its 
repeatable and the same in each one. Once the combined system is near 
the optimal 2.1-2.8 MW total power, the phase variation vs power level 
is a few degrees at any time.

Since we have VSWR all times, the length from the load (source of the 
reflected power) to each tube is off by 90 deg if the lines were all cut 
the same length. This would create a messy operation, as one tube would 
respond differently than the other due to the phase of the reflected 
power being different. To fix this, there is extra piece of line on the 
output of one amplifier, so that they are both in phase with respect to 
the load, but there is a similar 90 deg line on the opposite input 
splitter port. This maintains the 90 deg relationship of two feeds to 
the combiner, while staggering the amplifiers apart by 90 deg so that 
they both respond to VSWR the same. It works well like this.

We use the screen voltage (from separate power supplies) as a knob to 
adjust the gain of each PA to be the same. Could also use the ouput 
loading control, but that introduces a phase change. There is a 3-1/8 
inch coaxial phase trimmer, a line stretcher, after the first splitter 
that follows the driver stage, in the feed to one PA input side. On the 
other side there is a nearly identical line without adjustable phase. We 
have run these transmitters for 6-8 years and not seen problems due to 
phase variation. In fact, the curves of phase vs power falloff, of a 90 
deg combiner is quite tolerant. I monitor the 4th port of the hybrid 
combiner, which feeds a water cooled resistor. The power on this port 
occasionally gets to 5 kW peak which is nothing when the total power is 
2.5 MW.

It is important to note that our system is a fixed freq setup, where you 
will likely be tuning around. I can imagine that things get more complex 
when varying frequency and also maintaining phase. Using in phase 
combining and Wilkinson devices will help there. If you are driving two 
antennas for circular polarization, small phase difference should just 
alter it to be more elliptical polarization.

You can find some images from a 2017 paper on this system here. Let me 
know if you want any further info on this.

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwjE6uvZ8d76AhU9jokEHYEuCLAQFnoECBQQAQ&url=https%3A%2F%2Fwww.semanticscholar.org%2Fpaper%2FDesign%252C-Fabrication%252C-Installation-and-Operation-of-Lyles-Barkley%2Faa25180c0cc7b15ba88315673f220e915fe6cde7&usg=AOvVaw083Mcx-V0X_YruR9yfAYb2

73

John

K5PRO

> Date: Thu, 13 Oct 2022 12:37:10 +0000
> From: Conrad PA5Y <g0ruz at g0ruz.com>
> To: "amps at contesting.com" <amps at contesting.com>
> Subject: [Amps] Phase and amplitude of 2 tube amplifiers
>
> I am going to attempt to use 2 identical monoband tube amplifiers to generate CP but before I start I have a question. The 90 deg delay for CP comes from the array orthogonal physical boom separation. So the amplifiers must be in phase at the outputs.
>
> I presume that this has been done in broadcast applications and I would like to know, assuming everything is tuned correctly how close the phase likely to be?
>
> I have line stretchers for the inputs, I have extremely good phase matching on the Wilkinson splitter and 2 identical directional couplers that I have tuned for good phase and amplitude matching. I can also measure the relative amplitude and phase of the 2 amplifiers accurately. Small phase and amplitude adjustments are ok but I have no idea what to expect with tube amplifiers.
>
> This was not really planned but a 2nd amplifier became available, so I thought why not? The tubes are becoming scarce so this way I can be a lot more gentle on them.
>
> Regards, Conrad PA5Y