More on the PI network

[W4AMV]

Prior posts on Pi network prompted me to look at the peak currents and
voltages developed in an amplifier I built earlier. I put together a Pi net
similar to what is in use for a 4 parallel 807's. The simulation assigned a
peak RF voltage on a DC offset of 1000 V to mimic a 200 watt carrier. This
source is a rough approximation to the tubes in that its a sine wave and not a
pulse of current filtered by a tank. So these numbers may be a ball park estimate.

I might be better off using a current generator in place of V1 to simulate a peak AC fundamental current that is in line with class C operation for the 807's.

Anyway, I see RMS voltages and currents in the Pi network that seem reasonable. The one interesting surprise is the currents in the load and tune caps. They are similar. Of course the voltage across the tune cap is significantly greater than the load cap. So if it is selected to handle the required break down V, it probably will handle the required current.

Do these values seem reasonable at 200 W carrier and 1000 V DC?

73' Alan

[K9MB]

There are formulas in old Radio handbooks, but this online calculator allows instant output for different source and load impedances and peak voltages and currents are there to help in choosing components.

https://coil32.net/online-calculators/pi-matching-network-calculator.html

If you want to roll your own, Orr put variations of the article in this 1967 Radio Handbook in a lot of these Handbooks.
See page 258

https://worldradiohistory.com/BOOKSHELF-ARH/Handbooks/Radio-Handbook-17-1967.pdf

73, Mike K9MB

[W4AMV]

Excellent Mike. Good resource and thanks for the link.
I compared results and they are very close to simulation.

Next, I am using old school DUBILIER caps as padders in the
load cap slot. They are rated at 2.5 kV and 5 kV test. The ESR
at low frequencies is quite low, need to see what I have at say
40 meters and if the RMS current x ESR or cap dissipation is excessive.

I am trying to chase down an efficiency issue.

73' Alan

[K9MB]

Excellent Mike. Good resource and thanks for the link.
I compared results and they are very close to simulation.

Next, I am using old school DUBILIER caps as padders in the
load cap slot. They are rated at 2.5 kV and 5 kV test. The ESR
at low frequencies is quite low, need to see what I have at say
40 meters and if the RMS current x ESR or cap dissipation is excessive.

I am trying to chase down an efficiency issue.


73' Alan

Are you talking about the 8XX series doorknob caps?
The lower value ones are NP0, but most of them are N750 or even more, so if they heat up, your loading will drift.

Some guys put a second air dielectric variable in parallel with their main loading cap and switch it in. There are some nice 1000 volt rated variables on Ebay now from old HP generators.
A more modern, compact and elegant approach is to use some of the ATC100E porcelain MLCC caps available on EBAY now. They come in 2500volt ratings and some have silver plated tabs. They can handle 15-18 RF amps each at a wide frequency range. They are widely used in Class E transmitters.
Here is a shot from the data sheet:

https://datasheets.avx.com/100E.pdf

[W4AMV]

Hi Mike,

Thanks again,

These are so called bathtub style. Molded style, some with screw down and some with ribbon leads.
Attached a picture from the 50's handbook.

Various manufacturers, Sprague etc... MicaMold

73' Alan

[K9MB]

I have used those in the past. I built a 4x 811A 30L1 clone linear amp in the early 80s running 1500 volts at 700mA peak that put out 700 watts of carrier on tuneup and put a couple of these in parallel in the loading circuit and they ran cool.

They should handle several RF amps for short periods and they are temperature stable. I have a tub full of them and use them for bypassing  in RF circuits. They should work well for your purpose. If they stay cool under full loading power you will know they are fine.
73, MB