RF power question

Yes, 12pF is about 27 ohms at 500MHz. That's 19KW. I see a large klystron in your future, or a UHF Television PA? A number of full-sized racks and the water cooling equipment and power supplies.

Tell him to want something else.

By "interesting", you mean a bigger PA. No passive network is going to make 20kW from 1kW.

Clifford Heath

He's trying to Change The World. I'd like to help.

The load is capacitive so, in theory, needs zero power. In real life, it's water cooled. Still, some resonant network should make the true power requirement more reasonable.

But it's NOT 20 kW he wants, it's 20 kVA; a network with Q of 30 is not impossible. With superconductors, you can get higher.

If the capacitor is part of a resonant LC circuit, you could get quite high resonance voltages with a high Q.

Unfortunately, the 12 pF is quite large for 500 MHz, so only 6 nH is required to resonate it. A wire about 6 mm long has also an inductance about 6 nH, so not very helpful. Connecting a large number of usable inductances in parallel can be used to create the 6 nH total inductance.

Putting the parallel connected inductors around the capacitor is how the cavity resonator principle is sometimes depicted. Then there are issues, how power is connected into the resonator.

The load might be capacitative, but if it's water-cooled it's going to be physically large enough that there's going to more than 6nH of inductance in any simple pair of connecting leads.

This is one of those cases where one has to understand exactly what the load looks like - how big it is, where the current flows - and why it looks like that.

If you sort out what's actually going on, there may be a different - and less demanding - way of getting the desired result.

Upside...downunder mentioned a cavity resonator.

A 500MHz cavity magnetron

might be a bit big.

puts the bottom end of the frequency range at 600MHz, so 500MHz might be attainable.

The actual power level at 500MHz will be high, and the application will be deemed unworkable. Skin-depth and proximity-effect issues alone, will doom this project.

UHF TV transmitters have some sorts of matching networks, at megawatt levels. There must be a way to do this.

The guy is at MIT. Should I introduce you?

What is the water cooling for ? From where is that power coming that needs to be dissipated and removed by water cooling ? If it is from the RF source, so some real power is required, no impedance tricks will not help much, the loaded Q will be low.

The analog UHF TV transmitter power is given as EiRP (Effective radiated power) towards the horizon. Since the antenna system flattens the radiation pattern, the antenna gain is at least 10 dBi. Thus the actual transmitters are actually about 100 kW.

On a transmitter system, one tries to keep the impedance levels reasonable, say around 50 ohms, so not too exotic impedance matching is required.

On a sunny day (Fri, 22 Nov 2019 12:43:34 -0800) it happened John Larkin wrote in :

What else does the load look like apart from the 12 pF? If it is capacitive only then you could drive it at low power in resonance,

Not so fast, you're talking about losses in an 8.5nH inductor, running at 500V and 18 amps, at 500MHz.

Haha, no thanks! I worked very hard and made a resonant inductor/transformer for 10kV at 1MHz, and could likely do 1kV at 10MHz, but not 500MHz. A cavity or magnetron with some wisardry, maybe.

What about this? :

Version 4 SHEET 1 880 680 WIRE -64 64 -224 64 WIRE 96 64 16 64 WIRE 304 64 176 64 WIRE -224 112 -224 64 WIRE 304 112 304 64 WIRE -224 208 -224 192 WIRE 304 208 304 176 FLAG -224 208 0 FLAG 304 208 0 SYMBOL cap 288 112 R0 SYMATTR InstName C1 SYMATTR Value 12p SYMBOL ind -80 80 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName L1 SYMATTR Value 8.44e-9 SYMBOL res 80 80 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 0 56 VBottom 2 SYMATTR InstName R1 SYMATTR Value .1 SYMBOL voltage -224 96 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V1 SYMATTR Value SINE(0 16 500e6) TEXT -264 -64 Left 2 !;ac oct 1000 400e6 600e6 TEXT -264 -120 Left 2 !.tran 5u

The capacitive load device has some losses, so it is cooled. Presumably a water systen is available to cool other things, like the power amp and the matching network. A little water goes a long way.

From where is that power coming that

That's still about 50:1 over what we'll need.

I'm not looking for an argument, I'm looking for some helpful suggestions for a researcher.

I suspected that you might not want to volunteer to help yet another bewildered physicist. The limit of my volunteering is to ask here.

I had plenty of trouble making an inductor for my Pockels Cell driver, at some tiny fraction of the power*frequency that this guy needs.

The load is water cooled, so it does have some losses. Anything but a good vacuum will have losses at 500 MHz 2KV p-p.

The matching network would probably need to be water cooled too. Maybe a coaxial rod-in-cylinder thing, like you see in the old Radio Amateur Handbook.

I think I metioned that I'm not an RF guy.

When I was at the Vanderbilt FEL (electron accelerator) Pulsed power was from a 'big' klystron. Wave guide impedance matching was done by 'artful' placement barriers in the wave guide (I took a course, (but never designed anything) the artful part was a first good guess, then EM modeling software, then...)

maybe something with stripline? The 'right' length of transmission line and a stub tuner?

(Is there something in the literature RSI?)

George H.

Please comment on the LTSpice which I posted at 9:59. I assumed a Q of

On a sunny day (Sat, 23 Nov 2019 09:17:14 -0800) it happened snipped-for-privacy@highlandsniptechnology.com wrote in :

Yes cavities, if the 'load' has the right size put it in a cavity (just joking). hard to tell without more data.

A microwave oven is sort of a cavity, maybe a bigger housing for the 5 x lower frequency and a turntable? Are there 1/2 GHz magnetrons?

I seriously doubt that Q is possible, but you also assume a signal source impedance of 0 ohms at 500MHz. It's easy to make SPICE models do almost anything.