Developing a low frequency High Voltage

[Marco]

Yah, I mentioned the Trek amplifiers a little further up. If you're really determined, you could probably build a push pull amplifier with a cascode, especially since only 50 Hz is required.

Maybe something like this with optocouplers so you don't need PMOS and keep the number of MOSFETs somewhat reasonable (open loop here, but trivial to close). It's not going to have much bandwidth because the resistor divider and the gate capacitances, but you can try to decouple those if more speed is required. For optocouplers you'd need something like OPI1270 to withstand the voltage and for MOSFETs something like UnitedSIC so they can handle the linear mode. Even with 1700V MOSFETs, this requires a stack of 22 for 18 kV peak to peak.

The optocoupler approach would work with tubes too to make a push pull amplifier, would need isolated power supplies for the heaters though.

PS. the patent I linked is actually from the founder of Trek even if not assigned to the company, wonder if they use it. I had a similar idea with a NPN/PNP buffer, but was looking for a clearly explained example.

[Qmavam]

I'd like to have a low frequency, adjustable, say 3hz to 50Hz High Voltage 9kV sinusoidal AC signal.

Is it possible to de-rate a microwave transformer to work at such low frequencies?

If so, then I could series several to get my voltage.

Google finds this
https://electronics.stackexchange.com/questions/237714/low-frequency-1-10hz-transformer

it seems you just proportionally de-rate  (ωL is constant)  so a 50Hz mains transformer accepting 240V, will accept 14.4V at 3Hz
'several' may be a higher number than you imagined ? 

 This is a little bit tempting, I have a transformer out of a welder, the core is 10" wide x 6" tall and 5.5" deep, it's 60 cycle. The next thing is to find an amp that would drive a couple dozen watts at that low frequency.
  I suspect this more of a project then I want to take on. It's something my son is doing at work, if they want it bad enough they'll need to shell out the money for a commercial unit.                           
                    Thanks, for the input, Qmavam
     

[Qmavam]

Have you considered building an amplifier around the 6BK4C  https://frank.pocnet.net/sheets/107/6/6BK4C.pdf  or similar high-voltage vacuum tube?

 Funny you mention that, I'm actually awaiting parts to build a HV pulser using a 6BK4C.
 Do you have any info about radiation output when I'm running 8kV?
 I bout the C, because it is supposed to have leaded glass to help, but I wonder if I need
a tube cage to be safe?
                                        Mikek

[TimFox]

When operated at 25 kV, 1 mA the Sylvania datasheet estimates 1.5 mrad/hr (but forgets to mention at what distance from the tube, perhaps 1 meter?).
The RCA datasheet for the 6BK4C estimates 0.5 mrad/hr, and refers back to some JEDEC specifications for the x-ray problem.
At 8 kV, an x-ray tube rule of thumb (radiation proportional to V²) estimates 10x less radiation output, and of softer x rays (implying somewhere around 0.1 mrad/hr at that reference distance).
You should have a metal shield anyway for HV safety.
X-ray absorption for Bremsstrahlung sources is usually specified by "HVL" (half-value layer), but there are few measurements available below 10 kVp.
Being real conservative, a published number for 31 kVp is about 0.5 mm of aluminum:  3 mm of Al would give you a factor of 1/64 at that energy.
Safety specs keep getting lower (and harder to measure), but the natural background radiation on Earth is roughly 250 mrad per year, or 0.03 mrad/hr, depending on location (with additional man-made dose from medical, etc.).
Staying well below that level is a conservative safety measure.

[coppercone2]

steel is alot easier to solder then aluminum, if you even want do that. I would use spot welded steel boxes for this with edge overlap and a few spot welds to hold it together. Easy construction like a car

it does not need to be made like a EMI shield. No reason not to use JB weld to glue it together either... or just glue lead. It's only had when you want to combine radiation and EMI shield, and you can totally coat it in a thick layer of epoxy to be safe with lead.

[TimFox]

Yes, for metal shielding of an x-ray source, you just need to avoid gaps that line up with the direction from the source (which here is the internal anode of the tube).
No need for lead, steel or even aluminum at this energy should suffice.

[coppercone2]

well, to combine the HV shield and weak rad shield, I recommend nickel silver as a material that is very easy to solder, and quite easy to bend into a shape.

It comes together as easily as copper but it should give more rad shielding then copper. And its robust, which is nice for a safety shield, you don't want it crinkled or ripped

I think its better mechanically then copper anyway. but solders easier then steel, and it will be a better low frequency EMI shield too (better near DC)

I would recommend if you can do it for low price, have it background if your fingers are on the chassis and low when its open for repair. That IMO is solid construction that you can be proud of