Electronics > Projects, Designs, and Technical Stuff
how would a very high voltage fast AC switch look like?
(1/1)
coppercone2:
So while investigating my welder thread, I discovered that making a fast high voltage switch that can block RF and HVDC @ +20KV but then be turned on to pass 4A of LF (500Hz max) AC, in a square wave, with a band width of 5KHz is pretty difficult.

While I don't think its necessarily the best choice for a welder, since its supposed to be a practical instrument, at least in this point in time with existing technology, I became curious to discover how such a switch would work.

How would the circuit topology look like for a fast switch capable of transmitting AC once the HV died down?

The first thing that came to mind was to try to make some kind of series transmission gate. That would require high voltage NMOS and PMOS and a HV high value resistor to basically shunt the leakage current across them.

Also, when you really have alot of series mosfets, can you even rely on a passive circuit to match their speed? or would you need to measure voltage across them constantly to try to keep the gate current matched?

Has anyone seen anything like this, maybe for a physics supply or laser pulse circuit?

I know this is probably the hardest way to tackle the problem, and I am looking into tesla coil design, but it really got my curiosity.

The best NMOS on digikey is something like 4.5kV rated @ 2Amps, so you would need like 6 of em in series to do this job, for a DC signal.. not sure how to pass a bipolar a AC signal from a current source.

The current source would be limited to 100V peak.

Could you make a fet that has a very high gate voltage or something?

I know there wont be existing solutions this complicated but even if something approaches it I would be interested to know.

jbb:
20kV isn’t super high. Silicon Carbide (SiC) devices with 10 mag ratings are under development.

Cascode power JFETs, MOSFETs or BJTs might do the trick.

For a practical design, maybe use a saturable reactor (or transformer core) in the HF injection path? Or use the HF injection transformer as a stabilising output inductor?
coppercone2:
Yea thats what I will go for but I wanna know about odd switch topologies with alot of series swtiches.

If 20kV is not high enough how about 40kV? Do you have any schematics for a switch circuit that would be considered unique/beastly? I wanna get an idea of what it would look like when its actually engineered by someone that has a good idea it will work.

Can you bypass with varistors if you are using it as a pulse control circuit for HVDC with a indefinate off time? Or would you bypass with both varistors and resistors?
RJSV:
OK This is probably above your performance needs...  Maybe helpful.
   I worked on a 4 Meg Volt 'supply'.   The common switch was about the size of an oil lamp;  a sort round bottle, where the electrodes were brass rings and brass spheres.  I believe the input was 5,000 volts,  and the output (pulse) was very violent,  actually putting pitts on the brass metal, and emitting UV light!   You could try Physics International,  year 1995.
   There was no filament, and I believe not a vacume but there was a pressurized air flow, to clear the electrode(s), of any hot metal drops.
And so the device was run interrmittantly,  at 10 hz.
   The whole system was the size and looks of a large metal 'dumpster'.   ENJOY...
T3sl4co1l:
No idea what you're after; vacuum disconnects seem to do well enough for such loads in power distribution.  A standard part, but for welding as such, a good old fashioned mechanical switch, or plugging the tapered connectors directly, is more than adequate.

Tim
Navigation
Message Index
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod