Hello, I am going to be building a circuit that will be operating at high voltages
( about 100kv or so). The problem is I need to be able to switch it off quickly (about 1 nanosecond and under). Since the switching times are faster then any relay I know could handle, I figured I would end up using transistors controlled by a pulse generator. The problem is I can't have the transistors directly controlling the voltage as they would just short circuit. So What I thought about doing was dividing the voltage using a voltage divider down to a reasonable range of about 100 volts, then having the transistor control the low voltage but I just don't know how to step the voltage back up to 100 kv afterwards

Does anyone know how to, or maybe a better way of doing it without the voltage divider. I just need to be able to have a switching time of 1 nanosecond and under, So from off to full on in about 1 nanosecond.

 

You may have a bit of trouble getting there, 1 nanosecond is awful fast. Add the voltage to it and it becomes a real challenge. I'll be following this thread with interest.

 

Let's see, switching 100kV @ 1GHz.

Mind telling us why you think you need that high of a voltage, switched that quickly?

In the meantime, an LT1016 ultra fast comparator has a 10nS propagation delay time; and that's switching signals at logic levels.

Meanwhile, 100kV can ionize the air and jump a gap of ~1.31 inches at sea level. You're probably not going to break that arc without something like a knife switch from an old Boris Karloff movie, or contacts in oil - and your switching times would be very slow; hundreds of ms at best.

Also, 100kv can be extremely hazardous. We usually don't discuss such high voltage items unless it's purely theoretical.

 

motor driven mechanical commutators just sprang to mind.

Tesla used such things and he achieved very high freq AC transformer linkages.

I don't think any of us are in his 'league' when it comes to the mad scientists stuff though. Better off choosing something slightly less dangerous and cheaper to build.

 

I just had a quick idea, Would it be possible to set it up using a a bunch of transistors connected in series with some resistors. So that way I can control a high voltage of maybe 1000V with say 10 100v Transistors. Which would be operated by a timing circuit.

 

Controlling multiple switch devices in series is possible in principle, but difficult in practice. That sort of thing was done to a limited extent - perhaps two devices in series - in early transistorised TV sets, before satisfactory power transistors were available for use in line scan circuits.

It is difficult to ensure that the switching of each device is synchronised well enough to prevent too much voltage developing at any point. The requirement to switch these very high voltages so quickly would make synchronisation far harder.