Finding/Implementing a digitally-controlled high-voltage switch

[MarcoPoloExplorer]

I am looking for a switch or relay to switch high voltages on and off based on a digital input signal. Below are the characteristics of the device.

Switching frequency: 100 Hz to 100 KHz
Input (control) signal: square waves, 3 to 12 volts amplitude, 100 Hz to 100 KHz range
Input voltage to be switched: 0 to 5 kilovolts DC, 1 milliamp (max)
Output signal: square waves, 0 to 5 kilovolts amplitude, same frequency as control signal
Load: capacitive load (several picofarads)
Cost: not more than $100 per switch

Thanks for any info you can contribute to this!

 

[Skogsgurra]

The standard device is the Ross relay. You will need to add a circuit that condistions and amplifies your quite unusual input signal. Ross relays usually operate on 50/60 Hz AC. Ross relays are sturdy things that easily switch many amperes. I do not think you will be able to buy one at the price indicated by you.

Other, less spacious and less expensive relays generally have a problem with voltages in your range.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[VEBill]

Is the OP expecting to switch the high voltage on and off at 100Hz to 100kHz (!!) ??

 

[Skogsgurra]

I can't imagine that.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[VEBill]

Output signal: square waves, 0 to 5 kilovolts amplitude, same frequency as control signal

A non-trivial set of requirements.

 

[Skogsgurra]

Hmm.. Your watch

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[VEBill]

Sounds like a job for Nikola Tesla. He was pretty good at this sort of thing.

 

[MarcoPoloExplorer]

Thanks all for the response. Yes, the intent is to be able to set the frequency of the control signal to any value between 100 Hz and 100 KHz. I did find a MOSFET transistor that switched up to 1.5 KV at more than 100 KHz (STP4N150), but I've yet to find one that goes up to 5 KV and 100 KHz. I'm driving a fairly small capacitive load (10-50 pF).

If my calculations are correct, it should be possible to charge and discharge 5 KV at faster than 100 KHz from a voltage source, given a transistor with an on-resistance of a few ohms (<7 ohms in the case of STP4N150). The output of the voltage source has a stabilizing capacitor (2.2 nF) that is about 50 to 200 times the capacity of the load capacitor. This stabilizing capacitor recharges when the load capacitor is discharging.

I would like to find transistors or other devices that don't require bulk purchase (hundreds of dollars) as I would like to test them before I commit to larger quantities.

 

[MarcoPoloExplorer]

Actually, the load capacitance is closer to just 1 pF.

 

[Skogsgurra]

Are you aware that 1 pF is less than one inch of wire? A standard terminal pin or screw has several pF.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[MacGyverS2000]

I sooooo need to know what the application is now... those are some huge numbers when coming from both sides of the equation (speed and voltage).

Dan - Owner

http://www.Hi-TecDesigns.com

 

[LionelHutz]

My guess would be that you will have to impliment something along the lines of 5 x 1kV isolated power supplies and use each one as the power source for a 1/2 bridge. You connect the output of each 1/2 bridge to the common or negative of the next one to create a series string. If you switch the upper devices on then you get 5kV and if you switch the lower devices on then you get 0V.

Of course, this approach adds it's own problems in that these 1/2 bridges and the isolated power supplies will have to all have very low capacitance to earth ground to ensure you can switch the voltage that quickly. Best would be to reference the middle 1/2 bridge to earth and allow the 2 on each side to switch between earth and 2.5kV of earth. Still, I'm not convinced you'd be able to get the whole thing to work at speeds anywhere close to 100kHz.

 

[Skogsgurra]

Aha! This is what MPE is after: thread248-301947


Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Skogsgurra]

"Input voltage to be switched: 0 to 5 kilovolts DC, 1 milliamp (max)"

I do not think he wants a relay at all. Sounds more like some unrealistic overunity project.

Can you verify that MPE? So we don't waste our time.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[MarcoPoloExplorer]

Yes, I was considering using a piezoelectric transformer, which essentially acts like a relay. An input of low voltage is transformed to as much as several KV at thousands of KHz (ex. see this piezoelectric transformer: SMSTF68P10S9). However, in examining this approach, there appears to be some issues (ex. having to operate at the resonant frequency). I am still reviewing the details before I make a final decision.

The relays that I've found either don't switch fast enough or the voltage does not go as high as 5 KV.

The maximum current of 1 mA is mostly for safety reasons (threshold of human sensitivity to electricity) and is the output of the voltage source in use.

@Skogsgurra: I find that the capacitance depends strongly on the distance between the two terminals/wires, which are separated by a good distance. But it's possible that additional stray capacitance may come from a second switching device when it's off. One device charges and the second device discharges. I need to look closer at this as it could be an issue.

@MacGyverS2000: It's for a supporting circuit in a master's thesis project. Is Dan from hi-tecdesigns.com someone who would know something about this application?

@LionelHutz: Thanks for the 1/2 bridge idea. I had thought of something like that. I'll revisit that if I don't find something more simple.

All, thanks!

 

[Skogsgurra]

MPE

Yes, I think that we all know how capacitance varies with distance between wires/objects/plates. What I said is that you will never get down to 1 pF in any kind of real device. You will most likely end up with at least 30 - 40 pF. Even if you operate your thing in outer space.

Your thinking that a transformer is the same thing as a relay is completely wrong. Where did you get that from?

Whatever your project is - do not count on success. Your thinking is more wishful than based on reality. Sorry, but that's a fact.



Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[MarcoPoloExplorer]

Hello Skogsgurra, yes it's true that the piezoelectric transformer is not a "relay" to be exact. But the basic idea is that I'm using a small voltage to either control (relay) or generate (transformer) a larger voltage. Either way would work for my application.

It is possible to resonate out a large capacitance with a set of inductors so that it appears small or non-existent. However, this is a more complex approach and would probably work better with sine waves, not square waves.

This is part of a research project. I try to remember that people like Thomas Edison tried thousands of failed experiments before they succeeded (ex. commercially practical light bulb in the case of Edison). My opinion about R&D is that it's good to keep the mind open to possibilities. I think ideas start in a wishful sense, then they are reduced to practice. Albert Einstein is reputed to have said that "imagination is more important than knowledge." Sometimes an approach does not work, but in the process I sometimes make a useful discovery. In any case, I really appreciate all your help. Your post on capacitance was very helpful because it allowed me to bypass some experiments:

I think one solution is to use a larger capacitor at the voltage source and then charge and discharge the load capacitor for only a few cycles. This will allow the source capacitor to recharge before doing another run. To do this, I still need to find a good switching device which can have as much as a 200 pF of output capacitance, for example. I am currently researching this transistor (IXTV03N400S, up to 4KV) and will see if I can get a few test quantities from the manufacturer.

 

[MarcoPoloExplorer]

A few posts above it should say "tens of KHz" not "thousands of KHz."

 

[Skogsgurra]

Your references to Edison and Einstein are soo common in your circles that I was expecting them.

Edison was lucky a few times. But basically, he was a dabbler. I am not impressed by him.

Einstein had a deep knowledge, but was still open for what we today call thinking outside the box. He knew where the box was and he also had made observations that made him draw well-founded conclusions - which he also described accurately in equations - and tested against real world observations. There is no comparison with yours and other overunity fanatics clueless contraptions.

You cannot convince me or any other visitor on EngTips. Sorry. The sooner you realize this, the better for you.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[LionelHutz]

Gunner, I have a feeling your hunch is correct especially since there was no denial...