High Voltage Swithching by Mosfet. / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

High Voltage Swithching by Mosfet.

1 2 3 next

Move Thread

LAN_403

Patrick

Mon Jul 13 2015, 11:47PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Id like you all to use a "sane-o-meter" to check my idea.

looking here :

and

And researching some previous work by others. I need to be able to switch HV from a single location (source) to many different locations (sinks) at differing times. At the 10 to 20 Kv level.

Ive seen some work by power companies using long towers of bricks with toriods for field grading. I may not need all that, but id like 100uS rise and fall times (could be slower).

so is this possible practically or am i crazy, or should i just have an array of transformers driven by a microprocessor ?

Ross Allan

Wed Jul 15 2015, 12:28PM

Registered Member #55411 Joined: Mon Jul 06 2015, 04:57PM
Location: Glasgow, Scotland
Posts: 12

It should be possible to stack them. They like to fail short, so any issue may result in cascading failure of the entire stack. (and possibly high voltage through your delicate control circuitry?)

The following papers describe working circuits:

Series operation of power MOSFETs for high-speed, high-voltageswitching applications

stacking power MOSFETs for use in high speed instrumentation

HIGH-VOLTAGE SERIES MOSFET OUTPUT DRIVER FOR NANOMETER TECHNOLOGIES - presumably doesn't need to be nanometer scale? Same circuit layout should work otherwise. Maybe.

For IGBTs:

Multi-series Connection of High-Voltage IGBTs

You'll probably have quite high resistance through the stack, which could be a deal-breaker.

GrantX

Wed Jul 15 2015, 04:12PM

Registered Member #4074 Joined: Mon Aug 29 2011, 06:58AM
Location: Australia
Posts: 335

With regard to the stack failing and killing the control electronics, it might be possible to wind GDT's with some HV-insulated wire to avoid this.

Electra

Wed Jul 15 2015, 08:10PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

The thing I don't like about the source coupled stacks, is all the fets turn on and off sequentially, the bottom one is the only one driven directly.

Personally I would do it with a large toriod gate transformer with multiple secondary's (1 per fet) wound with the thinnest hv wire that will fit. Usual passive components back to back zenners e.c.t on the gates, and some form of static voltage balancing, resistors across each drain source or a few Tvs diodes in series instead.

Assuming your drive wave shape and frequency can go through a transformer that is, else you are going to need more complex secondary side driver circuitry.

Bored Chemist

Wed Jul 15 2015, 08:32PM

Registered Member #193 Joined: Fri Feb 17 2006, 07:04AM
Location: sheffield
Posts: 1022

I presume that relays are too slow.
This sort of thing.

Patrick

Fri Jul 17 2015, 06:02AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

I was thinking, about a special ferite toroid specifically engineered for high voltage multiple secondaries. i guess a toroid is the best shape.

then, some complicated support components on each gate. TVS's, rise and fall diodes and what not. im deciding on mosfets or igbts at the moment. not sure what the current average or peak is yet, for my time intervals.

Sulaiman

Fri Jul 17 2015, 06:27AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

I have serviced equipment with about 20 mosfets in series,
each mosfet had a ferrite toroid with gate winding (secondaries)
and a single well insulated wire passed through all toroids (primaries)
this gives voltage insulation and each mosfet gets the same voltage vs. time waveform

equivalent to having N single turn primary GDTs in series,
so you could prototype just one stage using a single turn primary
of course high permeability ferrite toroid cores are required as for any GDT
the longer the required ON time the greater the number of secondary turns required
Integral(e.dt) = integral(Vgate.dt/turns) = saturation limit of core. (e = volts/turn)

Wolfram

Fri Jul 17 2015, 06:30AM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

How fast does it need to switch? How long does it need to stay on?

Patrick

Fri Jul 17 2015, 08:53PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

I dont want anything crazy strict. like 1nS on/off times. Or 100 billion volt / second changes.

So, lets start with 20uS rise, 20uS fall and 100 or more duration high uS.

im starting to think 15 kV maybe to lowest i can go. but im looking for the lowest cost mosfets, like in the 2kV range. the 4kv ones really jump in cost. and if i blow a few the cost will be a problem.

so if i down rate the system by 20% that will mean 15kV x 1.2 = 18kV for each leg.

Looking here :

and at the datasheet :

and

edit:

Ash Small

Fri Jul 17 2015, 11:26PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Can you not use a transmitter tube, or something?

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob