Author Topic: a solid-state kilovolt (de)multiplexer?  (Read 1278 times)

0 Members and 1 Guest are viewing this topic.

Online ledtesterTopic starter

  • Super Contributor
  • ***
  • Posts: 3286
  • Country: us
a solid-state kilovolt (de)multiplexer?
« on: August 21, 2020, 11:23:36 pm »
I was watching this Electroboom video:

https://youtu.be/gqng2MPClO8?t=8m25s

and I was wondering how to do implement it without the motor (skip to 8:25 to jump right to the setup.)

Someone in the comments suggested that a car distributor might work. Is there a solid-state solution for this kind of high-voltage demultiplexing?

 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: a solid-state kilovolt (de)multiplexer?
« Reply #1 on: August 22, 2020, 06:09:47 am »
That's probably at the point you'd want to use x-ray or beam switched semiconductors.  Take a lump of intrinsic semi, blast it with energy, free up charge carriers, profit.  Just a photocell, but using the bulk.  Optical light doesn't have enough penetration depth, it's more of a surface thing.  Penetrating radiation is needed, such as x-rays or beta rays.  Which are obviously not the cheapest thing to come by, as you need a vacuum tube and high voltage.

A reminder what you're asking for.  When you say "this kind of" thing, you're talking about quite high voltages (10s kV), with peak currents of hundreds or even thousands of amperes -- capacitive discharge, fast rise times (air ionization occurs in a few nanoseconds).  To be able to mux that, you need some pretty nice switches.

If you were just distributing a little current here and there, you'd still have a problem switching it; transistors are driven with a few volts, which has to be isolated to fly on top of thousands.  You could do the same thing over again, just using optical light -- photocells or photodiodes, activating a surface layer so the conductivity kinda sucks but that doesn't matter for ~uA.  Photodiodes of course have to be used back-to-back; and their gain sucks but again, for ~uA that's fine.  (Photodiodes also having the downside of generating photovoltage, which should cancel out in a back-to-back arrangement, which can be trimmed by adjusting individual illumination.)

The way power companies do it (switching many megawatts) is with big stacks of transistors (usually thyristors, since the gain is much higher, the frequency is low, and the load is AC), and isolated drivers, which are either self (parasite) powered, solar powered, or transformer coupled.  Transformers that can isolate hundreds of kV don't perform well, so the use of solar cells really is pretty attractive, despite the poor efficiency!  (There are also direct optical triggered SCRs, which I'm not sure how common they are, but it's a thing.)

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: ledtester

Online KT88

  • Frequent Contributor
  • **
  • Posts: 334
  • Country: de
Re: a solid-state kilovolt (de)multiplexer?
« Reply #2 on: August 22, 2020, 08:18:36 pm »
If you have deeeeep pockets: http://www.behlke.com/
 
The following users thanked this post: ledtester

Offline TheMG

  • Frequent Contributor
  • **
  • Posts: 867
  • Country: ca
Re: a solid-state kilovolt (de)multiplexer?
« Reply #3 on: August 22, 2020, 10:53:20 pm »
If you want solid state the easiest and cheapest way to do this is going to be individually controlled high voltage sources (one high voltage transformer dedicated to each output). Switching such high voltages (and peak currents) by solid state components is no trivial (or cheap) task.
 
The following users thanked this post: ledtester

Offline jbb

  • Super Contributor
  • ***
  • Posts: 1265
  • Country: nz
Re: a solid-state kilovolt (de)multiplexer?
« Reply #4 on: August 23, 2020, 01:22:16 am »
Direct optically triggered SCRs are cool. Sadly, I heard from the semiconductor team at my last job that you had to make so many silicon compromises so the optical trigger works that the electrical performance isn’t very good.
 

Offline Pawelr98

  • Regular Contributor
  • *
  • Posts: 105
  • Country: pl
Re: a solid-state kilovolt (de)multiplexer?
« Reply #5 on: August 24, 2020, 05:25:33 pm »
Individual high voltage sources is the easiest way to do it.
Those chinese modules are fairly cheap to buy.

Another alternative to solid state is vacuum tubes.
Those are cheap and can take multiple kV.
Look up the soviet 6E15P. Nominal operating voltage 5kV and it should be able to take higher voltages.
Around 1USD each.

If AC conduction is desired then form a bridge rectifier out of HV diodes and connect the switching element between the "positive" and "negative" of the bridge. The same method could be used with a high voltage transistor.

If only slow switching is needed then DY87 or any other high voltage diode may be used with the heater as ON/OFF switch.
« Last Edit: August 24, 2020, 05:28:55 pm by Pawelr98 »
 

Online KT88

  • Frequent Contributor
  • **
  • Posts: 334
  • Country: de
Re: a solid-state kilovolt (de)multiplexer?
« Reply #6 on: August 24, 2020, 08:43:35 pm »
I worked on a lightning protection cicuit a while ago. For testing  we used triggered spark gaps. These consisted of two speres with one having a small hole in the center having an isolated trigger electrode through it. A weak trigger spark ionized the air between the spheres which then ignited the main discharge  - 50kV into 2Ohms :box:
Instead of the spheres a flash tube might also work... A simple flash trigger circuit could generate the trigger signal.
You can find an image of these bad boys here: https://www.md-automation.de/news/dehn-ist-akkreditiertes-labor-fuer-blitz-und-ueberspannungsschutz
...not solid state though...
« Last Edit: August 24, 2020, 09:06:30 pm by KT88 »
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: a solid-state kilovolt (de)multiplexer?
« Reply #7 on: August 24, 2020, 09:07:03 pm »
Flash tubes are actually rather slow, with a risetime of 5-10us (at nominal supply voltage).  Maybe doable for surge testing (and maybe they go faster at higher voltages -- but mind the peak currents, too!), but quite a bit slower than air (10s ns?).  The voltage drop is also fairly high (~100V), making them not great as switches.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Online KT88

  • Frequent Contributor
  • **
  • Posts: 334
  • Country: de
Re: a solid-state kilovolt (de)multiplexer?
« Reply #8 on: August 24, 2020, 09:17:16 pm »
Quote
The voltage drop is also fairly high (~100V), making them not great as switches.
...which isn't all that bad at 50kV and 25kA...
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: a solid-state kilovolt (de)multiplexer?
« Reply #9 on: August 25, 2020, 01:53:48 am »
Er, flash tubes wot you see in cameras are only 400V and probably pop at a few kA (hm, I haven't tested that, come to think of it).

Even the NIF's flashlamps are... don't see exact figures at a glance; one paper says 800uF at 10kV, with 17uH in series, which should give less than 69kA peak. Nice.

Oh here we go, a test setup: https://ieeexplore.ieee.org/document/1001850 wonder if they upgraded the lamps or ratings, that's more than double the voltage.  But less than half the current, and still a bit more energy than the other figure I saw.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Online KT88

  • Frequent Contributor
  • **
  • Posts: 334
  • Country: de
Re: a solid-state kilovolt (de)multiplexer?
« Reply #10 on: August 25, 2020, 05:52:05 pm »
I guess these flash lamps aren't quite small...
Ignitrons are another class of impressive devices given their size to current ration - an Ignitron the size of a soda can can handle 20kV and 100kA.
For the specific application Electroboom demonstrated,  the operating voltage of the system could be higher than the max. voltage rating of the flash tube because the spark gap and the flash tube would be in series.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf