High voltage switches.

[haveyouseenhim]

Kind of a off topic question, but I figure I would try anyways.

I have a small capacitor bank I use for various fun stuff, like frying junk components, coil gun, remote firework igniter etc.  I have a little circuit that I charge the bank with and it uses a single AA battery. I put a small switch to cut the battery on and off to charge it and another that is in series with the positive contact of the bank. The one in series with the bank can't handle the sudden load put on it and fries.

My question is where can I find a switch that can handle 350VDC. I have no idea what kind of amperage this thing puts out. The highest rated switch I have is 3A at 230V and it fries after two discharges. 

Should I maybe look into solid state switches or relays?   I'm a bit lost as usual.

[Johan]

Powerswitches in europe are often rated 500volt. ..so i guess what im saying is ...look across the pond...
J

[Seljer]

I don't think the way to go torturing any sort of mechanical switch or relay.

The most robust/cheap electronic solution is an SCR or thyristor. The symbol is a diode with extra a gate terminal poking out. I defaults as non-conducting, but if you apply a pulse of low current into the gate, the device begins conducting (but you then can't turn it off, it only turns itself off after current stops flowing through it).

Or a big power mosfet should do the job as well providing you haven't got a huuuuge capacitor bank (and MOSFETS happily turn off so you can generate short pulses instead of going all out)


I don't understand in your post if you're charging it just to a couple of volts from the batteries or are actually running the capacitors up to 350V. If you're running high voltage use some optoisolators!!!!

[haveyouseenhim]

The most robust/cheap electronic solution is an SCR or thyristor.

Cool. That looks like it would fit the bill.   EDIT    after thinking about it I think the power mosfet would do better. The whole reason for the extra switch on the bank is for a little extra safety.

Or a big power mosfet should do the job as well providing you haven't got a huuuuge capacitor bank (and MOSFETS happily turn off so you can generate short pulses instead of going all out)

It's very small. It is several disposable camera capacitors in parallel equaling about 1200uf at 350VDC.  

I don't understand in your post if you're charging it just to a couple of volts from the batteries or are actually running the capacitors up to 350V. If you're running high voltage use some optoisolators!!!!

I'm charging them to 350 with the charging circuit from a camera.

It's scary to think that this was one of my first electronics projects.   I found it and figured I'd get it working safely.

[teemuk]

1200uf at 350VDC

That's a lot of energy stored.

The highest rated switch I have is 3A at 230V and it fries after two discharges. 

Those ratings are better for switching applications of < 48VDC and <1A current, and in which transients due to switching are not likely to occur. Even a standard "standby" feature of tube amps eats those kinds of switches on regular basis due to high voltage and current transients, arcing and contact bounce. And that's way mcuh less energy involved.

I have no idea what kind of amperage this thing puts out.

That would depend on load. If you discharge it to 1M resistor... not much. If you discharge it to short circuit... impressive amounts.

My question is where can I find a switch that can handle 350VDC.

From a store that sells electronic components? From a store that sell components for electric wiring etc?

[haveyouseenhim]

That's a lot of energy stored.

I mean in the world of capacitor banks that is very small.

From a store that sells electronic components? From a store that sell components for electric wiring etc?

I don't have any stores around here. I used the highest rated one I could find at ratshack, but it's still not strong enough.

[Seljer]

Doing the math: (350V) * 1200 microfarads = 0.42 coulombs of charge
If you discharge that in a 1/100 of a second: 0.42coloumbs / 0.01s = 42A

And probably discharges even faster so ever higher peak current.

When glancing over tables of components you can usually see the rated current, but in the data sheet along with the normal continuous current rating theres also a pulsed current rating.
A 20A mosfet might do 80A pulsed, but as I mentioned, thyristors are much more robust and a 20A thyristor could do 200A (and be cheaper while doing that)

[GibsonGM]

That's a lot of energy stored.

I mean in the world of capacitor banks that is very small.

From a store that sells electronic components? From a store that sell components for electric wiring etc?

I don't have any stores around here. I used the highest rated one I could find at ratshack, but it's still not strong enough.

www.newark.com   or Mouser, or Allied...gonna have to do internet order, man.

[haveyouseenhim]

Doing the math: (350V) * 1200 microfarads = 0.42 coulombs of charge
If you discharge that in a 1/100 of a second: 0.42coloumbs / 0.01s = 42A

And probably discharges even faster so ever higher peak current.

When glancing over tables of components you can usually see the rated current, but in the data sheet along with the normal continuous current rating theres also a pulsed current rating.
A 20A mosfet might do 80A pulsed, but as I mentioned, thyristors are much more robust and a 20A thyristor could do 200A (and be cheaper while doing that)

I have a 600V 10A Triac on hand. Do you think that would do the job? I still have to learn how to apply it, but from what I can tell it just needs 12V on the gate to turn it on.

http://www.weisd.com/store2/NTE5645.pdf

[R.G.]

I dabbled with capacitive discharge circuits once upon a time. The problems mentioned by the replies are accurate. The switch has to handle very -very!- high pulse currents when the caps are discharged, and in many cases has an arc start before the contacts actually close, just as they come near one another. This means that a mechanical switch does self-microwelding and pitting of its contacts. They don't last long this way.

The answer for setups where you want to discharge the cap completely is thyristors in general and SCRs in particular. Triacs are OK-ish, but their specialization for conducting both ways makes them more fragile. The big banks of caps used in rail guns as opposed to coil guns are usually SCR switched. But SCRs can't be turned off in the middle of a pulse. They dump it all unless something else stops the current before the caps are exhausted.

SCRs have very high pulse current ratings. So do MOSFETs, but the MOSFETs are resistive and the channel resistance of the MOSFET tends to limit the peak current through it more than an SCR does. This is probably not an issue with your used, but coil and rail guns, capacitively driven electro-forming of metals, capacitive discharge welders, and so on may not want the limitation. MOSFETs let you turn the pulse OFF in the middle. You may or may not want the peak current limitation or pulse turn off - in fact, from your description, you don't - but you could.

I think you have it backwards. If you want a safety, use a mechanical switch for the *safety* switch and an SCR for firing. Using the mechanical contacts for safety only, never starting or interrupting an arc, will make them last a long time. Using an SCR to fire takes the pulse start off the mechanical switch.

With 1200V on the reserve caps, your mechanical switch needs to be rated for at least 1500 V breakover. Otherwise, it might let an arc reach out and get YOU. Most AC power line switches are rated for this much pulse voltage, so use a switch with a UL, CEL, TUV, etc rating for mains power. At least. You do not want to be the switch contacts that lets the 1.2kV out. Trust me on this point. I speak from experience.

And Seljer is pointed in the right direction on the math, but is way too conservative. Capacitive discharge pulse banks are used to generate pulses of thousands or millions (no, I'm not making that up or exaggerating) of amperes. When you fire a capacitive bank, the current is limited only by the resistance of the loop the current fires in, and the inductance of the path. If the resistance is low, just wires, and the loop area is small, the currents can be large enough to distort metals from the magnetic field-induced eddy currents. See http://webpages.charter.net/tesla/crushed_coin.htm - and note well the warnings. You're an order of magnitude lower energies at the moment, but it's easy to get hurt if you don't think about what could go wrong.

[haveyouseenhim]

I think you have it backwards. If you want a safety, use a mechanical switch for the *safety* switch and an SCR for firing. Using the mechanical contacts for safety only, never starting or interrupting an arc, will make them last a long time. Using an SCR to fire takes the pulse start off the mechanical switch.

With 1200V on the reserve caps, your mechanical switch needs to be rated for at least 1500 V breakover. Otherwise, it might let an arc reach out and get YOU. Most AC power line switches are rated for this much pulse voltage, so use a switch with a UL, CEL, TUV, etc rating for mains power. At least. You do not want to be the switch contacts that lets the 1.2kV out. Trust me on this point. I speak from experience.

Thanks RG   I did have it right on one point though. I'm using a mechanical switch for charging and I was inquiring about how to use a SCR to fire it.

I think you misread the voltage I'm working with. I'm not using 1200V. Just 1200uf at 350V

I'll take your's and Seljer's advice and use an SCR for firing. I don't think I would be cutting it off mid pulse, so I think that would fit the bill better.

[PRR]

Go to Home Depot and get a 10-pack of household wall light switches.

And a plastic switch-box and a STRONG plastic box cover.

Wall-switches will stand-off 350V just fine.

42++ Amps WILL burn the "15A" contacts. That's why you get a 10-pack. Should cost under $0.39 per dozen or more firings. If your fun is not worth 4 cents a hit, pick another hobby.

---OK, they raised the price to $5 for a 10-pack. Your choice of blah colors.
http://www.homedepot.com/p/Leviton-15-Amp-Single-Pole-Toggle-Switch-10-Pack-Light-Almond-M26-01451-2TM/100684041#

[tranceracer]

a little low tech, but what about a mini knife switch or diy contact plate.

[psychedelicfish]

If you still want to use some kind of electronic switching, perhaps you should look in to IGBTs. We Tesla Coilers use these in Dual Resonant Solid State Tesla Coils (DRSSTCs) because we're dealing with massive current pulses normally at around 1kV. IGBTs are perfect for this (way better than MOSFETs, anyway) because they have fast switching times, high voltage and high current ratings, making them great for low pulse width switching. Tayda sells one kind of IGBT, which might do the trick for your capacitor bank. It's rated for 120A 1mS pulses, 60A DC, and can handle 900V from its collector to its emitter.

In terms of switching this thing, an optoisolator might work, but I would prefer to use a transformer to provide a switching voltage. You could rig up an oscillator that gives you a low frequency, low pulse width square wave, and feed it via a switch into your transformer. The transformer could be 15 turns or so of bifilar wire, using one wire as your primary and the other as your secondary. The secondary should be connected from the gate to the emitter of your IGBT, and you should be careful to connect the two windings in phase.

[amptramp]

+1 on IGBT's

After Toyota forced the manufacturers to make better IGBT's for hybrid cars, they got a lot faster and better and you can get voltages into the thousands and currents into the hundreds of amps.  We used them exclusively for power inverters to take a 600 volt solar array and put out anywhere from 30 KW to 500 KW.  We used the same sort of design for 1.2 MW and 2.4 MW fuel cell converters.

[haveyouseenhim]

Should cost under $0.39 per dozen or more firings. If your fun is not worth 4 cents a hit, pick another hobby.

Trust me, I know about price per trigger pull. 

If you still want to use some kind of electronic switching, perhaps you should look in to IGBTs. We Tesla Coilers use these in Dual Resonant Solid State Tesla Coils (DRSSTCs) because we're dealing with massive current pulses normally at around 1kV. IGBTs are perfect for this (way better than MOSFETs, anyway) because they have fast switching times, high voltage and high current ratings, making them great for low pulse width switching. Tayda sells one kind of IGBT, which might do the trick for your capacitor bank. It's rated for 120A 1mS pulses, 60A DC, and can handle 900V from its collector to its emitter.

In terms of switching this thing, an optoisolator might work, but I would prefer to use a transformer to provide a switching voltage. You could rig up an oscillator that gives you a low frequency, low pulse width square wave, and feed it via a switch into your transformer. The transformer could be 15 turns or so of bifilar wire, using one wire as your primary and the other as your secondary. The secondary should be connected from the gate to the emitter of your IGBT, and you should be careful to connect the two windings in phase.

English please    I'll look into IGBTs for switching. I'm not looking to make it super complex. I'm just looking for the simplest most effective way to discharge the capacitor bank.

[psychedelicfish]

I'm just looking for the simplest most effective way to discharge the capacitor bank.

IGBTs are pretty much designed for what you want to do: short, high current pulses at high voltages. The type Tayda sell is fine for what you want to do. When it coms to switching, I've just thought of a simpler method: Connect the - of a 9V battery to the emitter of the IGBT, and use an ordinary low voltage switch to make/break a connection between the + of the battery and the gate of your IGBT. The battery will last for ages because the gate of the IGBT draws hardly any current (500nA max, according to the datasheet). This may be a bit more complex than PRRs suggestion, but in the long run it will probably outlast the capacitors. That, and it won't need replacing every 12 firings, which would be a pain.