6.5kV 8345 Microfarad Capacitor

[VE7FIM]

Not sure if this is the right place for this, but wow — *this* is a capacitor...

http://www.ebay.com/itm/General-Atomics-32947-6-5kV-8-1F-Energy-Discharge-Capacitor-/292085271551

If my calculations are correct, when fully charged, this thing would store 1.71 x 10^8 Joules, and if completely discharged in 0.1 seconds, would release 1.71 gigawatts of power.

[mariush]

It's not 8.1 Farads, it's 8160uF +/- 10% or 8345 uF  (measured) as you can see on the label.

That's 0.008345 Farads, quite a lot less than 8 Farads.

[VE7FIM]

That makes a lot more sense...

Good catch.

Still, 1.71 megawatts of power is still pretty impressive...

[T3sl4co1l]

FWIW, at 0.3uH ESL, the short circuit time constant would be on the order of 77us (1/4 wave time, of a ringing waveform).  The resonant impedance is 6mohm, so the peak current is a bit less than 1 megaampere, so the peak power (instantaneous V*I) is on the order of 5GW.

Conclusion: I don't recommend touching the terminals when it's charged.

You would normally use it at much less than this, since you accumulate inductance quickly with all that bus bar (what's shown is probably about 1uH, total), and more again from whatever the load itself is.

Tim

[KE5FX]

FWIW, at 0.3uH ESL, the short circuit time constant would be on the order of 77us (1/4 wave time, of a ringing waveform).  The resonant impedance is 6mohm, so the peak current is a bit less than 1 megaampere, so the peak power (instantaneous V*I) is on the order of 5GW.

Still need to take ESR into account, don't you?

[T3sl4co1l]

FWIW, at 0.3uH ESL, the short circuit time constant would be on the order of 77us (1/4 wave time, of a ringing waveform).  The resonant impedance is 6mohm, so the peak current is a bit less than 1 megaampere, so the peak power (instantaneous V*I) is on the order of 5GW.

Still need to take ESR into account, don't you?

As low as Zo is, ESR should be lower still -- it's a film cap.

The Q should be low, in the single digits, but not 1 or below.

Low Q causes the ringing frequency (pseudofrequency) to be slightly lower (the change goes as 1/(RC)), and the ringing to decay fairly quickly, so that the first peak is noticeably lower than V(charge) / Zo (which, if you check my math, you'll notice I made a guess at ).

Tim

[bitwelder]

Once fully charged, how long would you need to wait for it to self-discharge until it's down to safe levels, if nothing is plugged to its terminals?

[T3sl4co1l]

Once fully charged, how long would you need to wait for it to self-discharge until it's down to safe levels, if nothing is plugged to its terminals?

Take a regular 1uF 250V film capacitor, charge it to 250V.  How long would you wait for self-discharge?
Take 26 of them and put them in series.  Charge the series string to 6.5kV (assume they charge evenly).  How long would you wait for self-discharge?
Take 26 * 8345 of the strings and put them in parallel.  Charge the mass to 6.5kV (assume they charge evenly).  How long would you wait for self-discharge?

In each case, you have the same voltage, on the same film cap.  Size doesn't matter.

Tim

[AF6LJ]

FWIW, at 0.3uH ESL, the short circuit time constant would be on the order of 77us (1/4 wave time, of a ringing waveform).  The resonant impedance is 6mohm, so the peak current is a bit less than 1 megaampere, so the peak power (instantaneous V*I) is on the order of 5GW.

Conclusion: I don't recommend touching the terminals when it's charged.

You would normally use it at much less than this, since you accumulate inductance quickly with all that bus bar (what's shown is probably about 1uH, total), and more again from whatever the load itself is.

Tim

Looks like just what the engineer ordered for the aluminum can crusher I was dreaming of building. Now to get on designing that ballistic switch.

[SeanB]

I would be a lot more worried about dielectric absorbtion that self discharge, that size will still be able to develop a pretty big punch even after a week of being stored shorted, then left open circuit for a few hours, after soaking for a day at full rated voltage.

Best use i can see for one of these is to make a pulsed Bussard fuser though.

Does not ship to my country though, will have to see if I can persuade the scrappie to give some of those power line ones, just to use as decoration.

Looks like it also comes with a nice puck thyristor array switch as well, saves buying one if it is included.

[james_s]

You could probably make a hell of a coin shrinker with one of those. I'm no stranger to high voltage stuff but a capacitor like that even makes me nervous. There is no second chance if you have an accident with that. It's likely to turn whatever body part contacts the terminals into a pulpy mist and you'd probably be dead before you hit the ground.

[KE5FX]

I would be a lot more worried about dielectric absorbtion that self discharge, that size will still be able to develop a pretty big punch even after a week of being stored shorted, then left open circuit for a few hours, after soaking for a day at full rated voltage.

Best use i can see for one of these is to make a pulsed Bussard fuser though.

Does not ship to my country though, will have to see if I can persuade the scrappie to give some of those power line ones, just to use as decoration.

Looks like it also comes with a nice puck thyristor array switch as well, saves buying one if it is included.

It would make sense to keep a few megs of bleeder resistance permanently connected across one of these, I think.  I can't see much upside to leaving 150+ kJ worth of energy sitting around for months or even years at a time.  It'd be like keeping a live grenade in your desk drawer at work.

[PartialDischarge]

If I lived in the US I'd definitively buy this, not only the capacitor is included but clamped ABB thyristors+firing circuitry+oscillator board, so nice.

[AF6LJ]

You could probably make a hell of a coin shrinker with one of those. I'm no stranger to high voltage stuff but a capacitor like that even makes me nervous. There is no second chance if you have an accident with that. It's likely to turn whatever body part contacts the terminals into a pulpy mist and you'd probably be dead before you hit the ground.

You could but the secret to that is the same as the secret to making a good aluminum can crusher. You need a switch that will go from on to off in a fraction of a microsecond.
People have tried to build can crushers with SCRs but they take for ever to turn on. (milliseconds) during that time they operate in a linear mode and much of the useful energy is converted to heat.
In the 1980s the Russians were trying to build a a capacitive discharge EMP generator.
They tried all manner of switch mechanisms They settled on a ballistic switch, the moving contact was driven home with explosives, a compressed gas discharge system might have worked as well.. Both systems eliminate the rise time issues that exist with SCRs and other thyristers.

I don't know if it is still up there...
There was a YouTube video of a guy who built a can crusher with a big cap and a spring loaded switch... His system had other problems like the wire size used for the coil was too small and on one occasion his coil blew up.

This stuff has fascinated the hell out of me, I use to work next door to one of Maxwell Industries plants. They built an EMP generator that took a city block offline one day due to ineffective shielding. They also built a gun that shot a .4" aluminum covered Mylar bullet that could penetrate 3/4" steel decking. There was an article in the local paper about that gun, then nothing heard.... Makes me wonder if the third generation of that gun isn't orbiting up in space above us...

Some day I am going to build my can crusher. 

[james_s]

I saw a coin shrinker in person at a Maker Faire years ago, it used 3 suitcase sized capacitors that were charged by a large neon sign transformer. The switch was a spring loaded thing with large tungsten contacts, it seemed to work well but the coils were single use, they vaporized in the process. What surprised me was the noise, the discharge sounded like a large caliber handgun.

[PartialDischarge]

There was a YouTube video of a guy who built a can crusher with a big cap and a spring loaded switch... His system had other problems like the wire size used for the coil was too small and on one occasion his coil blew up.

This guy did it that way because he didn't know any better, the best way to build a can crusher, or anything that requires a fast total discharge of a kV capacitor with kAs of current flow is a thyristor: extremely fast on times, in 10us you can have thousands of amperes flowing provided the correct gate firing, and the voltage across will easily drop from a few kVs to a few volts (provided appropriate clamp force of press-pack devices and good metal to metal contact)

[Kilrah]

What surprised me was the noise, the discharge sounded like a large caliber handgun.

You don't need that much, in the past I've used a direct short of a simple large photo flash cap (1000uF 350V) charged to rectified 230V mains precisely to simulate a gunshot in a theatre play. Worked a treat, louder than those pyro caps... and reusable 

Pretty sure I still have the cap in a drawer. That was about 20 years ago and the cap was taken out of a broken flash that must have been at least 10 years old back then...

[AF6LJ]

There was a YouTube video of a guy who built a can crusher with a big cap and a spring loaded switch... His system had other problems like the wire size used for the coil was too small and on one occasion his coil blew up.

This guy did it that way because he didn't know any better, the best way to build a can crusher, or anything that requires a fast total discharge of a kV capacitor with kAs of current flow is a thyristor: extremely fast on times, in 10us you can have thousands of amperes flowing provided the correct gate firing, and the voltage across will easily drop from a few kVs to a few volts (provided appropriate clamp force of press-pack devices and good metal to metal contact)

The ones I have seen in the 10,000A and up take milliseconds to turn on.
In this day and age you may be able to do it with FETs, but that is a lot of current you are dumping into a one or two turn loop, make from copper buss bar. 

[PartialDischarge]

There was a YouTube video of a guy who built a can crusher with a big cap and a spring loaded switch... His system had other problems like the wire size used for the coil was too small and on one occasion his coil blew up.

This guy did it that way because he didn't know any better, the best way to build a can crusher, or anything that requires a fast total discharge of a kV capacitor with kAs of current flow is a thyristor: extremely fast on times, in 10us you can have thousands of amperes flowing provided the correct gate firing, and the voltage across will easily drop from a few kVs to a few volts (provided appropriate clamp force of press-pack devices and good metal to metal contact)

The ones I have seen in the 10,000A and up take milliseconds to turn on.
In this day and age you may be able to do it with FETs, but that is a lot of current you are dumping into a one or two turn loop, make from copper buss bar.

Without knowing the exact dimensions of the thyristors in the cap bank its difficult to know the model, but it is an ABB one and could be this one:

https://library.e.abb.com/public/f15d6377c6c2ffb183257c63004dc804/5STP%2026N6500_5SYA1001-07%20Mar%2014.pdf

Have a look at the specs and get marveled, that thing can pass 100kA in less than 1ms.

[RGB255_0_0]

How capable is this one compared to Photonicinduction's one he used to obliterate stuff? It was the quickest way of making a smoothie I've seen.

[KE5FX]

A hydrogen thyratron could be a possibility, but they seem to top out at 10K - 100K amps.  Sure can't complain about the rise (fall) time, though.

When you start looking at data sheets from EG&G, it's probably time to step away from the workbench for a while and get some fresh air...

[AF6LJ]

There was a YouTube video of a guy who built a can crusher with a big cap and a spring loaded switch... His system had other problems like the wire size used for the coil was too small and on one occasion his coil blew up.

This guy did it that way because he didn't know any better, the best way to build a can crusher, or anything that requires a fast total discharge of a kV capacitor with kAs of current flow is a thyristor: extremely fast on times, in 10us you can have thousands of amperes flowing provided the correct gate firing, and the voltage across will easily drop from a few kVs to a few volts (provided appropriate clamp force of press-pack devices and good metal to metal contact)

The ones I have seen in the 10,000A and up take milliseconds to turn on.
In this day and age you may be able to do it with FETs, but that is a lot of current you are dumping into a one or two turn loop, make from copper buss bar.

Without knowing the exact dimensions of the thyristors in the cap bank its difficult to know the model, but it is an ABB one and could be this one:

https://library.e.abb.com/public/f15d6377c6c2ffb183257c63004dc804/5STP%2026N6500_5SYA1001-07%20Mar%2014.pdf

Have a look at the specs and get marveled, that thing can pass 100kA in less than 1ms.

That's a good part but it takes 3ms to turn on...
That is going to get really hot in those three milliseconds, and that means that energy isn't getting to the load.
But it is faster than many I have seen that are in the 10-16 millisecond range. 

[AF6LJ]

In this day and age you may be able to do it with FETs, but that is a lot of current you are dumping into a one or two turn loop, make from copper buss bar.

FETs are voltage controlled current sources. They only behave like a resistor (Rdson) in deep linear mode, which means their drain current is not too high (determined by Vgs>>Vds).
When drain current is not limited, they behave like constant current source. Therefore, you cannot use a 1mR device and apply 1kv on D-S, and hope it will conduct 1MA. It will self limit at a few hundred amps, then desaturate, and then explode.
SCRs (high speed ones designed with filamentation mitigation and self driving firing mechanism) and firing IGBTs (designed specifically for firing applications, such as photo flash IGBTs and pulse power IGBTs) are still the best silicon device for this purpose. Vacuum switches are faster in terms of carrier storage and mobility, but have higher inductance.
I've never played with vacuum switches, but my boss was a pulse power guy a few decades ago, and he always talk about his fancy stories about vacuum switches.

I am thinking it would take many in parallel. 
The turn-on time would be as fast as the slowest one, and the drop across many in parallel would be manageable, but it wouldn't be my first choice.

[AF6LJ]

A hydrogen thyratron could be a possibility, but they seem to top out at 10K - 100K amps.  Sure can't complain about the rise (fall) time, though.

When you start looking at data sheets from EG&G, it's probably time to step away from the workbench for a while and get some fresh air...

The more you look into solid state devices, the more a ballistic switch looks like the right choice.
Unless you want to get into a electrically activated pneumatically actuated, vacuum switch.
And at that point; I'll stand back and watch the manufacturing and testing of the switch.
This could be a mechanical engineer's nightmare.

[james_s]

What surprised me was the noise, the discharge sounded like a large caliber handgun.

You don't need that much, in the past I've used a direct short of a simple large photo flash cap (1000uF 350V) charged to rectified 230V mains precisely to simulate a gunshot in a theatre play. Worked a treat, louder than those pyro caps... and reusable 

Pretty sure I still have the cap in a drawer. That was about 20 years ago and the cap was taken out of a broken flash that must have been at least 10 years old back then...

The coin shrinker was probably orders of magnitude louder than that. It was outdoors in the open, sounded like a .44 Magnum or maybe something bigger, it was *loud*!