6.5kV 8345 Microfarad Capacitor

[AF6LJ]

That's a good part but it takes 3ms to turn on...

I think it is 3us. It has interdigitated amplifying gates, it should be VERY fast and very resilient to filamentation.
It's rated 26*10E6 A2s peak I2T when fully depleting capacitor's energy, assuming load discharges cap bank in 100us, then the SCR can handle 500kA RMS discharging current.
Assuming RMS=average (engineering approximation) and current decays linearly, then peak current of 1MA can be achieved, as long as load inductance and resistance allow.

You are right I misread the datasheet.
It might do the job, I would feel more comfortable with two of them in parallel.

[AF6LJ]

What you're talking about are various kinds of things called "closing switches" which are NO switches as opposed to "opening switches" that are NC.

There are several kinds of high voltage / high power closing switches used for these sorts of applications for scientific research, things like university / lab scale physics experiments etc.

Here's some overview:
http://uspas.fnal.gov/materials/09VU/PPE_Switches.pdf
http://www.dtic.mil/get-tr-doc/pdf?AD=ADA639879

"Cheap and simple" would be something like a high pressure triggered spark gap which can switch in ns to us time frames.
There are physical scaling laws depending on the switching time, the hold off voltage vs. gap length and the pressure.  Higher pressures let you have higher stand off voltages in short gaps and less switching time.
http://ece-research.unm.edu/summa/notes/SwN/SwN28.pdf
http://www.iaea.org/inis/collection/NCLCollectionStore/_Public/20/085/20085510.pdf

Then as mentioned you have various kinds of thyristors, thyratrons, ignitrons, etc.
Some are designed to be quite fast switching and capable of high power / high voltages.

There are also light activated SCRs and light activated bulk semiconductor switches (not thyristors) that they trigger through fiber optics with some kind of somewhat high pulse power laser.

Then there are some exotic pulse power specific sorts of semiconductor switches such as the fast ionization dynistor:
http://www.fidtechnology.com/

It is an interesting field, maybe one day I'll get to work more with the "lab level science" stuff that uses such apparatus.

Very Cool.
Thanks for passing that along, the technology has evolved to quite a high level...

[amyk]

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*!

The sound is the shock wave caused by the rapid expansion of air, heated by the arc which forms when the coil explodes and vaporises.

[CatalinaWOW]

There is all sorts of fun to be had discharging high voltage capacitor banks.  In my formative electronics years I assembled a largish bank at about 600 VDC using every salvaged electrolytic I could lay my hands on.  I don't remember, but it was probably a couple of thousand microfarads.  The discharge "switch" was the contacts on the end of two wands.  Inductance?  ESR?  Anything like that was far into my electronics future.  Touch them to the object of interest and watch things happen.  What happened was widely varied depending on the materials and other things that were uncontrolled in my crude experiments.  One of the more interesting was discharge into a tungsten light bulb filament.  Left a beautiful star of filaments hanging in the air.  Each filament maybe half a meter long, finer than spider silk, but possible to catch and pull proving that they weren't just smoke trails.  Never could duplicate that.  Using fairly clean water and doing the discharge underwater had interesting effects as the mass of the water provided confinement, while the steam produced provided pressure.

Of course all of this was terribly dangerous and I couldn't possibly have survived.  As the Mythbuster's used to say - "Don't try this at home.   I was an expert. "