Re: The Geek Group High Voltage Capacitors, making a HV switch, (fwd)

["High Voltage list" <hvlist@xxxxxxxxxx>]

Original poster: Steven Roys <sroys@xxxxxxxxxx>



---------- Forwarded message ----------
Date: Thu, 09 Nov 2006 19:46:33 -0800
From: Jim Lux <jimlux@xxxxxxxxxxxxx>
To: High Voltage list <hvlist@xxxxxxxxxx>, hvlist <hvlist@xxxxxxxxxx>
Subject: Re: The Geek Group High Voltage Capacitors, making a HV   switch,
      (fwd)

At 01:55 PM 11/9/2006, High Voltage list wrote:
>Original poster: Steven Roys <sroys@xxxxxxxxxx>
>
>
>
>---------- Forwarded message ----------
>Date: Wed, 8 Nov 2006 21:44:26 +0800
>From: Peter Terren <pterren@xxxxxxxxxxxx>
>To: High Voltage list <hvlist@xxxxxxxxxx>
>Subject: Re: The Geek Group High Voltage Capacitors, making a HV  switch,
>       (fwd)
>
>There are interesting design considerations with high current pulses.
>I have two caps joined with 1 1/2 inch busbar, yet they were making big
>sparks between the cases when they were fired at 5kJ. A few calculations
>show why.
>
>Consider the inductance of 20 cm of busbar compared with 60 cm wire in 3
>turns of a can crushing coil (?300 nH).  Without calculating it I would
>guess at the busbar being 10 % of the inductance of the coil.  So at 6 kV
>across the coil (busbar) would be 600 V at perhaps 80 kA peak current. Short
>that out and you will have plenty of sparks.

Yes.. "interesting" stuff happens with high current fast pulses.


>Good luck trying to get 20 caps to discharge in 10us. I have wire lengths of
>21, 25 and 30 inches for the full short circuit discharge path.  The
>resonant frequency of my 3 caps with a short circuit is 25kHz ie period of
>45uS. The first rise of the half cycle is 12us but that energy gets stored
>in the wire inductance and so current will lag. With 20 caps your wiring has
>to be longer.

Typically, you have equal length buswork from the caps going to a 
common join point (assuming you're paralleling them).  I use copper 
pipe for buswork: it's round so it doesn't have edges for corona, and 
for fast pulses, skin effect means that the middle of conductor isn't 
used anyway.  Especially since copper has gotten so expensive!

If I were building something like this today, I'd seriously look into 
using aluminum.  For a given size and weight it's a LOT cheaper, so 
you can make up for the 60% higher resistivity with much lower inductance.