Re: Voltage multiplier capacitors (fwd)

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

Original poster: Steven Roys <sroys@xxxxxxxxxx>



---------- Forwarded message ----------
Date: Wed, 21 Mar 2007 20:19:03 -0700
From: Jim Lux <jimlux@xxxxxxxxxxxxx>
To: High Voltage list <hvlist@xxxxxxxxxx>
Subject: Re: Voltage multiplier capacitors (fwd)

At 07:47 PM 3/21/2007, you wrote:
>Original poster: Steven Roys <sroys@xxxxxxxxxx>
>
>
>
>---------- Forwarded message ----------
>Date: Tue, 20 Mar 2007 16:15:22 -0700 (PDT)
>From: Mike <megavolts61@xxxxxxxxx>
>To: High Voltage list <hvlist@xxxxxxxxxx>
>Subject: Re: Voltage multiplier capacitors (fwd)
>
>Are you going to have this NST multiplier with both a positive and a 
>negative voltage?   If so,  each stage's capacitors with need to 
>stand of 1.414x 12k/v  or about 17kV   As far as the capacitance 
>value,  I'm not too sure, but I would think the first stage should 
>match the impedance of the transformer to get the most from it. The 
>diodes should prevent the resonant rise, but again,  I'm not 
>sure  lol, so you might want the caps to be a little more than the 
>impedance match (which I calculated at 6.6nF...so maybe use 8 or 
>9nF).  If you want a single positive output, then you might want a 
>full wave rectifier on each output and use the ground for the 
>reference.  Then your caps for each stage would only have to be 
>about 8.5kV.   I hope someone else has a better handle on this than 
>I.   If I'm right about the full wave rectifier on each leg, then 
>you'd need four times the capacitance since the voltage is 
>halved.   On subsequent stages,  I think the value can go down because the
>  overall voltage goes up and current down, but all the CW 
> multipliers I've seen use a constant value for each stage, so it 
> would be safer to do that.
>   Mike

http://home.earthlink.net/~jimlux/hv/cw1.htm has some schematics and 
design equations, including one for a center tapped transformer.

It's a bit complex to figure out what size caps might be optimum for 
a hugely inductive source like a NST.     As the stack charges from a 
stiff low impedance, the current flow starts to look like narrow 
pulses (since it only flows when the transformer voltage is > the cap 
voltage, and so on, up the stack).  A ballasted source like a NST 
might act more like a constant current source, so the charging 
current waveforms might look different.

I'd just pick some caps of convenient values and build it and try it.