Re: term understanding: voltage reversal.

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: Bert Hickman <bert.hickman-at-aquila-dot-net> 

Hi Christopher,

This specification refers to the amount of voltage reversal during a 
"ringing" discharge. Most commercial pulse caps are not designed for use in 
high Q ("ringing") circuits. This is why you'll find that they are 
typically rated for only a 10% or 20% voltage reversal. This rating is 
defined as the maximum rated reverse voltage swing expressed as a 
percentage of the DC voltage rating for the capacitor. Rapid voltage 
reversals create significant stress on the capacitor's dielectric system, 
particularly at the edges of the metal plates.

A sudden reversal of the capacitor's voltage is equivalent to the total 
voltage swing between the initial and worst-case reverse voltage during the 
discharge. Excessive voltage reversals can induce partial breakdown of the 
oil or the surface of the dielectric film, resulting in degradation of the 
dielectric and premature failure. The estimated % voltage reversal can be 
estimated if you know the "Q" of the circuit:

    % Vrev = 1.57*Sqrt(1/(Q^2-1))

A relatively low Q circuit will create a relatively high voltage reversal. 
For example, a tank circuit with a  Q of only 15 (typical for a well 
coupled TC primary) will apply a 90% voltage reversal across the tank cap. 
If the initial voltage stored on the capacitor is Vo, the effective stress 
"seen" by the dielectric will be  1.9*Vo.

Because of voltage reversal limitations, commercial pulse caps used in 
Tesla Coils are usually heavily derated. Knowledgeable coilers design their 
systems so that the peak DC voltage applied to the pulse cap is only ~50% 
of the cap's DC voltage rating. A good rule of thumb is that a commercial 
pulse cap used in a ringing tank circuit should have a DC rating that's at 
least 3X the expected RMS voltage of the HV AC supply.

This is why 35 kV Maxwell caps that are continually offered on eBay have 
been observed to prematurely fail when used in 15 kV RMS systems:
http://cgi.ebay-dot-com/ws/eBayISAPI.dll?ViewItem&item=2595652751

More robust 45 kV caps will survive in a similar environment: 
http://cgi.ebay-dot-com/ws/eBayISAPI.dll?ViewItem&item=2596506003

Since the problem is associated with oscillatory discharge in the tank 
circuit, the method of charging the cap (AC or DC) is irrelevant.

Best regards,

-- Bert --
-- 
--------------------------------------------------------------------
We specialize in UNIQUE items! Coins shrunk by Ultrastrong Fields,
Lichtenberg Figures (electrical discharges in acrylic), & Scarce OOP
Technical Books. Stoneridge Engineering -- http://www.teslamania-dot-com
--------------------------------------------------------------------

Tesla list wrote:

>Original poster: "Christoph Bohr" <cb-at-luebke-lands.de>
>Hello All.
>I came  across something I am no longer sure I understood that right:
>In pulse cap data sheets there is usually a point called "voltage reversal".
>To say it less technical I understand that voltage reversal causes stress on
>the cap and is undesireable if you like longelivity.
>But what exactly is this voltage revesal in a AC, sync gap, TC? is it:
>1.: The changing polarity of the carging current, i.e. the fact that I once
>carge the cap with the one polarity and during the next half sine wave to
>the other
>or
>2.: The changing polarity during the HF-"ringing". As the changes happen
>more often and more rapidly here I feel that this is the main voltage
>reversal relatet Stress on the cap.
>I fear this relates to point 1 and 2 as well. I only hoped that if it was
>point 1, one could reduce the stress on the cap by full-wave rectifying the
>xformer output to charge the C1. Would this work? I am not talking about a
>real DC supply with smoothing and charging inductor, only charging with
>pulsed dc to avoid the changing in polarity of the charging current.
>Probably I am missing something but the longer I think about it the more
>unsure I get...
>Maybe some of you guys can explain that to me.
>Best regards
>Christoph Bohr
>
>.