Gap models

From:  Bill the arcstarter [SMTP:arcstarter-at-hotmail-dot-com]
Sent:  Tuesday, April 14, 1998 8:14 AM
To:  tesla-at-pupman-dot-com; bert.hickman-at-aquila-dot-com
Subject:  Gap models

Bert Hickman wrote:

>As I'm sure you know, modeling a spark gap is not a trivial task, and
>even the best models are only approximations due to the various
>non-linear characteristics! As a first approximation, try using
>back-to-back series connected 5-8 volt zener diodes in series with a
>controlled switch - and set the model so that the switch opens only 
>the primary energy (1/2 LI^2 + 1/2 CV^2) is zero. BTW, what value 

I don't think the gap completely opens when the net primary energy is 
zero.  It seems to open earlier.  If I had to model it I suppose I'd 
examine the peak primary current - once the peak drops below a certain 
(arbitrary) value per cycle - the gap goes fully open.   For example - 
if you operate a coil in 'single-shot" mode, ie slowly charge the cap up 
using DC until the gap fires, you will find that the cap still contains 
some energy.  

(Those retentive mathemeticians would also point out that, for a damped 
lossy system, the net energy is *never* zero, but is actually an 
exponentially decreasing function which approaches zero.  But I won't 
mention that.  :)  )

For the particular experiments I did, the cap ended up with about 80-100 
volts, down from about 15000 (gaps fired at 15000).  Another thing - 
final voltage was randomly either positive or negative - depending on 
exactly when the gap decided to quench out.  

Alternately, you could set the gaps wide, and charge up the cap.  Then 
remove the charger and trigger the gap.  I did this once by dropping 
some iron filings through the gap, but there are certainly more elegant 
ways of accomplishing this!

I'm curious what other folk have used for a gap model...?

My $0.02.  This is just one of many nasty nonlinearities of the tesla 
spark gap.

-Bill (a.k.a. The Arcstarter)