Re: Gap Dwell Times (formerly: Beating Solved)

To: tesla-at-pupman-dot-com
Subject: Re: Gap Dwell Times (formerly: Beating Solved)
From: Bert Hickman <bert.hickman-at-aquila-dot-com>
Date: Tue, 01 Oct 1996 07:20:53 -0700
Tesla List wrote:
> 
> >From MALCOLM-at-directorate.wnp.ac.nzMon Sep 30 22:00:22 1996
> Date: Mon, 30 Sep 1996 08:56:14 +1200
> From: Malcolm Watts <MALCOLM-at-directorate.wnp.ac.nz>
> To: tesla-at-pupman-dot-com
> Subject: Re: Gap Dwell Times (formerly: Beating Solved)
> 
> Hi Bert, all,
> 
> > > >   I did some research this weekend, pouring through a Corum brothers
> > > > monograph and some old posts to this newsgroup and found out that the
> > > > optimal dwell is 1/(k*Fr), *not* 1/(2*k*Fr). In other words, the spark
> > > > gap should conduct for only 1/2 of the superimposed beat-frequency
> > > > envelope. Primary current peaks at that point, and maximum energy is
> > > > transferred into the secondary.
> > >
> > > I started from k approx dF/F, and ringup time = 1/2dF so you can see
> > > how I derived that. The problem when cutting the gap off (if you
> > > could) when Ip is maximum is that with k<1, most of the primary
> > > energy is coupled to the primary. I tried doing exactly this with the
> > > MOSFET gap and the spikes hit the roof. Virtually none of that energy
> > > was coupled to the secondary. The spikes in a real gap would have re-
> > > ignited it anyway. I wonder if they have actually tried doing this?
> >
> > <SNIP>
> > Malcolm,
> >
> > It looks like they did - in the appendix of their booklet "Vacuum Tube
> > Tesla Coils", page IV-10, they present a chart which shows experimental
> > data points connected with a curve. Their system had Fu=123.33kHz and
> > Fl=74.18 kHz, and k=.46(!). The predicted "best case" quenchtime was
> > 1/(2*(Fu-Fl))=10.17 uS. They tested dwell times from slightly less than
> > 5 uS through 55 uSec, finding that the peak secondary output occurred at
> > the expected value of 10 uSec.
> 
> Well that sounds right. 1/2dF or 1/2kFr. Important to note though that
> the time quoted is related to operating frequency. Even at k = .46,
> quenching at the first zero crossing makes the transfer far from
> complete. At this k they would have been quenching around the end of
> the third primary half cycle. Funny k to use though. The transfer
> while done in an integral number of half cycles would not go entirely
> to completion IMO. Did they do that with a spark gap or some other
> device in its place?
> 

Unfortunately, they don't say. The entire section discusses spark-gaps
and quenchtimes, so I would guess they used a hihg-speed rotary.


-- Bert --

<SNIP>