From: Bert Hickman
Date: Fri, 18 Oct 2002 06:08:05 -0500
Subject: Re: [TSSP] Top Voltage
Paul, Please don't beat yourself up on this little bump in the road. Heck, if this stuff were easy, someone else would have done it already. Remember that the present level of understanding and the powerful modeling tools, while not quite perfect, are a far cry from that which existed even a year ago. And lots of very smart folks have looked at Tesla Coils for well over 100 years without coming close to achieving the depth of understanding reflected in the present TSSP effort. Relax, and have a well deserved Guinness. You WILL find this minor SW bug and the quest will continue. Thanks, and very best regards, -- Bert-- -- Bert Hickman Stoneridge Engineering "Electromagically" (TM) Shrunken Coins! http://www.teslamania.com Paul wrote: > Hi All, > > There's something wrong with the modeling software. > > I sat down last night to run some thorough tests > against the formula for max gradient between sphere > and plane, North rpt, section 7, figure 7-2 and the > equation at the top of page 57. > > Thinking that previous tests have shown good agreement > I expected this to be a formality, but what I'm finding > is that tssp diverges from north for spheres that are > small compared with the sphere-plane distance. In fact, > tssp is ending up almost a factor of two high on the > field (which leads to breakout almost a factor of two low). > That factor of two sound familiar? > > This problem bypasses any issues over what the breakout > field actually is. The model should match the gradient > given by the equ on p57 very closely indeed. It doesn't. > > (It does diverge at small sphere-plane separation too, > but this is to be expected when the gap gets down to > the dimensions of an electrode surface element. However, > it should definately not diverge at large distance. > > This is baffling because I'm pretty sure I checked this > (quite some time ago, when we first started calculating > surface fields). The surface field comes out wrong, but > the C is accurate...equals a program screw up, somewhere. > > Something's wrong and I'm an idiot. I'll figure it all out > and report back later today, but for now, hold off on worrying > further about model/measure discrepancies on breakout. Was > too tired to figure it out last night, so either the mistake > is in the model, or the stuff I set up last night to run > the tests. > > -- > Paul Nicholson, > -- >
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.