From: Paul
Date: Fri, 11 Oct 2002 12:24:54 +0100
Subject: Re: [TSSP] Quiz, Sec diameter question
Greg Leyh wrote: > with very thin [< 1:10 aspect] coils, it becomes almost > impossible to set up a *practical* primary geometry > that provides more coupling than k=0.10 or so. Indeed. Perhaps this underlies the problems experienced at high k factors. It's not the high k per se, but the concentration of coupling onto one small fraction of the secondary. It certainly causes more energy to go into the higher modes of the coil, because nature selects the mode amplitudes so that their sum gives a total voltage profile that includes the 'hump' caused by primary induction. It's the same process that occurs in Fourier synthesis of a waveform with a peak in it. Lots of harmonic sinusoids must be added together in order to construct the peak. But in our case, the 'waveforms' being added are the spatial distributions of voltage and current along the coil (as opposed to the time domain) and the component signals are coil overtones rather than harmonics. But the principle is the same. The mathematician would say that the overtones (plus perhaps a DC term) form a complete orthogonal set of basis functions. Thus, even without unpleasant breakdown, efficiency is reduced. The question of the coil's ability to withstand transients must also have a bearing on primary design. We might not be able to do much with this until we can predict with some confidence, the likely potential that some part of the system will break down. So the directly relevant issues are a) Maximising efficiency by minimising the energy that goes into higher overtones. b) Avoiding direct turn-turn breakdown due to primary induction. c) Avoiding radial pri-sec breakdown. d) Response to discharge transients. and indirectly, the usual primary design issues that involve primary and gap losses. There's much work to be done using modeling to try to find primary shapes which spread out their induction without themselves becoming targets for discharges from the topload or coil. Frustrating that we have the tools but not the time at the moment. Would be a good topic for anyone who wants to get stuck in to some serious modeling. -- Paul Nicholson, --
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.