From: Paul
Date: Sun, 12 May 2002 21:18:39 +0100
Subject: [TSSP] Racing arcs
Hi All, I thought I'd start this seperate thread on racing arcs, as it diverges from the stress factor and topload breakout threads. We see Bart's coil developing at most 550kV averaged over the coil length of 103cm to give 5.3kV/cm. This h/d=4.8 coil has an unloaded stress factor of 1.38 and with topload and primary drive, the model reports a peak gradient of around 7.6kV/cm occuring at about 60% height. None of these figures approach a 26kV/cm gradient, even if we assume quite a large 'corrugation factor', so we wouldn't expect to see breakdown along the coil. Bart wrote: > I ran the coil as is at several different voltages. I got racing > arcs every time and multiple streamers, relatively short. Ok, that's for the in-tune coil at 13 turns... > I then detuned out about 7% to turn 14 from 13. Still > got racing arcs, however, brighter. Ok, the racing arcs got stronger as you detuned the primary. > lowered the toroid to center at 70" above ground. This puts the > sec top winding about an inch below the toroid center (quite a > bit of shielding at this point). This pulls the detune to > 10.7%. Low and behold, no racing arcs. So a further detune (by raising sec Fres) stops the racing arcs. Bart, were the racing arcs localised to one part of the secondary? > Notice here, I didn't change coupling. Indeed. I'm beginning to think that coupling doesn't have a lot to do with racing arcs. None of the models in pn040502 show any sort of drastic increase in the voltage gradients. I'm interested in exploring the idea that the energy for the racing arcs originates in higher frequency primary resonances that are being coupled through to the secondary. Let me try and summarise the hypothesis. At frequencies well above the operating frequency the primary appears as a distributed resonator grounded at both ends. It will therefore have a spectrum of resonant modes and some of these may receive energy from the bang due to the step nature of the spark gap closure and perhaps the subsequent non-linearity of the gap arc. These modes will couple both inductively and capacitively to the secondary and may induced longitudinal voltage stresses sufficient to cause racing arcs. We have an example of strong ringing at the start of the event in http://www.abelian.demon.co.uk/tssp/md110701 with a set of component frequencies typical of a distributed resonator, and much too far apart to be due to the secondary. Thus they are most likely to be from the primary, since that is the only strongly coupled object with a chance of having its lowest mode of circa 2Mhz. In Thor we see the initial base current ringing peak at around 10 amps, which is only 50% below the peak current at the operating frequency! By looking at the secondary's forced response at this frequency, this base current will develop around 60kV induced voltage across the bottom 6cm of the secondary. We may find that we have to consider capacitive pri-sec coupling, and a series resonant circuit involving the primary distributed winding, the pri-sec cap, and the lower bit of the secondary distributed winding. Further, we may find that the racing arcs manifest differently if the secondary happens to have a mode that coincides with one of these excited primary modes. That might explain the racing arcs coming and going as Bart's secondary tuning was raised by lowering the topload. Adjusting the coupling may appear to affect the racing arcs simply because it is adjusting the frequencies of the primary modes relative to the secondary. I tried making the model process the primary as a distributed winding but it broke due to the few turns. It would only have been a qualitative test anyway, because I've no way to estimate the boundary conditions at the ends of the primary. Perhaps an application of tcma to a typical primary is called for? -- Paul Nicholson, --
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.