From: Paul
Date: Fri, 10 Nov 2000 23:53:44 +0000
Subject: [TSSP] Transients in primary waveforms
Terry wrote: > My paper at: > http://63.229.238.62/TeslaCoils/MyPapers/spark1/sparkgap.html > has some waveforms in it that may be of use. Terry, Referring to sparkgap.html, last image, can you tell us where abouts in the primary circuit the current probe was inserted? I'm not surprised at the apparent amplitude of these current spikes, although I disagree with the short wide strips suggestion - that seems to me to be a recipe for greater parasitic capacitance and lower series inductance, thus bigger and shorter current spikes. I'm assuming the current spikes are undesirable from the RFI point of view, and from an efficiency point of view the high currents are a source of un-necessary losses. Just a few rambling thoughts about these current spikes... If the primary tank, gap, and coil were all connected by negligible lengths of wire, there would be no problem. Realistically there must be some separation, so there is a 'line' capacitance to charge up, and a line inductance which determines the rate at which the charging occurs. The result is the transmission line ringing seen in the last but one image of sparkgap.html. The frequency and peak current could be dropped by introducing some series inductance close to the gap, but the energy in the ringing would be unaffected. Of course the ringing could be suppressed by dampening the extra series resistance with some resistance, and finding the right value would be equivalent to finding the right terminating impedance for the primary transmission line, but that's of no benefit to efficiency. Off hand I can't see a way to prevent the initial ringing as the gap first fires. Let's try to estimate the energy lost due to the ringing. From the last image in sparkgap.html, and taking the amplitudes of the initial spike as, say, I = 100A and V = 2000V, and if L and C are the parasitic reactances involved in the ringing, we have roughly, V/I = 2000/100 = sqrt(L/C). and, if the resonance is at around 50Mhz, sqrt(LC) = 1/(2.pi.50e6) >From these we get L = 60nH, C = 150pF. With a peak V of say 1000V, thats 75 uJ injected into the ringing twice each cycle. Terry, you don't say what the bang energy was in these measurements, and the figures above are for spikes beyond the frequency response of the probes, but on the face of it, it looks like the parasitic ringing loss is probably quite small in comparison to the energy carried at Fres. Cheers, -- Paul Nicholson, Manchester, UK. -- -- Paul Nicholson, Manchester, UK. --
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.