From: Paul
Date: Thu, 19 Oct 2000 23:47:09 +0100
Subject: Re: [TSSP] Surprising secondary voltage profiles
Terrell W. Fritz wrote: > The profiles look remarkably different. Yes, the hand plotted profile shows the maximum gradient near the bottom and follows the expected convex V profile, in which the gradient steadily decreases with height, while the computer plot reflects the concave profile used by E-Tesla5 in which the gradient steadily increases up the coil. I just don't know whether creeping up on the coil with a grounded electrode is a valid way to plot the field. > If the E-Tesla5 secondary voltage contour was far off, it should > give a significant calculation error. The program gives an Fo > frequency of 1566.4 kHz. I measured the coil at Fo = 1597.6 kHz > and Mark got 1.615MHz. Errors of -2.0% and -3.1%. Reasonable agreement of E-Tesla5 predictions - obviously the V-profile coded into the program is appropriately weighting the capacitance somehow. > One thing I worry about is how the capacitance is distributed on > the coil. Near the base, the capacitance is the ground plane is > high since it is so close. Near the top of the coil, the > capacitance is also high since the top of the coil needs to charge > the large top area above the coil. Yes, the laplace solver which provides the external capacitance distribution in 32 steps along the coil always arrives at the 'bathtub' shaped distribution in the absence of a toroid. When a toroid is fitted, the rising coil capacitance at the top almost disappears. > The top turn of the coil has substantial current in it since it > is charging so much area. Yes, there is some current left at the top, but its smaller than that lower down, ie it doesn't increase towards the top - there's no mechanism to supply the current in the top half of the coil. > Robert Jones (wherever he is?) was struggling with how to determine > this capacitance distribution in his secondary model last I heard. Well we've moved on quite a bit now from that particular problem. > I think the high capacitance near the base of the coil suppresses > voltage buildup and the high currents spraying into space near the > top of the coil tend to make that nasty concave curve: I disagree, I would expect the high capacitance near the base to induce a steeper rise of voltage in that region. See http://www.abelian.demon.co.uk/pn1710/ example 5. > http://users.better.org/tfritz/VoltDistBare.jpg > > It is interesting that this curve implies the voltage is greatest > just before the end of the coil!! I still don't belive that... Me neither, would imply a phase reversal of the current at that point. > By setting the capacitance distribution in various ways, you can > probably get just about any curve you want. I'm don't think thats the case. Due to the form of the coefficients in the differential equation (pn1310 5.5) for the V and I profiles in the 1/4 wave, there are constraints on the shape of the solutions. I've been investigating these but I'm not prepared to commit to any conclusions just yet. > I think it can be equated to the spring ruler analogy but I have > never pinned that down. I doubt it, the bent ruler analogy is just a bit of handwaving, I've demonstrated the differential equations just dont come anywhere near. > I am still working on ways to measure the profile. So far most > things tend to see field stress rather that absolute voltage. Yup, we really do need some unequivocal V profile measurements, I hope all you experimenters out there are getting down to the job. Us 'theory guys' need to be brought down a peg or two now and then, and unassailable measurements are a good way to do that. I reckon the measurement can be approached in more than one way, but short of tapping into the coil, its going to be some sort of field measurement, its just a choice between close-in and far-out probing. My vote goes for the close-in approach as there are fewer unknown quantities involved. Regards, -- Paul Nicholson, Manchester, UK. --
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.