From: "Malcolm Watts"
Date: Wed, 27 Jun 2001 08:40:01 +1200
Subject: Re: [TSSP] Some Considerations
Hi Bert,
On 26 Jun 01, at 7:42, Bert Hickman wrote:
> Malcolm and Paul,
>
> I agree with virtually all of Malcolm's previous post except the part about
> recommending a spherical topload terminal. My thought process is as
> follows...
>
> It would seem that the key to getting an accurate measurement is to have a
> method of calibrating the divider ratio, or accurately computing the
> effective capacitance between the topload and the probe plate above to
> determine the divider ratio. In addition, the topload must be large enough
> to shield the resonator's E-field from the measurement plate, and that the
> measurement plate must be far enough away to avoid being struck or from
> forming corona on the plate. Clearance may need to be several leader
> lengths (or more?). There are some things we can do to improve measurement
> accuracy. For example, suppose we add a grounded plate (a "guard ring")
> around the sides of the capacitive probe and also behind it. This will
> virtually eliminate fringing fields or other extraneous E-fields, and
> should improve measurement repeatability once the unit has been calibrated
> with a given toroid size and spacing. If the top of the toroid had a flat
> section tangent to the outer curved surface, the toroid and measurement
> plate would form an (almost) ideal flat plate capacitor, the value of which
> should be computable with fairly good accuracy.
>
> However, the E-field drop off near the center of a regular toroid may
> actually be desirable. I suspect that using a toroidal topload would be
> more desirable than using a sphere especially for measuring voltages after
> breakout. As Paul mentioned, once breakout begins, the presence of leaders
> and streamers (particularly in the region between between the top terminal
> and measurement plate) will significantly distort the E-field and our
> measurements in changing and unpredictable ways. A spherical topload may
> actually cause leaders to concentrate precisely where we don't want them -
> at the top of the terminal - due to thermal effects.
>
> Suppose instead that we use a toroid of relatively large outer diameter
> combined with a comparatively small radius of curvature, or we purposely
> add a calibrated breakout sphere to stimulate breakout at a known e-field.
> In this case we should be able to direct most of the discharges laterally,
> or downward, minimizing E-field distortion between the toroid and
> measurement plate, and improving accuracy.
>
> This is getting very interesting... :^)
I made a tacit assumption that the sphere would have a greater
diameter than the coil. I wasn't clear about that sorry. However, I
subsequently realized that that shape doesn't actually matter.
Whatever terminal is used is going to be "calibrated" by measurement
with a known voltage. In fact the absolute capacitance doesn't matter
either. I realize streamers from a sphere could be a nuisance, hence
the bit about breakout points. Those would cause undesirable
distortions also. Point taken about the "ideal" flat plate except for
field concentration at the edges. Actually I build all my "toroids"
that way :) The probe might be prevented from corona formation by
enclosing it in a dielectric (hot melt for example).
So what are we looking at? A terminal suspended at the height
it would on the resonator fed with a voltage source running at Fr
with the probe sitting some distance above it for calibration. Then
an actual measurement. I guess the task now is to design a suitable
calibration source. A linear amp fed from an oscillator and stepped
up by HF transformer to the point where a meaningful scope reading
can be taken should do it. Termination of the topload by known
impedance might be part and parcel of the arrangement.
Regards,
malcolm
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.