From: Paul
Date: Sat, 17 Feb 2001 14:45:40 +0000
Subject: Re: [TSSP] Top voltage testing
We have, for Terry's small coil:
measured modeled error
Q: 72.5 102 +41%
Rin: 416 ohms 293 ohms -30%
Vgain: 83.46 119 +43%
Interestingly, 102/72.5 ~ 416/293 ~ 119/83.46 ~ sqrt(2), which is
food for thought. Once again the Qi is a little different to the Qv,
so I've used the Qv for comparison.
Terry wrote:
> This coil uses white PVC for the form that is 1/16 inch thick.
>
> The diameter is 4.250.
> The length is 26.125
> The turns are 1175.5
> Rdc = 33.40 ohm
Revised comparisons are
measured modeled error
Rdc: 33.4 ohms 34.09 ohms +2.1%
Ldc: 22.1 mH 22.39 mH +1.3%
Zft: 34450 ohms 34874 ohms +1.2%
Les: 17.64 mH 17.85 mH +1.2%
So not much change there. The overall similar positive errors is
interesting. Perhaps if we can solve the discrepancy in inductance,
the transfer impedance will follow.
> I tried hooking the generator directly to the base of the coil and
> got over 1kV peak to peak on the top but the Q was still around 70.
which shows that the generator source impedance is not the main source
of series resistance. That's good.
> I tried using the amplifier but that only raised the Q a small
> amount.
Yep.
> However, I raised the coil about 2.5 inches off the ground plane and
> the Q did go up significantly. I will have to clean up the setup to
> know how much but perhaps eddy currents in the foil are causing to
> low Q.
Did the inductance go up too? The Q will go up anyway because the
frequency will go up as the external C is reduced, so to gauge the
extra effect of the eddy loss, divide Q by f1 and see if the ratio
improves as you raise the coil.
This is very reminiscent of the behaviour of your large coil when we
first started testing. Confounded by the sonotube loss, we were unable
to come to a conclusion about how much loss was occuring in the foil
eddy currents. In view of the quite large shortfall in measured Ldc
compared with the Nagaoka value, is sure looks like there is sub-
stantial eddy coupling. I've not been able to calculate the expected
loss yet - I have some relevant papers[*], but I've not got around to
doing any arithmetic.
> If I we to make a radial counterpoise out of #24 wire, do you have
> any idea how many radials I would need and their length for an
> effective plane?
I have absolutely no idea how well a wire radial artificial ground
actually works. Such things are commonly used in antennas, and things
like chicken wire for reflectors on quite high frequency antennae, but
in none of these applications is the 'wire ground' inserted into such
a high Q circuit as we find in the TC. So I daren't trust to my
intuition, which is all gained at much higher frequencies and far
lower Q factors. I plan to do some simple tests to answer the main
question: To what extent does a wire mesh impersonate a continuous
metal sheet in its ability to intercept the arriving E-field. My
baseline standard is tinfoil - cheap, guaranteed to catch all the
E flux, and because it's thinner than the skin depth at any reasonable
TC operating frequency, the Rac is equal to the easily measured Rdc.
One of my experiments will be to measure C between a square of foil
and the TC, and compare this with a same-sized piece of mesh at the
same distance. There are other experiments along the same lines, eg
using a plane detector shielded behind a) tinfoil b) mesh, etc.
If anyone beats me to it with experiments like this, I'll be more than
glad. I've settled for foil as the bulk of my CW groundplane, using
sheets 0.75m x 2m, twelve of them, sandwiched between heavy poly-
ethylene for robustness. Laid out in a pattern, each insulated from
neighbours and separately connected to a central collecting point,
that gives me a continuous ground out to around 2m radius. I've yet to
test this, so I don't know if I need to go further out. The foil does
not run underneath the coil, so I've about 1 sq m which I need a
different plan for. Maybe some big sheets of the copper-strip vero
(matrix) board, with all the strips joined at one edge. The aim of it
all is to intercept as much as possible of the coil's external flux
and return it to a central collecting point, a ground hub if you like,
with minimal loss. Any flux missed by the groundplane has to make it's
own way back to the signal generator or coupling box, and therefore
introduces an uncontrolled loss. The best (lowest) effective ground
resistance that I've been able to achieve over a grass lawn was around
50 ohms, with the lawn well soaked, riddled with many stakes, and
covered in a fairly coarse mesh. That's still high enough to halve my
Q factor, and therefore halve my output voltage. Hence the desire to
experiment with foil sheets - the potential benefit from 100$ worth of
foil and plastic, etc, is several times more cost effective than
trying for the same gain from a bigger power source.
> That would eliminate a lot of ground plane effects. I could also
> raise it off the floor to hide test stuff under it. I could probably
> only build one big enough for the 26 inch coil.
The closer to ground the coil is, the greater the portion of it's E-
field intercepted by the ground plane. So lower is better, but it
means putting together a ground structure which will not couple
significantly to the B field. I think that a dense radial pattern of
#24 might be overkill and there ought to be a way to construct an
effective ground much easier than this. A double-sandwich of foil
and plastic might do, for example, in cross section,
------------------------------------------------- top plastic
************ ************* foil radials
------------------------------------------------- middle plastic
===== ============== ========= foil radials
------------------------------------------------- bottom plastic
and in plan view, something like
[*****************]
[=================]
[*****************]
[=================]
.etc
.etc
[*****************]
[=================]
with all the foil strips joined along one edge. I use the heavy duty
plastic sheet that builders use to waterproof floors. Maybe the bit
directly beneath the coil should be something like matrix board, as
the foil is probably too wide for use that close in. Unfortunately
this arrangement does not prevent the coil B field from stirring up
currents in the earth beneath the artifical ground, so putting it all
above a wet lawn, say, there will still be some unaccounted losses.
[*]
Dodd and Deeds,
"Analytical solutions to eddy-current probe-coil problems",
J. Appl. Phys., vol. 39, pp. 2829-2838.
Theodoulidis T.P., Tsiboukis T.D., and Kriezis, E.E.
"Analytical Solutions in Eddy Current Testing of Layered Metals with
Continuous Conductivity Profiles"
IEEE Trans. on Magnetics, Vol. 31, No.3, pp.2254-2260, 1995.
--
Paul Nicholson,
Manchester, UK.
--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.