From: Paul
Date: Thu, 21 Dec 2000 09:47:17 +0000
Subject: Re: [TSSP] Coils at high elevation
I wrote:
> > Has anyone heard from Mark lately?
Mark S. Rzeszotarski, Ph.D. wrote:
> I have been out of town and am still catching up on posts.
Hi Mark - glad you're still with us, haven't heard from you for ages!
> I may actually have some time for coiling measurements next week.
> Can you outline exactly what you want?
Well some time ago you sent me results for two coils measured at high
elevation and I've been unable to obtain satisfactory agreement on
the resonant frequencies with the model. The higher elevation results
from Malcolm's Q test data are also proving to be difficult to
reconcile. The predicted frequencies are some 10-20% high of the
values measured, and I've been unable to resolve this problem.
Some possibilities for the cause:
a) We're not taking sufficient account of capacitance to objects in
the environment around the coil - supports, feedline, distant walls,
etc.
b) We're ignoring relative permittivity of the secondary former and
insulated platform.
c) Physical capacitance is not being calculated correctly, although
two utterly different methods agree to within 3%.
d) Physical capacitance is not employed correctly in the model. Does
fine at normal elevations - some fault that shows up only when Cext
is small compared with Cint?
Apart from the Q factor, this is the only other major outstanding
defect in the model at present. I'm inclined to believe in option (a)
since if the coil is supported high up in the middle of the room,
we can reasonably expect the external capacitance to be sensitive to
the entire room and thus be hard to quantify.
In order to confirm this, I suggested in another post that it might be
possible to measure the f1, f3 and f5 on a coil suspended from say
a fishing line or similar, a couple of coil lengths or more above a
good ground, outdoors, and with a very thin wire feedline. Perhaps
this would eliminate all external capacitance except for the direct
capacitance of coil to ground plane. If such a measurement came out
OK, then it would be safe for us to attribute frequency errors from
indoor measurements to difficulty of gauging the external cap.
The pattern of errors for your small28a, 875 turn coil, are currently
meas tsim
kHz kHz
f1 885.0 1029.0 +16.3%
f3 2338.0 2473.2 +5.8%
f5 3436.0 3479.6 +1.3%
Mark, you described your coil as supported on an insulated platform.
Can you think of anything nearby that could add a substantial amount
to the coil's capacitance? Was the coil energised through a thin wire?
I did some rough checks and decided it would take quite a wide copper
strap coming out at right angles to the coil base to create enough
stray C to account for the error. Maybe somthing near the top of the
coil - a field probe perhaps?
> I can wind appropriate H/D if needed. I have most
> even wire sizes from 16 to 28 AWG and a few smaller than that.
> The constraints at my end are equipment based: my vector impedance
> meter works from 400 kHz to 10 MHz so I have to work with relatively
> small coils with that unit (Fres > 400 kHz).
Appreciated. I think measurements on smaller coils are more delicate
in view of the relatively greater effect of stray capacitance.
> It is useful for measuring odd harmonic Zin.
For now we're just interested in getting the right Fres.
> I can make a series of measurements around these values as
> well, but not at the even harmonics, where the meter will not
> stabilize due to voltage peaks at the drive point.
Yes, also the location of the even overtones is very dependent on
stray reactance presented to the base by the feed arrangement, for
that reason we don't use them.
> I can also take some measurements versus height above ground
> outside if you like (if the snow lets up).
Yes, although not exactly the best time of year for us in the north.
Malcolm mentioned he might be able to try such. One thing about
outside is that the physical cap becomes easy to model accurately.
I do all my tests outdoors, partly because my coils are too big for
the house, but mainly to get a well defined environment. Trouble is
it has rained almost nonstop for months now, so no measurements from
my end! We have a nice big tractor shed here on the farm - metal
walls and a high metal roof - would make a great HV lab but
unfortunately it's full of tractors!
> Alternatively, I can use an inductive feed with low coupling to
> energize the coil and observe it using a plane antenna a la Terry
> Fritz. In this case I can measure resonant frequencies accurately,
> but not Zin.
Yes, lets forget Zin for now. Maybe a very thin wire feed direct to
base would be less invasive than a coupling winding from the point of
view of capacitance? Tune for minimum Z magnitude or max current on
the feed rather than introduce a probe?
Ultimately I won't mind too much if we cannot get exact Fres at
high elevations, so long as we can prove that this is just inability
to quantify external capacitance well enough. The main worry is
option (d) which might indicate a fault in determination or employment
of Cint. Cint seems to produce some interesting phenomena in the coil,
features of the resonance which are a surprise to me, eg elevated
current maximum and in some cases a forward transfer impedance higher
than that of a 'lumped' inductor. For this reason eliminating
possibility (d) is most important I think.
Regards,
--
Paul Nicholson,
Manchester, UK.
--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.