From: Paul
Date: Sat, 19 Aug 2000 18:22:23 +0100
Subject: [TSSP] News 19 Aug 2000
Hi All,
Measurements
------------
I'm very keen to collect measurements on tesla resonators for
comparison with the simulator. Trawling through coiler websites
produces little useful info, sometimes resonant frequencies are
given but the authors dont say if/how they were measured. The
only decent set of figures I've found are those of the Thor
system.
See
http://www.abelian.demon.co.uk/tmc.html
for details of the required measurements.
If you can help out please get in touch. Please let me know
if you come across any websites containing well documented
specs and measurements.
Inter-turn capacitance
----------------------
The v0.1 software included a term for the 'Palermo'
capacitance. In every case the resulting dispersion
came out too high. I've now removed this, or rather, I've
made it optional. I dont think anyone will be surprised
by this - the usual argument is that current migration to
the inner surface of the coil renders the simple parallel
conductor capacitance invalid, although I'm not exactly
convinced by this.
For example the Thor system:
Model error at: f1 f3 f5
With Palermo: +1.8% -1.1% -6.3%
No Palermo: +2.0% +0.1% -3.6%
and comparisons with my own coils show a similar
excessive dispersion when the Palermo inter-turn
capacitance is enabled.
If this remains the case, then it would be possible to
significantly reduce the number of LCR sections used
in the model, which up to now has been set to one section
per turn in order to examine this inter-turn capacitance.
Internal and external capacitance
---------------------------------
An essential next step is to separately assess the model
error in estimating the contributions of internal and external
capacitance to the total charge storage. In order to decouple
the two, I would very much like to take measurements on a small
coil placed at a high elevation above the ground plane, so that
the external capacitance is minimised.
Complementary to this, I'd like some measurements of a small
coil surrounded by a fairly close fitting conducting grounded
cylinder, so that internal capacitance is minimised and external
capacitance maximised.
If anyone is willing to do some experiments along these
lines, please get in touch - the results will be very valuable.
(Surely these kind of measurements have been done before -
websites?)
Project name?
-------------
Someone asked me what this project is called. Any suggestions
for a name?
Website
-------
Maybe its time the project had a website - still interested
Calvin? Can anyone suggest a suitable peer review process for
any posted results?
Summary
-------
The computer model is beginning to show some quantitative
resemblence to real coils with respect to the resonant
frequencies.
Input impedance and Q factor still have a long way to go.
Model errors are currently:
System f1 f3 f5 f1:f3 f1:f5
abelian +6.2% +2.5% -5.6% 1:2.3 1:3.1
no toroid
abelian
with toroid +4.2% -0.9% -7.3% 1:2.8 1:4.1
thor +2.0% +0.1% -3.6%
The model consistently exhibits too much dispersion, and the
predicted f1 is always on the high side. The excess dispersion
is worse when external capacitance is reduced by removing the
toroid, suggesting that internal capacitance is over-represented
in the model.
It's interesting to see what happens if we switch off internal
capacitance between regions of the coil, leaving the rest of
the capacitance matrix intact. Then we get, for two of
the systems,
System f1 f3 f5
abelian
no toroid +12.2% +39% +55% 1:2.9 1:4.9
thor +3.7% +13% +33% 1:3.7 1:6.8
In the case of abelian without the toroid, the overtones
are almost harmonic, suggesting that the mutual inductance,
is not providing a great source of dispersion. These figures
suggest that the implementation of internal capacitance is at
least of the right order, but is maybe 10% too high. To help
fix this, those measurements on an elevated coil are required.
And finaly...
-------------
The measured inductances of both thor and abelian both fall
significantly short of the Nagaoka inductance, which is usualy
good to 1% or better. If the Nagaoka inductance is used
instead of the measured, and an estimate of the excess dispersion
is removed, the resulting errors just happen to come out rather
good.
Regards All,
--
Paul Nicholson,
paul@abelian.demon.co.uk
--
To be removed from the TSSP list, email to
majordomo@abelian.demon.co.uk and put
unsubscribe tssp
in the body of the message.
--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.