From: Paul
Date: Wed, 09 Oct 2002 10:14:27 +0100
Subject: Re: [TSSP] Curiouser and curiouser!
All is revealed. Great set of results.
> http://hot-streamer.com/temp/OLTC10-08-01.CSV
The above from the pinger is fine. Very clean ping of
the secondary. Appears unaffected by the presence of
the primary.
> http://hot-streamer.com/temp/OLTC10-08-02.CSV
This is the running coil, in which tcma is having
trouble analysing the f3 mode (3/4 wave overtone).
If I discard the first 0.1mS after the ping and
analyse the remainder of the ringdown, we get
PK FREQ kHz (Error +/-) Q FACTOR (Error +/-) LEVEL
1 37.148 (0.01%,2Hz) 84.00 (0.29%, 0.2) -0.0dB
2 113.893 (0.01%,7Hz) 43.15 (7.90%, 3.4) -35.6dB
3 74.368 (0.01%,4Hz) 100.82 (26.56%,26.8) -35.5dB
The second harmonic of the secondary f1 is visible at 74.368kHz,
it has a very narrow line width on the FT compared with f1,
which means that it's amplitude is almost constant, while f1
is exponentially decaying. f3 is pretty much the same as in
the secondary ping, which gave
PK FREQ kHz (Error +/-) Q FACTOR (Error +/-) LEVEL
2 113.864 (0.01%,8Hz) 46.26 (0.70%, 0.3) -12.3dB
The 'trouble' occurs in the first 0.1mS while the primary
'gap' is closed. Analysing this short section gives
PK FREQ kHz (Error +/-) Q FACTOR (Error +/-) LEVEL
3 117.956 (0.09%,102Hz) 14.95 (11.17%, 1.7) -18.5dB
4 174.305 (0.09%,151Hz) 31.46 (15.79%, 5.0) -21.9dB
5 230.793 (0.09%,200Hz) 42.60 (11.17%, 4.8) -23.6dB
6 278.122 (0.09%,241Hz) 41.49 (18.78%, 7.8) -26.9dB
for the higher modes. Notice that f3 is slightly higher in
frequency (118 kHz) compared with the secondary ping (114 kHz).
This occurs because as far as the higher modes are concerned,
the f1-resonant primary looks a bit like a shorted winding.
The reaction from this slightly reduces the effective
inductance of the sec, raising the mode frequencies a little.
We see the following frequency shifts of the higher modes
when the primary resonator is switched in.
Sec only Sec+pri Change
f3 113.864 kHz 117.956 kHz +3.6%
f5 172.448 kHz 174.305 kHz +1.1%
f7 223.862 kHz 230.793 kHz +3.1%
f9 273.390 kHz 278.122 kHz +1.7%
The breakdown of tcma is caused by the pattern of modes
changing as the gap opens. The 'running' signal contains
a mixture of two mode sets. It shows up as 'sidebands' of
each mode peak in the spectrum. The mode peaks themselves
are broadened. Only with f3 are the 'sidebands' above
the threshold that tcma tries to analyse them. Hence f3
appears broken and the higher modes appear cleaner, although
if anything they are affected almost as much.
tssp is able to model this effect. I've seen this many
times in the model outputs, but until tcma came along, there
was no prospect of ever detecting and quantifying the subtle
changes of Fres.
Predicted shifts for this coil are
Sec only Sec+pri Change
f3 113.355 kHz 117.844 kHz +4.0%
f5 173.724 kHz 180.341 kHz +3.8%
f7 226.585 kHz 234.117 kHz +3.3%
f9 285.218 kHz 292.168 kHz +2.4%
A most interesting result. Very pleasing to see this subtle
point so nicely (and accidentally!) confirmed.
(I put 'sidebands' in quotes because of course they're not
really sidebands, just a mix of two signals that aren't present
at the same time. They look a bit like (one-sided) sidebands on
the FT).
Cheers All,
--
Paul Nicholson,
--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.