[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Small coil frequency sweeps



Original poster: "Dr. Duncan Cadd by way of Terry Fritz <twftesla-at-uswest-dot-net>" <dunckx-at-freeuk-dot-com>

Hi Richie, Terry, All!

My apologies if this repeats - I sent it 24 hours ago and I think the
server ate it ;-)

>Original poster: "Terry Fritz" <twftesla-at-uswest-dot-net>
>
>>Very interesting plots Terry,  I have meant to do something similar
>>myself.
>>
>>I am surprised by the rising current at higher drive frequencies.
Maybe
>>this is due to the close proximity between the bottom of the
resonator and
>>your ground plane ?
>
>Gee, I don't know :-))  I don't remember anyone ever measuring it
before.
>Maybe Paul's project would have the answer to that.
>


I have just posted a couple of gifs ("I fear the Geeks when bearing
gifs." Who said that? But no Trojans in Terry's virus-scanned temp
directory ;-) to hot-streamer-dot-com.

http://hot-streamer-dot-com/temp/basefeedi.gif
shows the variation of simulated basefeed current with frequency on my
NEC2 model of Skip's 240 turn coil.  Several things to note.

1) The frequency resolution is coarse, 250kc steps.  (Each data point
represents 40 minutes of computer time, so with 45 points that's
around 30 hours for the whole graph!)  This means that the true peak
frequencies will not correspond to the calculated peak frequencies,
and the true peak values of current will be higher.  Also, the
harmonics which appear are not on their correct centre frequencies.
Think of it as a "quantization" error.

2) The frequency sweep is from 500kc/s to 11,5Mc/s.  StarOffice 5,1 is
too much of a pig to stick the frequencies on the x axis :-(

3) The model is half a metre above perfectly conducting ground.

4) The current which is plotted is the total magnitude, combining the
real and imaginary components vectorially.

5) For a variety of reasons, the NEC2 model does not exhibit
resonances at the same frequencies as Skip's real coil !  The
frequencies for resonance according to NEC2 are in every case
significantly higher than the real thing.

6) In this case, it is a much higher harmonic current than second
harmonic which *seems* the largest.  This can easily be an artifact of
the frequency stepping.  If the frequency step happens to land near
the maximum of an harmonic (which it could be) but is off the peak of
the fundamental (which it is) expect strange things to appear.

7) However, the trend is clear, the basefeed current does indeed
increase to higher frequency.  Classically, I recall that above Fres,
a coil exhibits _capacitive_ reactance and of course as the frequency
goes higher, so will the current.

The second picture
http://hot-streamer-dot-com/temp/efield.gif

shows the variation in electric field simulated at the surface of the
toroid vs. frequency.

The same frequency steps and comments as per (1)(2)(3)(5) above apply.

8) The field reported is the vectorial sum of the x,y and z
components.

In both cases, an excitation of 10V was applied.

I can re-run the simulation on the harmonic spot frequencies, but
that's another N hours computing . . . if you _really_ think I should,
say so :-)  It is also possible to look at the current in the coil.
However, the output data files for coil current are _massive_ and
there is no way I can post the raw data to hot-streamer-dot-com, even
generating a graph will be mildly amusing and will require some
fortran hacking (too much data to try it manually via a spreadsheet -
6000+ data points).  Bearing in mind 40 minutes per frequency, I am
open to further suggestions on this (gulp! What have I let myself in
for?!)


Dunckx