Original poster: "Barton B. Anderson" <bartb@xxxxxxxxxxxxxxxx> Hi Greg,Sorry not to respond earlier. I've been lazy at reading emails the last couple of days. Derek is correct. Javatc and Fantc are the 2 online programs which integrate all capacitively coupled elements into the whole. Paul Nicholson's Geotc code is behind the scenes doing the number the crunching. When a topload is placed on top of the coil, the capacitance from topload to coil, coil to ground, topload to ground, topload to surroundings, coil to surroundings, coil to primary, topload to primary, topload to topload, etc.. etc.. all affect C. Based on size and geometry, the affects may be small or large. In any event, the program is working hard to spit out an accurate number (and does take time to complete because of the slow processing of non-compiled code, javascript).
There are times when both Wintesla and Javatc are close, and other times when there is a large discrepancy. The surroundings inputs can be confusing. It does assume a given C value in relation to the coil based on the distance and size of the surroundings and the coil. The surroundings are in reality different for most of us. There are usually objects around the coil scattered here and there that the program cannot account for. They affect C also. I've found that once the coil is actually built and measured within it's surroundings, there is usually a need of adjustment to the surrounding inputs to pull out an exact Freq match. But, once that is done, you then have good numbers for your surrounding inputs which helps ensure accuracy with further modifications to the same coil and with new coil builds. In other words, the surrounding inputs can be modified to account for all those other objects around the coil.
Regardless of any program, I am a believer in measurement during the design process and updating the inputs as real data comes in. Programs and measurement go hand in hand.
Take care, Bart Tesla list wrote:
Original poster: "Derek Woodroffe" <tesla@xxxxxxxxxxxxxxx>
Greg,
I can t explain all of the anomalies, but I suspect the biggest
difference between the two is the capacitance of the top load.
JavaTc gives a much more accurate figure for the top load that includes its
close surroundings. WinTesla assumes a free space topload. It is quite
possible that the loser frequency given by JavaTC is because of the wall to
topload calculations giving a higher top load capacitance.
There is no way to tell for sure unless you look at all of the values for
all of the calculated parameters e.g.
Top load capacitance, Secondary Inductance, primary inductance, primary
capacitance etc. given by both programs this should give you some pointers
to where the differences are.
JavaTc is the more accurate model. If in doubt and you are sure of all of
your parameters, then trust that one.
Cheers
Derek
Specifically the most far-off value was the frequency for the Secondary
circuit; WinTesla tells me it is around 362kHz, whereas JAVATC suggests a
more conservative 246kHz. That is quite a difference. I have double and
triple checked my inputs, yet these discrepancies persist. Curiously the
Primary circuit, while still off somewhat (200kHz for W.T. vs. 232 kHz for
J.T.C.) the values are much closer, and WinTesla suggests it is lower than
JAVATC, rather than higher, unlike for the Secondary.
Thanks,
Greg