I have a spreadsheet which does pretty much what Jim just described.
One thing that interested me when I started looking at the "best fit"
equivalent-circuit values was that the value of C varied depending on
the core material. In other words, you couldn't assume that 10 turns of
a particular coax on a "240-size" core, say, would always give the same
value of equivalent capacitance - it varies considerably depending on
the core material.
I've not yet yet found a mechanism that will predict the variation in C
to my satisfaction, but clearly a simple model based on inter-winding
capacitance doesn't get close.
If we had a good way to predict C, predicting the CM impedance of a
particular choke design would be relatively straightforward.
Steve G3TXQ
On 30/10/2014 23:48, Jim Brown wrote:
Values of parallel R, L, and C can be computed from the plots of
choking Z using conventional curve fitting. Rp is the measured Z at
resonance (the peak of the curve), L can computed from the slope of Z
curve at low frequencies, C can be computed as the value that
resonates at the peak frequency with L. I go one step further and put
those values in a spreadsheet, plot the impedance of that network on
the scale as the measured data, and tweak values for best fit.
This method is based on materials like #43, which have only one
resonance. #31, which has two resonances, requires more reliance on
curve-fitting. #31 with a lot or turns is like a double tuned IF. I
discuss this in the tutorial).
Thus, C is that which matches curve on the higher frequency side of
resonance, the value of L near resonance is that which resonates with
C, and the value of L at low frequencies is that which matches the
slope of the curve well below resonance. Values obtained are a fair
approximation of what might be obtained from more detailed curve fitting.
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|