RE: Theory self-C

Hello Malcolm (and any other interested parties)

You recently commented

TL>Hi Pike and all,
TL>                You asked....
TL><snip, points noted>
TL>> You've mentioned Medhurst's formula several times and
TL>> I've looked for it but not found it, at least not by
TL>> that name.

TL>I originally found it in the the AWA Radiotron Designer's
TL>Handbook ed. F. Langford-Smith. 

TL>In simple form :
TL>                 Cself = H x D    where D is the coil
TL>diameter in cm and H is a value depending on coil h/d. H
TL>ranges from 0.51 for aspect 2, to 0.81 for aspect 5 in a
TL>more-or less linear fashion.  :

This expression is also discussed in the "Radio Engineers' 
Handbook" by Terman: it is interesting to note the comments 
( 1st ed., 6th impression 1943 p 85 and p 923) in the 
footnotes, namely:

" The effective distributed capacity will also depend to 
some extent upon the current distribution in the coil, and 
will in general be larger when the coil is shunted with a 
large external tuning tuning capacity than when the coil is 
resonated with its self-capacity."

[I have seen that typically the top-loading capacity hat 
(toroid) has a value of Cself/2 and 2*Cself so I would argue 
that the actual C self of the coil appears to be larger when 
it is terminated in a toroid].

"Distributed capacity is sometimes calculated from the 
frequency at which the coil is in parallel resonance with 
the distributed capacity. This does not give the distributed 
capacity effective under practical conditions, however, 
since it corresponds to a different current distribution 
within the coil".

This second quote has particular relevance because most TC 
experimenters check for self-resonance by base driving and 
looking for minimum Zin, or by using an E-field probe and 
looking for maximum Vout. The preferred method would be to 
hang external lumped C across the coil, and find the new 
resonant f. Then you plot a curve (straight line) of 1/f^2 
against the external tuning C. You get a line that cuts the 
y axis at some value and you can then extrapolate this line 
to some value of -x which represents the actual self-C.

What I argue is that the external influencing factors vastly 
swamp the measured values of self-C and whether you use C=Hd 
or one of the other empirical formulae (all of which are 
based on the C of an isolated cyclinder), the fact that no. 
of turns, turns spacing etc are ignored doesn't matter.

Richard Craven, England.