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

Re: More Coupling...



Original poster: "by way of Terry Fritz <twftesla-at-uswest-dot-net>" <paul-at-abelian.demon.co.uk>

Just a summary of the state of play.

The acmi program under consideration is at

 http://www.abelian.demon.co.uk/acmi/

and the current version is 0.1j, 24th December.

A summary of test comparisons for flat primary coils is
available from 

 http://www.abelian.demon.co.uk/tmp/acmi-compare.html

Bart also maintains a table of comparisons at
 
 http://www.classictesla-dot-com/temp/k.html

Advice on how to measure mutual inductance is available at

 http://www.pupman-dot-com/listarchives/1997/november/msg00898.html

acmi computes self and mutual inductances for concentric air cored
coupled coils by representing the windings as a large number of
circular current filaments and summing their contributions. This
approach allows arbitrary winding cross section profiles to be modeled
and also gives the mutual coupling in distributed form if required.

The approximation taking place here is likely to be weakest when large
conductors are used. Consequently we expect the worst results to occur
with primary self inductance and the best to be achieved with
secondary self inductance. Error on mutual inductance is expected to
lie between these extremes and the coupling coefficient k is at the
mercy of all three sources of error.

On the whole the program seems to over-estimate primary and mutual
inductances, and the error in mutual inductance is worse at closer
coupling, as expected in view of the primary conductor subtending a
larger angle as seen from a secondary conductor.

Attempts will be made to reduce the program error by seeking a better 
representation of the thick primary conductor.

If this program finds a use in the tesla community it will be largely
due to the efforts of Bart Anderson, Terry Fritz, and others, whose
continuing work to gather measurements and improve techniques involves
far more man-hours than is required to code and maintain the program.

--
Paul Nicholson,
Manchester, UK.
-