Re: Generating k

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Barton B. Anderson by way of Terry Fritz <twftesla-at-qwest-dot-net>" <tesla123-at-pacbell-dot-net>

Hi Malcolm -

Tesla list wrote:

> Original poster: "Malcolm Watts by way of Terry Fritz
<twftesla-at-qwest-dot-net>" <m.j.watts-at-massey.ac.nz>
>
> Hello All,
>            This is a general enquiry, directed in particular to the
> author of MandK.
>
>      Since a program exists to calculate k given a few physical
> variables, would it not be possible to write a routine to give
> relative coil positioning given a required k and the physical coil
> attributes?

Absolutely. If I get the gist of what your asking, basically there would be
an input into a program for
a predetermined coupling value (say from personal or list experience) and
the secondary to primary width
and/or height would be an output.

> Anybody out there already done it?

Dont' think so, but, I've been working this last month on a hard equation
(by hard, I mean not
attempting a filament structure such as with Acmi) but I want it to be very
near Acmi or MandK. I'm
doing this for JavaTC. I'm getting excellent results with flat and helical
coils (small to large) and
currently working on conical primarys. I'm within a couple microhenry's if
I compare with Acmi in the
flat or helical cases. The angles are in process. I need a hard equation
due to Javascript limitations
in the programming arena.

On a side note, John Corture once mentioned the conical equations we use
are not quite right. I've found
this to be very true. Inductance is furthest off when a 45 degree angle is
used, and gets better as the
angle is nearer to flat or helical. I've found a way to correct this (very
important after determining
mutual inductance and then determining K from the individual inductances -
all part of the process I
guess).

> While operational
> variables might dictate the k one has to live with (e.g. racing
> sparks), it would be a useful program for anyone wanting to set their
> coil up to a particular k for whatever reason. I am not familiar with
> the algorithmic particulars but could be pushed in that direction
> *sometime*, not necessarily soon if no-one takes up the challenge.
>
> Malcolm

Very useful and easy to implement. I say, go for it if you get the urge. I
may try this approach in
JavaTC after all is said and done. Good idea! (and your timing is perfect).

Take care,
Bart