Re: [femm] How to get the leakage inductance when very bad coupling and litz wires.

[David Meeker <dmeeker@xxxxxxxx>]

john fox wrote:

> Attached two examples at 25Khrz.
>
> The geometry is the same
> 7 turns with one amps each.
> 4 turns that I want to put on short circuit .
>
> In case1_1.FEM the wire is a 0.9*1 square wire with
> 1A or 0A depending on the shortage .
> In case 1_2.FEM the wire is a split in 2 wires 0.9*0.5
> carrying each 0.5A . In fact 1A for the litz wire.
>
> So the results from the two cases should be close when
> applying Dave's summing circuit method. In fact not !
> It seems to be a factor 4 coming for the value
> impedance values and a factor 2 for the voltage.There
> must be an adaption to make in the summing lua program
> in this case.
>
> This is for a 2 wires Litz. What factor should I be
> applying for 5 or 13 wires ?
>
> So , I need help and advise !

Be careful that sum the impedances correctly. If you have split one 
turn several wires, each of which carries part of the current in the 
turn, you have to add the all of the impedances of the wires that make 
up the turn in parallel, i.e.

Zturn = 1/(1/Zwire1 + 1/Zwire2 1/Zwire3 + ...)

Then, you'd sum up the impedances of each turn in series to get the 
total impedance:

Ztotal = Zturn1 + Zturn2 + Zturn3 + ...

Rather than modeling each component wire in the Litz wire (which can get 
old real fast), you might try the sort of continuum approximation 
described in the example at http://femm.berlios.de/proxapprox.zip

Dave.

--
David Meeker
dmeeker@xxxxxxxx
http://femm.berlios.de/dmeeker