DearJohann,
thanks for your support.
First, I'm not a specialist of modelica, and I put parameters in the package in order to use them in many variant of the same model in one package.
Is there a best solution ? Thanks for your advice ...
Second, the parameters are relative to a transformer where the coupling coefficient are close to 1.
By calculation, I grouped L1-L2, L3-L4, L5-L6 and the coupling coefficient are around 0.998 or more.
Then I simplify the model with the calculated value and obtained a similar results.
So you're right, the system seems not stable.

Any idea ?
Thanks for your reply
Fred.

Dear all,
please find here after a small diagram with 2 x 3 coupled inductors. The model does not work compare to the attended results and compare to the same model in LTSpice.
The goal of this model is to have a look on V_FI ...

And the relative netlist spice ...
L1 0 N002 1.57050669468473
L2 N002 N005 0.033714301106239
L3 NC_01 N003 0.05521312598501
L4 N003 N006 0.009481871379016
L5 N001 N004 245.391671044489
L6 N004 N007 0.937385890525789
R1 N005 0 0.061620942082922
R2 N006 0 0.00628903159971
R3 N007 0 14.5262976395218
V1 N001 0 SINE(0 1250 50)
K1 L1 L3 L5 1
K2 L2 L4 0.834491038835512
K3 L2 L6 0.617984954096065
K4 L4 L6 0.758762907589478
.tran 10
.backanno
.end

PACK1.mo

Thanks for your comment.
Regards
Fred

Thank you  very much.
It works with OMshell.

Dear sjoelund

Thank you for your answer.

Yes, I'm using MDT, only to have a check of the syntax.
But I'm using the "console OMC" inside Eclipse to generate the cpp and exe code by a call to simulate(model).

Is there a difference between "omshell" and the console OMC inside eclipse ?

Thank you for your support

F CHEVALIER

Dear All,

I'm trying to write a model with an external object and and an external function, very close to example "ExtObj.mo".
Then I receive a message from the compiler and I find this in the model_functions.cpp :

  mmc__unbox__mmcExpTypeShort:ERROR_rettype tmp3;

Could somebody help me ?

Regards