How to get a valid inductance matrix from maxwell when there are parallel winding

Hello, I'm trying to design a 200W transformer. I have reached a valid design in PEmag and I got the 3D FEA from it and the results seem to make sense so I transferred the design to the maxwell to double-check the inductance values (I want a specific value for the leakage inductance) and calculate the core loss. In my design, I have two windings for the secondary side to limit the current density. In the maxwell first I tried to put three windings and join the secondary and tertiary winding in the parallel in post-processing but the values don't make any sense at all. So I tried to put all the secondary terminals in one winding and set the number of parallel branches 2 but it didn't help as well. ( even the effective turn ratio doesn't make any sense) for example, in PEmag the L11, L12, and L22 values are 157uH, 16.68uH, and 12.47uH respectively but in the maxwell when I enter the number of parallel branches 2 for the secondary winding my inductance matrix would be 368.84uH for L11, 0.228uH for L12 and 0.540uH for L22. I used the eddy current link to the maxwell feature of PEmag so I didn't change anything in the geometry I just changed the excitation and assign coil terminal excitation and then transfer them to two windings. Does anyone know how should I set up my windings to get the desired values from maxwell which are similar to PEmag? Thanks. 


  • icellb1icellb1 Forum Coordinator


    Could you attach a screen capture of your windings configurations in Maxwell, and your post-processing window showing how you assigned parallel with the two methods you mentioned: (1) put three windings and join the secondary and tertiary winding in the parallel in post-processing (2) put all the secondary terminals in one winding and set the number of parallel branches 2?

  • Hello,

    Thanks for your response. I've attached the configuration of windings in PEmag, How do I set up the windings in maxwell and the results. In 2 winding configuration, all 26 turns of the secondary winding are in one winding but the number of parallel branches is set to be 2. But in 3 winding configuration, the number of parallel branches is 1 each of the secondary and tertiary windings have 13 turns but in the post-processing, I have joined them in parallel.

  • icellb1icellb1 Forum Coordinator

    Hi, MohammadShc ,

    Could you try attaching screen captures by "upload image" option rather than "upload file" option? Your images don't show directly, and we are not allowed to click into images that do not show. Thank you.

  • Hello again,

    Sure here are the pictures in the same order.

    2 windings with two parallel branches:

    3 windings:

    Winding configuration in PEmag:

    Also, today I found out that the direction of the coil terminals for secondary and tertiary winding were different so I changed one of them so that both have the same direction now the answer is better but it's still different with the PEmag here is the picture of new inductance matrix.

  • icellb1icellb1 Forum Coordinator

    Hi, MohammadShc,

    Both the two ways you used to set excitations for the parallel secondary windings in maxwell are correct. You should definitely set the directions of the excitations the same for the two parallel secondary windings, otherwise the inductances will cancel.

    From your last picture (the one you got today), the inductance values obtained by maxwell make sense because L11, L12, and L22 obey the turn ratios (e.g. L11/L12≈N1^2/N2^2, L12/L22≈N1/N2) while the inductance values you got from PEMag do not obey turn ratios (L12=16.68 μH does not make sense).

    Since the inductances obtained from PEMag and maxwell are very different, I would suggest you to roughly calculate the inductances to check which simulated inductances are correct, by simply using the equation L11=N1^2*μ0*μr*cross-sectional area/length.

  • Hi @icellb1 ,

    Thanks for your response so I will double check it with the mathematical equations.

Sign In or Register to comment.