Wolfspeed Power Applications Forum

Wolfspeed Technical Community:

SpeedFit™ Design Simulator

Simulate and evaluate the performance of SiC-based power circuits and determine the right SiC device in seconds.

Start a Simulation

Categories

PLECS Model CAB400M12XM3

luke13
luke13 Contributor Level 1
edited August 2023 in Module Products

Hello,

I'm currently running a simulation with a VSI powering a PMSM in PLECS. Therefore i'm using your thermal model for the CAB400M12XM3 modules. The temperatures and power dissipation of the switches look good. The only thing is that when i try to calculate the junction temperature of the sic-mosfets by hand there is a difference of nearly 50 K. The power dissipation (conduction and switching losses) for each mosfet in the simulation and the calculation by hand only differ by 10W, so I don't where the huge difference in the temperatures are coming from.

For the calculation by hand I do:

TH = 6*Pv*RthHF+TF then TC = 2*Pv*RthCH+TH and TJ = Pv*RthJC+TC

Pv is power dissipation for one mosfet and i do the calculation 3 times to adjust the power dissipation to the rising junction temperature.

Is it right that i need to enter 0.15 K/W for the thermal interface resistance in the mosfet model (see picture below)?

I also attached the plecs model.

Appreciate your help.

Best regards,

Lukas

2023_07_29_18_14_38_Inverter_PMSM_thermo_.png
Inverter_PMSM_thermo.plecs


Tagged:

Comments

  • Forum_Moderator
    Forum_Moderator Wolfspeed Employee - Contributor Level 4

    Thank you for your post, it has been approved and we will respond as soon as possible.

  • BlakeNelson
    BlakeNelson Wolfspeed Employee - Contributor Level 2

    Hey Lukas,

    You are on the right track, but my guess from a quick look at your circuit is that you are doubling up on the module Rth,jc by adding it to the thermal interface. The module model has a cauer network inside which accounts for the Rth,jc. See this diagram:

    image.png

    You set RthCh1-3 to 0.006, which is a much more reasonable value for Rtim than 0.15. RthHF of 0.008 is also a reasonable selection for high performance liquid cooling. Basically you just need to update the value in MOSFET interface and drop RthCH1, RthCH2, and RthCH3.

    Otherwise I don't see any issues with your calculation. The simulation runtime is long enough for the system to settle, which is another common mistake which leads to expected Tj to not match simulated.

    If you have other questions, please feel free to let me know!

    Blake

  • luke13
    luke13 Contributor Level 1

    Hey Blake,

    thanks for you answer, now I understand.

    Did the simulation again and the values are looking good now.

    Best regards,

    Lukas

This discussion has been closed.