1
\$\begingroup\$

I have a full-bridge inverter whose output is connected to an LC filter, like in the following image:

The power supply is at 20 V, the switching frequency is 100 kHz. The output is a sine wave with a frequency of 40 Hz, for an output power of about 10 W.

In my simulations, everything works fine. However, in reality I have much more losses than expected. For instance, with each half bridge at 50% (0 V differential output), the power dissipated by the filter is about 2 W; the result doesn't change with or without load.

For this case, I measured the current through the inductor and it's 1.3·Ipeak, the same we'd expect from the simulations. However, the 55 mΩ parasitic resistance of the inductor cannot account for the losses. I also verified the resistance at high frequency (100 kHz) and it's 200 mΩ, not enough to explain it either.

Do you know of any effect that I might be missing?

\$\endgroup\$
5
  • 1
    \$\begingroup\$ How do you drive the FETs? Do you have protection circuitry like a flyback for the coil? \$\endgroup\$
    – Ralph
    Commented Feb 1, 2022 at 11:41
  • \$\begingroup\$ What is the ESR of C2? \$\endgroup\$
    – user16324
    Commented Feb 1, 2022 at 11:53
  • \$\begingroup\$ Any cross conduction of your MOSFETs? What inductors or more specifically, core material are you using? \$\endgroup\$
    – winny
    Commented Feb 1, 2022 at 12:23
  • \$\begingroup\$ The ESR of the capacitors is 120 mOhms. There should not be any cross-conduction, as the gate driver is supposed to avoid it and I don't see it when probing the gates. \$\endgroup\$
    – fs.cor
    Commented Feb 1, 2022 at 16:59
  • \$\begingroup\$ Good. Are both the inductor and capacitors hot? What capacitors are you using and what tan fi (loss angle) do they have? \$\endgroup\$
    – winny
    Commented Feb 2, 2022 at 13:01

1 Answer 1

1
\$\begingroup\$

You are likely experiencing losses in your inductors. The most likely culprit is core loss, and you may also possibly have some skin effect-related resistance losses in the coil itself. When you switch an inductor at high frequency, eddy currents in the inductor core are induced. These currents, combined with the resistance of the core material, are losses that cause the core to heat up.

The skin effect losses are caused by the fact that as you go up in frequency, the current in a conductor tends to be concentrated near the outside surface and little conduction occurs in the center of the coil, so the wire is more resistive at high switching frequencies.

The switching frequency is not the only frequency when you consider the core loss effect. Your bridge switches at dv/dt orders of magnitude higher than a 100-kHz sine wave, and eddy currents occur from this instantaneous rate of change. However, you can reduce the number of times per second these losses occur by reducing the switching frequency. Try reducing your PWM frequency to a lower value and I bet the losses will be diminished.

Good luck!

\$\endgroup\$
2
  • \$\begingroup\$ Hi! Thanks for your answer! Actually, if I reduce the switching frequency the losses go up. Do you think of any way to confirm the heating is due to core loss? The inductor I use is the SRP6060FA-220M, which should properly tolerate the ripple current I see. Thanks \$\endgroup\$
    – fs.cor
    Commented Feb 1, 2022 at 16:49
  • \$\begingroup\$ Can you see if you are ringing? \$\endgroup\$ Commented Feb 1, 2022 at 21:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.