0
\$\begingroup\$

I am trying out an H-Bridge inverter using GaN FETs at 140 kHz. The schematic is as follows. (In the experimental prototype, I used TI's LMG5200 and the load is connected via a full-bridge rectifier.

enter image description here The switching waveforms are enter image description here

The circuit works fine when it is not loaded. The output of the inverter: enter image description here

But when I load the inverter with the resonance LC circuit and the receiver, the output waveform becomes like this. The output voltage is not stable.

enter image description here

even in the LTSpice simulation, I can observe a similar waveform: enter image description here

But if I simulate at a higher frequency (by changing the value of the resonance capacitor), the output looks stable. See the below waveform at 640 KHz enter image description here

What am I missing here?

\$\endgroup\$
  • \$\begingroup\$ Why do you have output cap? Also why leg A and B are not 180 phased? \$\endgroup\$ – jalaffo Jul 6 '19 at 8:36
1
\$\begingroup\$

You are asking for trouble; the drive frequency is 140 kHz and your primary series resonant tuning is precisely 140 kHz therefore, the L and C act as a short circuit at the switching frequency. You need to run the primary circuit either from an output stage designed to handle the series resonance or use a different approach - maybe add a current limit resistor in series with C4(C2).

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ thanks, In fact, it will not be fully short-circuited. There will be a reflected impedance \$(\omega M)^2/R_L\$ in series with the ESR of the inductor. But at 140 kHz, it is extremely small. I understand that at a higher frequency, the reflected impedance will be higher and it makes the current lower. I may have to use larger L1 and L2 (so as the M) so that I can limit the current. The use of a series resistor is not an option due to high losses. What do you think? \$\endgroup\$ – Pojj Jul 3 '19 at 15:01
  • 1
    \$\begingroup\$ I have done a similar job. I filtered the square wave to a sine and parallel tuned the primary. I don't know if this is possible for you. I was operating at 600 kHz FYI. \$\endgroup\$ – Andy aka Jul 3 '19 at 15:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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