How to protect STM32G491REI6 from EMI Generation in Inverter Application??

[Champak Kundu]

I am using STM32G491REI6 in the application of sinewave Inverter. The input of the Inverter is Battery. We have given provision of variation of input battery voltage from 96V to 192V. When the battery voltage is under 120V. Everything is going nicely. But when we connect 120V battery, the PWM signals and all the output signals got stopped. We are thinking that due to EMI generation in Inverter everything is going to happen. We are seeking suggestions from ST Microelectronics

[Simon V.]

Hello,

You may face EMI issue relative to an external oscillator present in your application. If the external oscillators operation is stopped, then the MCU clock tree may be also stop. Are you able to connect a debug probe when battery voltage is above 120V ? Can you check the reset status flag for example or the program counter location (hardfault for ex) ?

Regards,

Simon

[Champak Kundu]

Thanks Simon for your response.

Can you suggest the debug probe of any manufacturer by which we can easily debug as said by you?

and my next question is how to check reset status flag?

My another question is, if this problem has raised then how to overcome this?

[Simon V.]

Hi,

You can use for sure STLINK-V3 debug probe and STM32CubeProgrammer to simply access to the STM32G4 registers (Control/status register (RCC_CSR) cf product reference manual RM0440).

But before, may you share more information relative to the issue you observed like Andrew mentioned?

Regards,

Simon

[Andrew Neil]

"We are thinking that due to EMI generation in Inverter everything is going to happen"

Why do you think that would be any different depending on whether the battery voltage is above or below 120V ?

What tests/measurements/investigations have you done to verify/disprove this hypothesis?

Have you observed best EMC practices in your design? Have you measured the emissions?

[GLASS]

Hi.

Do you have galvanic isolation to drive bridges of stage 1 and stage 2?

May be it can avoid some parasitic return current​...

Same for debug probe​?