XM3 Inverter Phase Current Periodic Noise

rjae

Hi all,

We are developing a motor controller for a traction application so we bought the CRD300DA12E-XM3 to get acquainted with the XM3 series SiC power modules as well as with the embedded circuitry. For our design we use Matlab 2022B with simulink and Embedded Coder to generate code that is running on a RTOS.

At the moment we haven't even run the inverter yet, because we ran into trouble with the ADC phase current measurements. We have noise spikes over 10A and the strange thing is that is looks very periodic. The ADCs are running in single ended (12 bit) mode with a sampling frequency of 20kHz (triggered by ePWM module) and conversion window of 16 ADCCLKs. I did some debugging:

1. First thing I tried is lowering the ADCCLK and increasing acquisition period to see whether this would make a difference. Lowering the ADCCLK from 40MHz to 50MHz made no impact. Increasing the acquisition period from 8 to 16 ADCCLKs made no difference as well.
2. I checked the whether the module was grounded properly, there is no possibility of ground loops.
3. I measured the noise on the ADC input pins with our oscilloscope. The result was actually not what I expected, I can see noise, but it is just white Gaussian noise. Though when I run the software debugger and view the raw outputs of the ADC, the periodic noise is visible.
4. Then purely by accident I changed the CAN bus message rate of CAN bus A from 5000 to 50 messages/s. Suddenly almost all the periodic noise was gone. Trying out CAN bus B at 5000 and 5 messages/s also showed huge improvement, although there is still some periodic noise. For reference have a look at the attached figures where you can see all phase currents with their noise and the moment I turn off CAN A or B.

Hence my question: is there a known bug in the launchpad and/or Wolfspeed board that causes cross coupling between the CAN bus and ADC inputs? Is there a way to improve this?

Hope to hear from you! Thanks in advance!

Kind regards,

Remi

Forum_Moderator

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

JHayes

Hi rjae,

To my knowledge, there is no known bug with the launchpad or the Wolfspeed control board regarding coupling between the ADCs and CAN bus. Below you can see the trace mapping for both the launchpad and control board, with each set of traces highlighted.

Can you provide more information or pictures of your setup? How is the CAN connection being made back to the host PC? If you disconnect the LEM sensor cabling, is the noise still present?

Any other information you are willing to show would be helpful as we try to debug this issue.

Thanks,

Jonathan

TBhatia

Hi, I hope that this answered your question. I will close this discussion for now but if you have a follow up question, please "Start a New Discussion" and we would be glad to support you further.