Interaction of CHAN and ADC modules - problems with noise in measurements

Dear Sir or Madam,
For the needs of one of our projects in the previous period, we ordered an inverter Wolfspeed-CRD300DA12E-XM3. So far, the inverter has served us really well and has had very nice behavior.
We currently have a problem that manifests itself on the digital measurement channels (ADC inputs) when the CAN modules are working, so please advise us on what we can do.
In order to clarify the problem, I am providing several experimental results in the attachment.
All of the above tests were conducted without voltage on the DC link, and without power, the inverter power stage was not switching. The auxiliary electronics was powering the drivers and sensors. The sampling frequency is 400 kHz. Eight consecutive samples are averaged to reduce oscillations noise (i.e. oversampling is applied). When no CAN module is running, the measurements look like shown in Figure 1.
When there is communication via the CAN_A module, large spikes occur in current measurements, as shown in Figure 2. A more detailed depiction of the phenomenon can be seen in Figure 3. Offset is not a problem and is eliminated when there is voltage on the DC link. However, faster oscillations and high-power spurious spikes mean that we cannot use an inverter for our application.
When the CAN module B is turned on, very slow dips in the measured current occur, which are shown in Figure 4. These dips without oversampling go up to 15A. Also here, faster oscillations occur exactly at the communication (CAN) operating frequency of 5 kHz, as shown in Figure 5.
Is there any known problem with the inverter design or its interaction with the TI board and do you have any advice on how to overcome these problems?
I would like to mention that I have seen similar phenomena with another of your clients at the link :
Thank you very much for your time and help.
Best regards,
Aleksandar
Comments
-
Thank you for your post, it has been approved and we will respond as soon as possible.
0