S32K344 BMS example with MC33774

[sandesh]

Hello,

I'm trying to get the example project from MBDT BMS, specifically s32K344_665_774_s32ct.mdl, working on hardware.  I have the following hardware setup:

• S32K3X4EVB-T172 (default jumper configuration) with power connector (12V through a benchtop power supply) and micro-USB (to a Windows machine with MATLAB & MBDT)
• FRDMDUALK3664EVB wired to the S32K344EVB-T172 based on the recommended schematic in the project (see attachment FRDMDUALK3664_PinMapping). It is also connected to the RD33774PC3EVB via two twisted pair TPL cables.
• RD33774PC3EVB connected to the FRDMDUALK3664EVB as described before and connected to the BATT-18EMULATOR using the provided 3x 32P cable
• BATT-18EMULATOR connected to RD33774PC3EVB as described before and the provided 5V power supply.

I can build and flash the model (after changing the NumNodes input in initialization to 3) without problems, and I do see the LED blink. Furthermore, I added a counter in the model to confirm code execution, which I do see incrementing in FreeMaster. However, I see no data exchange or measurements made by the MC33774. The SendStatus is always 5 and the TDStatus  reads PHY_TS_IDLE. 

Is this a working model or are there changes that need to be made to get the model working? 

[mariuslucianand]

Hello @sandesh 

First of all, I’m very sorry for the experience you had when getting started with MBDT on your setup. We understand how frustrating this must have been, especially given the expectation that the provided example should work out of the box. I’d like to clarify what happened and provide some context around the issues you encountered.

When this example was developed (around April 2024), we did not yet have access to the 

FRDMDUALK3664EVB board to validate the setup directly on that hardware. As a result, the development and internal validation were performed using the previous board revision,
FRDMDUAL33664EVB. Unfortunately, this also led to a naming inconsistency in the documentation. Although the difference appears to be only a single letter, these boards target different platforms: the FRDMDUAL33664EVB was designed for use with the S32K144 EVB and that board is not pin‑compatible with the S32K344. This discrepancy resulted in the FRDMDUALK33664EVB being referenced in parts of the documentation, even though the wiring and testing were described separately and validated using the earlier board revision. 

I can confirm that the example models are/were tested prior to the releases. You now have access to a newer, pin‑compatible board revision, which no longer requires the additional wiring that was necessary at the time the example was developed.

The K board, which comes with a new pin mapping requires revisiting certain aspects of the setup, in this case adjusting the PINS configuration and also changing the SPI configuration  instance in the external configuration tools, as you correctly identified in your solution.

Unfortunately, the MBDT development team does not always have immediate access to the latest hardware revisions, and in this particular case the hardware evolved while the MBDT example were not updated due to no releases in between. To prevent similar issues going forward, I will open an internal ticket to review and update this example. At the moment, however, there is no new BMS release scheduled.

Now that you have successfully adjusted the model to work on your side and the MBDT blocks have been confirmed to work please let us know if you need any additional clarification or support from our side.

Regards,

Marius

[sandesh]

Thank you for the context, Marius. I appreciate it.

As I understand, FRDMDUALK3664EVB was released sometime in 2022. I understand the validation was done using the previous version of the board, the FRDMDUAL33664EVB. However, the readme document in the MBDT example folder clearly mentions the 'K' version. I recommend revising the document at the very least in the next BMS release if reconfiguring is not within scope.

As far as questions go,

1. Reading through the documentation for the AFE, there are different balancing and measurement modes available.  I understand that balancing is paused automatically in certain measurement modes. Are there examples that go over managing cell balancing and measurement on the MC33774A?

2. For the measurement with multiple AFEs, I know the examples show measuring one AFE at a time by changing the device address every time step. As I understand, the examples use the periodic measurement registers, which update every 4ms or so, and the actual communication time is ~0.5ms/AFE. I was curious if we could stack all the queries and responses for multiple AFEs within a for or while loop to receive data faster than the 100ms in the example? 

3. For the ring topology to work with multiple AFEs, do I need to add a new chain address? I was thinking I could just use the TPL12TXCS instead and capture data from TPL1RX and TPL2RX. However, I wasn't sure if the device enumeration had to be unique and if it needed a different initialization. Do you have any other pointers?

Thanks,

Sandesh

[sandesh · Accepted solution]

I managed to solve this problem. There are multiple things wrong with the example (s32k344_664_774_s32ct.mdl) - from the documentation to the pin configuration in S32CT. I'm not sure if anyone from the MBDT team did their due diligence in checking if this was a functional model before releasing it almost two years ago. In short, this is a bad example to use.

Additionally, I'm not sure why the example says the boards are not stackable or pin compatible, as the documentation for the FRDMDUALK3664 clearly states they are meant to be used together. Perhaps someone from NXP can shed some light. 

[sandesh]

The biggest problem I've seen so far is that the TPL1 Enable pin stays low (I presume this is an active high pin on the FRDMDUALK3664EVB?). There might be others...

@mariuslucianand, any thoughts?