TMS320F280039C: When I divide the clock, I find that the HWBIST test is not working properly.

With the increased time needed with a slower clock, is there any chance there is an interrupt that occurring during the run?  Especially something like the Watchdog ISR?

Best,

Matthew

The address 0x5E040 returns 0x29, which means it has experienced a time-out failure. How should I correctly perform 2x and 4x clock division in this situation?

I will need to check with some others to see if there is a known limitation here.

Best,
Matthew

Sure, thank you! If there is any news, please let me know promptly!

Apologies, I was OOO last week for US Thanksgiving holiday.  Let me follow up again now that I am back and see if we know of any limitation here.

Best,

Matthew

Ok, please let me know if you have any information. Thank you!!

After some internal discussion, there is a known issue with running HWBIST with reduced clock rate.  Currently we are evaluating if we can support reduced clock rate for power savings or this needs to be removed as option.

For background, do you have power consumption concerns at full clock rate that don't allow to run HWBIST at full clock, or can you revert the code back to the original clock settings for HWBIST to run?

Best,

Matthew

Currently we do not have conclusion if HWBIST can run at reduced clock rate on F280039 or any other C2000 device.

For #3, I think you are asking based on some of my earlier comments.  I was not yet aware of this issue in general of reduced clock + HWBIST when I made that statement.  This is not the root cause of this issue, so that is no longer valid problem.

I was speculating that if the watchdog was active, and HWBIST clock was slower, then perhaps accommodations had not been made to increase the WD timeout window.  But again, this is not the reason HWBIST does not work with reduced clock rate.

Best,

Matthew

Ok, if you have any information, please let me know in time, thank you!

Part Number: TMS320F280039C

Tool/software:

https://e2e.ti.com/support/microcontrollers/c2000-microcontrollers-group/c2000/f/c2000-microcontrollers-forum/1437888/tms320f280039c-when-i-divide-the-clock-i-find-that-the-hwbist-test-is-not-working-properly

In the previous post, I asked about clock division in HWBIST. I’d like to know if there are any new updates on this.

There is no new updates from TI side on this, and is not being actively pursued.  I would consider running HWBIST at reduced clock to not be supported.

Best,

Matthew