• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 2 answers
  • Subscribers 30 subscribers
  • Views 443 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMS570LC4357: Watchdog not mentioned as a safety measure for detecting latent faults in VIM?

Stephan HERMANNS
Intellectual 590 points
Part Number: TMS570LC4357

Dear Team of TI Hercules support,

Walking through the Safety manual spnu540a of the TMS570LC4357 MCU, i faced that watchdogs are not called as measures for detecting latent faults regarding VIM.

(Q1) Could you please explain reasons, why watchdogs are not called for detecting latent faults in the scope of the VIM?
I could imagine that a execution of VIM6 "Periodic Software Readback of Static Configuration Registers" missed by an error could be detected by a watchdog.

Thank you very much for your support!

Stephan

  • 0
    TI__Mastermind 49120 points

    Hi Stephan,

    Internal / external watchdogs are called out as tests for the periodic checking of statis config registers for VIM as well. For example, VIM6 in SPNU540A shows:

  • 0 in reply to Sunil Oak
    Intellectual 590 points

    Hi, Sunil,

    Thank you very much for answering.

    Indeed, i forgot to take the appendix A of the safety manual into account.
    Unfortunately for every "Possible Test for Diagnostics" Table 4 doesn't call whether a Test is on single point faults or on latent faults.
    E.g. for VIM6 the means CLK5B and CLK5C are called but i cannot discover whether they are to detect single point faults of VIM6 or latent faults.

    Section 6.19 of the Safety manual gives a hint that internal and external watchdogs are suggested as diagnostics for single point faults

    but not for detecting latent faults, as they aren't contained in the following list taken from section 6.19:

    Sunil, maybe section 6.19 should call internal and external watchdogs as diagnostics for latent faults too?

    Again, thank you very much - Enjoy the remaining day!

    Stephan

  • 0 in reply to Stephan HERMANNS
    TI__Mastermind 49120 points

    Hi Stephan,

    The actual software diagnostic mechanism is VIM6, which is a periodic check of the static configuration registers in the VIM module. This check is required to detect latent defects in the VIM registers, and must be executed by the application software periodically. Since this is an entirely software-based diagnostic mechanism it can be verified using the tests for the processor's code execution sequence such as watchdogs, core compare module and Flash/RAM ECC checks.

    Hope this helps.

    Regards, Sunil

  • 0 in reply to Sunil Oak
    Intellectual 590 points

    Hi, Sunil,

    Maybe i'm wrong, but i wouldn't call VIM6 as a test for latent faults because a wrong configuration of the VIM
    could directly cause that a safety function isn't working well.
    Just as an possible example a stuck-at-low in [REQENASETx] could lead to that an enable request is blocked even it shouldn't. 

    I agree that watchdogs CLK5A/B/C are tests for latent faults, e.g. that VIM6 isn't carried out periodically.
    And - except my concern above - i agree to the procedure you've outlined.

    The current problem with the Safety manual spnu540a i have is, that i cannot draw a DOORS link to the internal watchdog measure in section 6.19 for
    detecting latent faults because in this section the internal watchdog is suggested for detecting single point faults in VIM only.
    But that topic would need a change of the Safety manual to be resolved.

    I appreciate your support very much!

    Stephan