• Locked Locked
  • Replies 6 replies
  • Subscribers 62 subscribers
  • Views 658 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.

Original question:

BQ76942: POR reset by internal watchdog after exit of config mode

BQ76942: POR set after exit from config mode - fullscan check enough?

Erik Almqvist
Prodigy 140 points
Part Number: BQ76942


Hi!

We have some issues related to watchdog resets and POR. This is somewhat a continuation on the thread that is available here:

BQ76942: POR reset by internal watchdog after exit of config mode - Power management forum - Power management - TI E2E support forums

We are trying to confirm that our sequence for performing a reset via pin works. We've had a partial reset as a part of our configuration sequence. As we remove this - we want to confirm that a partial reset sequence works. I'm currently running loops of reset AFE via PIN every 30 seconds or so.

Currently, we randomly see internal watchdog triggers by the AFE after exit of our configuration sequence. Our configuration sequence is described in the original post, but basically our configuration sequence do:

--------------------------------

* Check AFE data memory registers for consistency.
  If not consistent - configure:


* Enter config mode
* Write parameters
* Exit config mode
* Clear full scan bit
* Wait for fullscan bit to be set.

* Re-check AFE data memory registers for consistency.

----------------------------------------------

In our tests, we've seen that the POR bit is set ~500 ms after the fullscan check completed. We also see that we on occasion won't get response when querying data settings ~300ms after full scan has completed



A.) What should we wait for after exit of config mode - to make sure AFE is in a good state?

We've seen that query of DA STATUS 5, subcommand and data memory too soon after config mode can trigger the watchdog. Checking battery status and control status request seems OK. 

B.) We've seen that waiting for fullscan bit to be set after clearing it takes around 150ms up to 1.5 seconds. 150 ms seems to fast to us, since a full measurement round should take 250 ms according to the manual. Doesn't the reset of the fullscan bit reset the measurement cycle? 

In the original post, we discussed that the time it takes for AFE to store the parameters upon exiting config mode is dynamic. When returning from shutdown mode, all parameters needs to be written. But after a partial reset, the parameters are likely only re-stored. Does this make the sequencing in the AFE different?

Best regards
//Erik Almqvist

Another related question regarding POR:

BQ76942: POR bit not set after shutdown - Power management forum - Power management - TI E2E support forums

  • TI__Guru* 81035 points

    Hi Erik,

    For question A, checking the Battery Status (direct command) is the best way. Subcommands are slower and use more processor bandwidth - I have not seen the watchdog trigger before like this, but this seems to indicate the device is not ready if you are seeing this trigger.

    For B, is the device in Sleep mode? The voltage measurement loop is much slower in Sleep mode as shown in this technical article: https://e2e.ti.com/blogs_/b/powerhouse/posts/improving-voltage-measurement-accuracy-in-battery-monitoring-systems Maybe you can try to send the SLEEP_DISABLE() command after checking the Battery Status and before checking the FULLSCAN timing. Then you can send the SLEEP_ENABLE() command to allow the device to re-enter Sleep mode.

    Regards,

    Matt

  • in reply to Matt Sunna
    Prodigy 140 points

    Hi Matt!

    A.) Sorry, it's hard to write a good post specifying enough details without making it an essay. We do check the battery status as a criteria for exiting the config mode. After this, we also clear the fullscan bit and waits for it to be reset. This doesn't seem enough in our case though - and we are looking for the right criteria to continue execution.

    B.) We are not allowing sleep mode. While we don't send the disable sleep mode in this sequence, I think it's unlikely to be set (but will check when I can).

  • in reply to Erik Almqvist
    TI__Guru* 81035 points

    Hi Erik,

    It will be good to double-confirm the part is not in Sleep. I cannot think of another reason for slower timing on the FULLSCAN bit.

    Regards,

    Matt

  • in reply to Matt Sunna
    Prodigy 140 points

    Hi Matt!

    We had some progress on this from our end. In this test, we initiated the config mode by RST_SHUT.I t seems like when we do, the INITCOMP bit goes low. If we wait for this bit to set again before we continue, the configuration sequence seems more stable. So, one thesis is that the command sequencing after the RST_SHUT sequence ended AFE watchdog to trigger - maybe the AFE was not done with the internal reset. 

    RST_SHUT sequencing is not that well documented, but it's stated that the AFE should exit sleep mode when pin is asserted so it seems weird to have the AFE in sleep mode during this time.

    (It also seems like the INITCOMP bit resets by enter / exit of config mode). 

    We'll not use the RST_SHUT sequencing as part of normal configuration - but I think it's important that the sequencing work for edge cases and error handling. 

    Best regards

    //Erik Almqvist

  • in reply to Erik Almqvist
    TI__Guru* 81035 points

    Thanks for sharing this update Erik. I was not aware of this behavior on the INITCOMP bit, so this is good to know and something I will look into more.

  • in reply to Matt Sunna
    Prodigy 140 points

    Thanks Matt!

    I hope this is the last we see of the POR bit and the watchdog...

    Best regards
    //Erik Almqvist