• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 63 subscribers
  • Views 551 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.

Whoa! - Look what needs a reset.

Michael Banak
Intellectual 915 points
Other Parts Discussed in Thread: BQ29330, BQ20Z90, BQEVSW

I understand that the Z-Track system requires diligence and observation to apply properly. The depth of battery mnagement utilities in this fine system is simply breathtaking, and management types have sat glassy-eyed as I explain my design decisions. The complexity doesn't scare me, given the benefits. So it comes as no surprise when I discover an unexpected state wihin the Z-Track.

I have carefully selected the thresold at which I want to declare a SC failure. This is a combination of time delay (dwell time, attack ime, call it what you will) and a current level threshold at which to start counting.

No problem.

Since I am operating in an NR=1 mode, the flag for the SC condition will clear after the appropriate wait time.

Still no problem.

However, when operation restores, the SC threshold and dwell time revert to TI's default values. Only a pack resest (0x0041) will restore my desired prgram values.

Question: When the SC fault occurs, is EVERYTHING returned to default values, thus requiring a pack reset?

I ask for any light an expert can cast on this.

Regards,

Michael A. Banak

http://eclectic-engineering.net

 

  • 0
    TI__Genius 14985 points

    Hi Mike,

    Glad you like the algorithm.  Can you tell us which device and firmware version you are observing display this behavior?

  • 0
    TI__Intellectual 825 points

    Michael,

    This brings up an interesting topic - How AFE (such as bq29330) registers work.

    At device reset (either a POR, or a software full reset through MAC command 0x0041), the GG (such as z90) writes to the AFE registers with what it wants. The configuration for the SC thresholds and dwell time, etc can be found from 1st Level Safety class in the data flash. Once committed, the content in the AFE for the short circuit protection (and other protection thresholds) will not change. Even if there is a reset later on, the GG would still write the same configuration to the AFE.

    Now, if you are evaluating the chipset, and change the AFE settings from the GG/1st Level Safety, these changes are not committed to the AFE. It takes a reset to update the GG content to the AFE. So this is probably what happened: in your test, the reset step is omitted after changing the values in the AFE.

    This information is documented in literature SLUA421 in the bq20z90 v150 webfolder:

    http://focus.ti.com/docs/prod/folders/print/bq20z90-v150.html

    Direct link: http://focus.ti.com/general/docs/litabsmultiplefilelist.tsp?literatureNumber=slua421a

  • 0 in reply to dMax
    Intellectual 915 points

    Thank you for kindly acknowledging my inquiry. The following was copied from the header of my gg file. I trust that this will tell you everything: Also, my reply to "SW", below, might be of ineterst to you.

     

    Regards,

    MABanak

     

    [Header]

    bq EVSW Version = 0.9.59

    DeviceName = bq20z90 v1.10

    Time = 1/14/2010 4:44:04 PM

     

  • 0 in reply to SW5307
    Intellectual 915 points

    Thank you "SW" for your suggestion.

    Your suspicions are justified, but I don't think they apply in this case. I did the SC fault several times, and observed the following sequence:

     

    1 - Set my SCD threshold configuration in bqEVSW

    2 - Write the flash with all those updates

    3 - do a sc test at a known excess current value (just a carefully selected load resistor, actually)

    4 - observe that the wait time matches my target values

    5 - Observe that the SCD flag annunciates

    6 - Wait for the SCD flag to self clear (NR = 1)

    7 - Redo the test

    8 - observe that the wait time now matches TI's default

    9 - RESET the pack (0x0041 command)

    10 - Redo the test

    11 - observe that the wait time matches my target values

    Now repeat #5 through #11


    You can try this too by finding some 1-ohm resistor somewhere, set the SCD threshold to bring this out.

    I think I will look download your references now. I am always willing to redo the test, of course.

    Regards,

    MABanak

     

  • 0 in reply to Michael Banak
    TI__Intellectual 825 points

    Mike,

    I have tested and was unable to see what you saw. I would not use a 1 ohm resistor for this test because the AFE comparitor is very sensitive to the large offset caused by the 1 ohm resistor. I always used the 10 mohm sense resistor for this test, and got consistent result.

    1 - Set my SCD threshold configuration in bqEVSW

    2 - Write the flash with all those updates

    //sw: add a RESET step here to commit the DF values into the AFE register; This is critical.

    3 - do a sc test at a known excess current value (just a carefully selected load resistor, actually)

    4 - observe that the wait time matches my target values //sw, without a RESET above, this does not make sense. Is your delay time very close to TI default value? Can you indicate the value you used?

    5 - Observe that the SCD flag annunciates

    6 - Wait for the SCD flag to self clear (NR = 1)

    7 - Redo the test

    8 - observe that the wait time now matches TI's default

    9 - RESET the pack (0x0041 command)

    10 - Redo the test

    11 - observe that the wait time matches my target values //sw, Since we had a RESET in step 9, this again does not make sense. The delay time should match TI default.