• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 323 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.

BQ77905: Fault sampling timing

Lon Glazner
Prodigy 40 points

Part Number: BQ77905
Other Parts Discussed in Thread: BQ76930,

Provide case details or comments:

We have a battery management system that incorporates a BQ76930 under processor control and a pair of BQ7790508 in parallel with the battery cells and BQ76930.  Due to a battery manufacturing implementation we have found that cell balancing in the BQ76930 causes the BQ7790508's to detect erroneous over voltage conditions.

Unfortunately the batteries can not be rebuilt at this time.  There are several firmware workarounds that might alleviate this issue.  One of which is to limit balancing time in a way that the BQ7790508 fault detection does not count up to the 5 samples as described in the datasheet, and has the opportunity to count down with balancing turned off.

In order to validate that I need to know the approximate sample time implemented on the BQ7790508.  I can't find that information in the datasheet.  This diagram does describe the fault detection method.



And I did find this image in an application note that seems to indicate the approximate sample time as ~100mS.



Can you please provide some insight into the sample time implemented in the BQ7790508, i.e. how long it takes to measure 5 consecutive samples and trigger the over-voltage fault?

  • 0
    TI__Genius 12765 points

    Hey Lon, 

    The internal device clock frequency combined with the actual fault and recovery delays can be used to estimate the counts. This was never intended to be relevant as is not discussed in the datasheet. 

    Furthermore, disabling cell balancing before the BQ77905 can reach 5 fault counts would be impractical, as cell balancing does not disable immediately after the command is send to the BQ76930.

    What is the manufacturing error in the batteries? Do they have internal protection systems?

    Do you have input filter resistors and capacitors before the voltage measurement pins on the BQ77905? Those should filter out any waveforms created when the BQ76930 is balancing.

    Thanks,
    Caleb

  • 0 in reply to Caleb Perez
    Prodigy 40 points

    Hi Caleb,

    Thanks of the response. 

    The battery packs have PTCs that introduce an IR loss under high currents.  When the cell balancing is enabled in the BQ76930 the balancing resistor can act like a voltage divider with the PTCs on the input to the BQ77905.  This causes errors in measurement by the BQ77905 and over voltage faults.  It has not been a problem for the BQ76930 likely because it measures the cell voltage with balancing off.

    We do have filter components between the BQ77905 inputs and the battery cell connections.  They are not sized to filter the 175mS balancing "on" time used by the BQ76930.

    Balancing doesn't have to turn off immediately in the BQ76930, as long as it turns off <=250mS.

    Duty cycling the balancing by firmware has been seen to reduce the BQ77 over voltage faults.  Turning off cell balancing eliminates the BQ77 over voltage faults.  Some systems seem to have very few of these faults.  A small number have them constantly.   

    Since we're using the 1S delay part (BQ7790508) can I assume sample time min = 0.8S/5 samples and max=1.4S/5 samples? 

  • 0 in reply to Lon Glazner
    TI__Genius 12765 points

    Hey Lon,

    Yes you are correct. The BQ7630 does not see the voltage error because it does not measure voltage while balancing.

    Our data shows that worst case for balancing to turn off is 275ms between command arrive and all balancing turning off. Best case is 0ms. It really depends on if the cell balancing loop is mid balance or off balance.

    I cannot give you an accurate estimation of sample time.

    Thanks,
    Caleb