• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 64 subscribers
  • Views 1511 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.

BQ76930: Coulomb Counter ADC calibration

Sebastian de Leon
Prodigy 190 points

Part Number: BQ76930

Hi, in my BMS I use the AFE bq76930, and I read the battery pack current from the AFE registers, for doing both over-current protection when charging the battery, and data-logging.

I'm finding an error of approx 5% in the cc ADC reading (the current's value read is 5% higher than the real value). It looks like an ADC gain error, 5% for all current's values, no offset error.

My collegue has the opinion that, as this is a coulomb counter, with a long integration time band (250ms), it can be expected to read values higher than the real ones. Is that make sense?

I tried only with one IC so far, didn't replace it to compare. Do I have to expect a similar 5% error if I replace the afe IC for a new one?

Should I do a calibration routine at assembly time, with a std load of, for example, 1A, and rom the ADC gain for using it later at runtime?

Thanks very much,

Sebastian

  • 0
    TI__Genius 10830 points

    Hi Sebastian,

    Please note that the BMS engineers are presenting this week during our yearly BMS Deep Dive in Dallas and may not be able to answer quickly to your question. Thank you for your understanding!

    Please check out more information about this event here, in case you want to join next year.

  • 0
    TI__Guru 54715 points

    Hi Sebastian,

    No error is expected in the current due to it being a CC.  It will integrate, if current is constant it will be a constant value.  If it were a ramp you might expect an average value, similar with a pulse.

    A common source for a linear error is the trace or solder resistance of the sense resistor. Of course check the tolerance of the resistor used, but also check to see where your traces route; coming from the pads will typically give the best result with a 2 terminal resistor.  Using a meter you can often see a voltage difference between the board terminals, sense filter resistors, and sense resistor with a current in the board. There are 4 terminal resistors made so that the sense filter can be connected to the resistor with a low current path, these are less common but can reduce the connection error if used. 

    You may note the CC has a 1.5% full scale gain error, so if you replace the part there will be some difference, but I would expect the result to be similar.  If you have other boards you might compare results and see if the result is dependent on the unit, board or resistor variation. A calibration per board is a good practice, using a low current is usually a relatively low current is suitable since the CC is typically linear over most of the range beginning to curve near the end points.

  • 0 in reply to WM5295
    Prodigy 190 points

    Thanks very much for your reply, it was more than what I was expected. 

  • 0 in reply to David Hien
    Prodigy 190 points

    Hi David,

    Didn't see those training videos before, they look great, and I'll start today with them.

    Thanks.