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.

BQ34110: BAT measurement resolution

Part Number: BQ34110
Other Parts Discussed in Thread: BQSTUDIO

Tool/software:

Hi TI support team,

We use the BQ34110 device to gauge a 10S NiMH battery pack with voltage ranging from ~8V to 16V. The device is configured with an external voltage divider for BAT measurement. The bottom resistor is 16K and the top resistor is 270K giving a full scale range of 17.875V.

According to table 6.13 of the datasheet the ADC resolution is at least 14 bits which should give a battery voltage resolution of 1.09mV but the battery voltage reported by BQStudio has a resolution of 10mV.

Is it normal operation ? If yes, what is the reason because it reduces the resolution and thus the precision to 10~11 bits ?

Regards

  • Hi,

    The resolution should be in mV.

    Ex: The gauge can measure 5.792V and will be reported as 5792mV.

    Regards,

    Diego

  • Hi Diego,

    You didn't answer the question at all. My question is not about unit, I saw that the voltage reported in BQStudio is in mV, but about the resolution or granularity or step.

    As I say, according to table 6.13 of the datasheet the ADC resolution is at least 14 bits which in my case with the voltage divider indicated in my first post, should give me a battery voltage resolution of  ~1mV but the battery voltage reported in BQStudio has a resolution or step of 10mV which corresponds to an ADC resolution of 10~11 bits and not to the 14 bits announced.

    For example if I apply a BAT voltage of 11998mV, the voltage reported by the gauge in BQStudio is 12000mV but if I slowly increase the BAT voltage millivolt by millivolt, the voltage reported by the gauge stays at 12000mV until the actual BAT voltage reaches 12004mV where the voltage reported by the gauge jumps to 12010mV denoting a 10mV resolution and not a 1mV resolution as expected.

    I hope I've been clearer in my explanations.

    Regards.

  • Hi,

    What is your voltage divider? Are you doing any voltage scaling? You are probably losing resolution due to this.

    Regards,

    Diego

  • Hi Diego,

    As I told you in my first post, the divider bottom resistor is 16K and the divider top resistor is 270K. The battery voltage range is 8V to 16V (see picture below).

    I don't use voltage scaling. Data flash "Number of Series Cells" is set to 10 for a 10S NiMH battery pack and data flash "Voltage Divider" is 17973mV after calibration (please find attached the .gg.csv file).

    So correct me if I'm wrong but if the ADC has a 14 bits resolution with a full scale input range of 1.0V, the resolution is 1000mV / 2^14 = 61µV

    With the voltage divider above, at the pack voltage level it will correspond to a full scale range of 1.0V x (1 + 270/16) = 17.875V and with 14 bits of resolution this gives a resolution in mV of 17875mV / 2^14 = 1.09mV but the voltage reported by the gauge has a resolution of 10mV and not 1.09mV as expected.

    BQ34110_DF.gg.csv

    Regards.

  • Hi,

    So, when I tested with a small voltage divider, I am able to get mV accuracy on my end, however with a large voltage divider the resolution on the BAT pin is reduced. When you apply 12000mV thru your voltage divider, the voltage on the bat pin is about 671.329mV, increasing your applied voltage to 12010 only increases the voltage on the BAT pin to 671.888mV, due to the large size in your voltage divider you lose resolution on the BAT pin. However, I do not think this will cause any issues in terms of tracking SOC/SOH.

    Regards,

    Diego

  • Hi Diego,

    Happy to see that you get the same result on your side.

    I agree with you about the resolution in mV, it increases with the size of the voltage divider but I do not agree about the resolution in number of bits, this one should be strictly the same as the ADC one. If the ADC has 14 bits of resolution, the reported voltage should also have 14 bits of resolution, only the full scale range and the resolution in mV should be dependent of the voltage divider.

    So as you specify an ADC resolution of 14 bits in the datasheet, the voltage measurement reported should also have a resolution of 14 bits which with my  voltag divider gives: 1.0V x (1 + 270/16) / 2^14 = 1.09mV and not 10mV.

    I agree with you, it is unlikely that this 10mV resolution would generate SOC or SOH problems but the reason for my question is that the 1mV resolution on the voltage measurement is a customer requirement so when I designed the circuit, I assumed a 14-bit resolution which should have given me the required mV resolution but I end up with only 10mV resolution and I have no explanation.

    With a 14 bits ADC and a 16 bits register for voltage measurement, I still don't understand where and why the resolution is degraded to 10-11 bits on the reported measurement and that's the heart of the problem. Do you have an answer to this question ?

    If any factor likely to degrade the resolution of the final measurement in bits, it should be mentioned in the data sheet.

    Regards.

     

  • Hi,

    I will ask our systems team about their input on this. Please allow me some time.

    Regards,

    Diego

  • Hi,

    OK, I'll wait for your feedback. Thanks.

    Regards.

  • Hi,

    Our system team is currently OoO, we will get back to you once they return.

    Regards,

    Diego

  • Hi,

    Any news from your system team about the resolution of the voltage measurement reported by the BQ34110 ? I didn't hear from you since 13 days and the post has been changed to resolved but I didn't get any answer.

    Regards.

  • Hi,

    After aligning with the system team, the reason has to do with the internal working of the I.T. algorithm, which is proprietary information, I could not share even with an NDA in place.

    Regards,

    Diego