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.

BQ27441-G1: How much temperature reading affects the prediction?

Part Number: BQ27441-G1
Other Parts Discussed in Thread: BQ24232, BQ27426, GPCCHEM

Hello,

I am writing because we have run into a problem and I would like to evaluate the real need of a redesign.

Our device is composed of 2 pcbs connected together by 2x10 ribbon cables. In PCB 1, we have the processor a BQ24232 and BQ27441. PCB 2 has the battery which has a 10k NTC (I have omitted other components not related to this issue).

We are connecting the NTC to the BQ24232 but we are not telling the temperature of the battery to the BQ27441. The internal temp sensor can not be used as the battery is not next to the gauge. It would be hard to move the gauge to the other side as we do not have many available wires.

How much does the temperature affect the prediction? Will not having an accurate temp reading make the gauge inaccurate after using it for some time?

Thanks.

  • The temperature is an important and integral part of our algorithm. It is important to have the gauge close to the battery or use a thermistor placed on the battery to measure temperature. That being said i am afraid you will have to rework your design.
    thanks
    Onyx
  • Thanks Onyx for your response. We will then have to let the gauge know the temp of the battery.

    Is there any required accuracy/tolerance or desired update rate for the temperature? 

    We already have an NTC (10k) in the battery (it already comes with one). We are connecting it to a BQ24232.

    In order to not have 2 NTCs, could we use this NTC for the gauge and not have any for the BQ24232? We will be able to control the charging by using the CE pin of the BQ24232. Is there any alternative part for the BQ27441 that has an NTC input which would work best for this situation?

    Thanks for your help!

  • Hi Leinho.

    +/- 2 deg error  should be ok. Your method of connecting the thermistor to the gauge and then controlling the charger via your host is a good one and what most customers typically do. Why  do you need another gauge other than the bq27441? You can use the bq27520 flash gauge .

    thanks

    Onyx

  • Hi Onyx,

    No, I do not need another one. The BQ27441 has been working well so far. We were just thinking that there might be an alternative for our situations. If you think that this method is what´s normally used, It is going to work fine then.

    How fast should I be checking the NTC and updating the gauge? One of our main requirements is low consumption and we would need to check temperature just as much as required.

    Thanks for your assistance!

    PS: I asked you about the deg error as the datasheet of the gauge says that the internal temp sensor has a resolution of 0.1 K deg

  • The number i gave is the acceptable error for the algorithm. Our temperature measurement accuracy is better than that.


    If you will use the NTC ffor the gauge why do you plan on checking it? Except you are planning on using it for the charger and then writing that information to the gauge. The gauge measures temperature once every second.

    thanks
    Onyx
  • The idea is what you said which it it also what´s normally done (I can not move either the battery or the gauge):

    - The Microprocessor would have to determine the temperature of the battery using the NTC
    - The Microprocessor is going to write the temp in the gauge using I2C.
    - The NTC will not be connected to the charger (BQ24232) as I do not think it can be shared.
    - The gauge would detect if the temp is too high and notify the microprocessor (is this possible?).
    - The microprocessor will control the CE of the BQ24232.

    Thanks for confirming that the gauge measures the temp once every second. I can start with that both while charging and discharging to see how this impacts our overall power consumption.
  • Leinho
    Most batteries have a 10k thermistor (NTC) which can be connected to the TS pin of the gauge and ground. That seems to be your case here. Why do you need the uC to measure the temperature and communicate to the gauge. You seem to be missing my point.
    -The thermistor is connected to the guage.
    -The uC measures the temperature from the gauge and communicates that to the charger or simply controls the CE pin.
    -it there is an over-temperature condition, the over temperature flag will get set. The host can be monitoring this flag for a safety condition.
    pls see the technical reference manual for additional details.
    thanks
    Onyx
  • Onyx,

    Yes, that's my case. Yes, I must be missing something. I have just reread the datasheets of both components but there is now way to communicate with the BQ24232 to tell the temperature and there is no pin in the BQ27441 that I can directly connect the thermistor to in order to be used to measure the battery temperature. The only pin that is shown as where the thermister can be connected to is BIN but it seems to be used for battery insertion detection.

    Please, correct me if I am missing something.
  • yes, the BIN pin is where the thermistor will be connected to on our gauge. On our other gauges it is called TS pin that is why I used TS.
  • How to instruct the gauge about this? I have checked in both the datasheet and the user guide (SLUUAC9A) and nothing is explained on this regard.
  • Leinho,
    That is captured in technical reference manual. Battery insertion as well as temperature measurement. Have you looked into that?
    thanks
    Onyx
  • Yes, I have looked into that (SLUUAC9A) but according to this document, BIN is not used for temperature measurement. You said I could connect the NTC to bin and the gauge would measure its value to predit the temperature. According to what I have read, this would not be possible.
  • Leinho
    I am so sorry. I just double checked the document, you are correct, this is one of our few gauges where the BIN pin doesn't double for battery insertion and temperature measurement. That is how most of our other gauges work so i just automatically assumed that was the situation here.. In this case, your host will have to write the temperature to the gauge. My apologies once again for the wrong information.
    thanks
    Onyx
  • No worries Onyx. Would not it be better to start using one of those gauges? Which one would you recommmend knowing what I have shared about our device?
    I mean, it would work as TI designed it and not using our own algorithm for temperature measurement.
  • Leinho
    the bq27426 would be one of such system-side rom based guage with the ability to have an external thermistor for temperature measurement.
    Pls endeavor to run the tests outlined in gpcchem to verify these rom based gauges would be a good fit and you would not have too much errror in accuracy.
    my choice of preference for a flash based sysstem side gauge would be the bq27520. It Also has the option of using an external thermistor
    thanks
    Onyx
  • Thanks for your help Onyx. I am going to look into these options then. I am well aware of the fact that the gauge parameters would need to be updated in the lab in order to have more accuraccy for specific battery. I think this type of testing can be performed before going to production and the best values can be modified in firmware and sent to the gauge using i2c interface. We would be performing such testing in the future with the new hardware.

    Kind Regards,