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.

BQ34Z110: Gauging Accuracy for Multiple Cells in Series.

Part Number: BQ34Z110
Other Parts Discussed in Thread: BQ34110, BQ34Z100-G1


Aren't the gauging algorithms based on single cell models?
If so, gauging accuracy would seem best with single cell target batteries.
It would seem that gauging accuracy would be reduced with each cell added in series.
But at least the single cell modelling is accurate when a single cell can be tested.
Even better if multiple single cell tests were used to provide a stochastic cell model.
Has stochastic modeling been considered in gauge development?
I'm curious because I have some expertise with this.
Bottom line is what kind of results can be expected when characterization for ChemID selection and gauge learning is for multiple cells in series?

Each 12V Pack of my target PbA battery is 6 cells sealed so I can't collect data on individual cells.
I'm concerned about this because I understand the precision of the measurements are important for gauging accuracy.
It would seem that characterizing six cells in series as if they were a single cell, significantly obscures the desired single cell behavior even with the best measurement practices.
But perhaps not as badly for PbA chemistry over LiFePO4 for example.
Does the slope of the voltage discharge curve have a bearing on gauging accuracy?
PbA chemistry has a more pronounced slope than say LiFePO4.
If all else is equal in properly applying the gauge to single cells, would the gauging accuracy of the PbA chemistry exceed that of the LiFePO4?

  • Hello Rom,

    The ChemIDs are indeed based on single-cell models. That includes lead-acid where finding a 1s cell instead of a 6s pack is rare. The main variation between cells is in internal impedance and capacity; there should be little variation in OCV. Gauges can learn impedance and capacity in the system, so the algorithm is fairly good at accounting for cell variation. 

    Ultimately, measuring the 6s pack and finding the 1s cell's characteristics from that could cause some minor inaccuracy. However, you will ultimately use it as a pack, so in a way we are measuring the whole pack, scaling it down, and scaling it back up. In this way, it actually works to the benefit of accuracy, since the model will include all of the minor variations internally.

    The slope of the OCV is very important to gauging accuracy. You are correct that LiFePO4's OCV is much flatter compared to PbA. This does work to the benefit of lead acid for gauging. Lead acid has some other quirks that can harm gauging accuracy though. For example, one mentioned in the BQ34Z110 datasheet is the charge efficiency. It has a non-linear charge efficiency which could lead to the coulomb counting being wrong. This gauge has correction factors for this.

    Also, it is worth considering CEDV gauges for lead acid. The BQ34110 is a good choice if you wish to pursue this option. In addition the BQ34Z100-G1 has a very similar implementation to the BQ34Z110 and also supports PbA.


    Alex M.

  • Hello Alex,

    Thank you for providing an outstanding response and making me aware of the Charge Efficiency compensation and CEDV features.
    They have reassured me that BQ34Z110 is the best choice for my light electric vehicle application in hot desert climates.
    Are there any BQ34Z110 electric vehicle or outdoor application reports that might be helpful?

    Many Thanks,