• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 63 subscribers
  • Views 203 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

BQ76952: DC Ripple

Paulo Santos
Prodigy 10 points
Part Number: BQ76952
Other Parts Discussed in Thread: BQ79616-Q1

Hi,

I'm developing a BMS based on BQ76952 and everything went well until I connected it to an inverter.

Using an electronic load I could test all the functions and everything worked as expected. With the inverter all the readings became very instable!

For the current measurement I could fix the problem by using the CC3 current present on DASTATUS5 with Settings:Configuration:CC3 Samples = 80. But for the voltages I couldn't find any internal ADC post-processing that could help smoothing (avg) the readings. Then, with an oscilloscope and a differential probe, I measured the ripple I was getting at the BMS terminals (between VC0 and VC16), and I got 1.93Vp-p @ 100Hz (almost a sinewave) when the inverter was taking 22A of current.

Not discussing here how well the battery pack was constructed, I would like to have the BMS capable of handling this situation. Some inverters in order to reduce size and cost they save on input capacitors, this means this could be recurrent situation even with a good battery pack.

Because of the 100Hz ripple, I don't think I can have enough sample rate from the BQ76952 to make post processing and even if I did, probably the voltage ripple on the cells would mess with the internal algorithm for autonomous cell balancing (because the readings are done sequentially on each cell, thus getting different readings over time due to the ripple voltage).

So I tried to increase the values of the differential RC filter with R = 1K and C = 4.7uF and it improved some measurements.

I got the following voltage variations after this change (making readings once a second):

- the Battery Voltage Sum on DASTATUS5() with a variation of 0.3V, 

- the Stack (VC16 pin) voltage with a variation of 1.72V,

- the PACK pin voltage with a variation of 1.93V,

- the LD pin voltage with a variation of 1.94V,

- the Max Cell Voltage on DASTATUS5() with a variation of 44mV,

- the Min Cell Voltage on DASTATUS5() with a variation of 156mV.

I think these results make some sense, the Battery Voltage Sum benefits from the differential smoothing provided by the filter on each cell, the Stack (VC16 pin) voltage not being a differential measurement, the filter doesn't affect so much, and the PACK and LD pins don't have any filters at all, so basically they are showing the ripple measured with the oscilloscope.

My questions are:

Is this a good approach? Any other suggestions to tackle this issue?

Does increasing the values of the filters will decrease the precision of the measurements on each cell?

Does the internal autonomous balancing algorithm gets affected? What if we increase even more the current and consequently the voltage ripple? What is the ripple that the IC can handle?

Thanks,

Paulo Santos

  • 0
    TI__Guru* 80964 points

    Hi Paulo,

    Increasing the filter value like this will reduce the accuracy of the cell voltage measurements and will introduce other issues. The maximum cell input resistor allowed by the datasheet is 100 ohms. Higher values will result in accuracy issues. A high RC time constant will have a large settling time which will probably have a significant affect on cell voltage measurement variation during cell balancing. I normally would suggest 100 ohms and 0.22uF for your cell input filters. Additionally, the VC16 pin draws a little current during balancing, so we often recommend using a smaller resistor on VC16 to reduce the IR drop - so 20 ohms on VC16 is usually best.

    The IC is measuring voltage every 3 ms. The decision for which cells to balance is on a 1 second interval. Can you apply filtering to the battery PACK pins?

    Best regards,

    Matt

  • 0 in reply to Matt Sunna
    Prodigy 10 points

    Hello Matt,

    Thank you for your help.

    During my research on this topic I found a battery manufacturer that also have BMS, that looks like they had faced a similar problem and their solutions was, like your suggestion, adding big capacitors on the inverter terminals to lower the ripple on the DC bus.

    Additional DC Capacitors required

    Even though this would be a workable solution, is not very elegant!

    In order to size this capacitors, what would be the maximum allowed ripple for the BQ76952 to work properly? I couldn't find this information on the datasheet.

    Using a more expensive part would fix the problem? I was looking at this part: BQ79616-Q1 and in the description it says:

    "The integrated front-end filters enable the system to implement with simple, low voltage rating, differential RC filters on the cell input channels. The integrated, post-ADC, low-pass filters enable filtered, DC-like, voltage measurements for better state of charge (SOC) calculation."

    Any other recommendations for similar parts?

    Of course this would make the product more expensive and basically I would have to start design almost from scratch again! I would like to avoid going this at this point!

    Best regards,

    Paulo Santos

  • +1 in reply to Paulo Santos
    TI__Guru* 80964 points

    Hi Paulo,

    I think the BQ76952 is measuring the voltage at the pins of the device correctly, so it is working properly. The problem is that the voltage is varying on the pins due to the ripple. It seems like you would want to address the ripple on the DC bus. 100 Hz is pretty low frequency, so I can see how this is challenging without large caps in the battery pack.

    I think another IC would have the same issue unless it allows for larger RC filter components on the input pins. But with large RC components on the cell inputs, it will affect the voltage observed by the IC and slow the reaction time if the voltage changes rapidly.

    Regards,

    Matt