• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 62 subscribers
  • Views 383 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.

Suitability of BQ34Z110 for a small solar off-grid inverter

Guy Ovadia
Intellectual 380 points
Other Parts Discussed in Thread: BQ34Z110, BQ34Z100-G1

Hello,


We are designing a small off-grid solar power system for single homes in rural areas. I'm considering using the BQ34Z110 for battery SOC monitoring, and I have several questions regarding its application and suitability. answers from TI would be much appreciated!

First, a brief description of the system: a solar panel array feeds a charge controller, which charges a battery pack. The load is an DC-AC inverter which supplies AC to the home. a board will be made that would contain the BQ34Z110 and an MCU. This board will monitor battery SOC and health, will disconnect or limit the load when the battery is too discharged, and will also log various battery parameters for later reference.

preliminary specs:

200Ah, 24V battery pack
40A max charge / discharge current

I have the following questions:

1. Battery compatibility: we'll probably use one of the following batteries (two of them in series, and possibly two series strings in parallel).

http://www.narada-ap.com/products/htb-313k-series-12v
http://www.hoppecke.com/energy-storage/lead-acid/opzv-bloc-solarpower.html

Do you expect that either of these batteries would fit one of the existing chemistry profiles? I understand you cannot guarantee it just from the datasheets, but I'm just trying to understand what is realistic to expect (i.e. probably yes / probably no)

2. In case the existing profiles do not fit, I understand I will need to have two batteries shipped to you so you can analyze them and create a new profile. Does TI charge anything for this process? what would be an expected time-frame for this process if it's needed?

3. Is there a C software library to interface the BQ34Z110? if there is, for which MCU was it written?

4. Evaluation module: I want to use the EVM to test the setup before going to a custom board production, but I see the EVM is not designed for my current levels (sense resistor of 10 mOhm on board seems to be unsuitable for 40A). can I use an external sense resistor instead? is it realistic to get decent accuracy from such a setup?

5. How do I tell the BQ34Z110 which sense resistor was installed? I noticed a register for the ratio of my chosen voltage divider, but not for my chosen sense resistor value (or maybe I missed it?)

6. Algorithm suitability: considering my target application, I'd expect that the charge "cycles" would be short and incomplete, i.e.rapid switching between charging and discharging. This is because both my energy source my load are unpredictable: sunlight comes and goes as clouds move across the sun, and users turn appliances on and off, changing the inverter's consumption from the battery. Generally, I will not have complete, uninterrupted charge cycles, but more like a few minutes of charge, a few minutes of discharge and so on. Now, the datasheet mentions the concept of "charge cycles" a lot (e.g. "the BQ34Z110 will calculate XYZ according to the previous discharge cycle), but as I explained my application will have very intermittent charging. Will this erratic behavior mess with the algorithm?

7. Production: what would I need to calibrate per each manufactured unit? will I need to do a charge / discharge cycle for each manufactured unit for calibration, or just once during development? will changes in the resistance of the high-current path (battery wiring to the board) affect system accuracy? if so, how can I compensate?

thank you for taking the time to read my long post... looking forward for answers so I can go forward with this project!


thanks,

Guy.

  • 0
    TI__Guru 73830 points

    Guy,

    I recommend that you switch to the bq34z100-G1. This device was released in January and supports PbA batteries and offers some gauging enhancements over the bq34z110. The bq34z110 will also be moved to NRND soon.

    We have not characterized your cells, so I cannot comment on which ChemID could be a match. You can collect log data and we can use a Mathcad program to try to find a match. We need to check for a Mathcad match before pursuing a full characterization. You can use the attached document for guidance in collecting the data.

    We do not have any C program examples to help with programming the device.

    The sense resistor will need to be sized to keep the voltage drop across the sense resistor from exceeding 125mV. Furthermore, the current register can only support 32A, so the current and capacity will need to be scaled to stay below this level. The EVM can only support about 7A, so you will have to remove the sense resistor from the EVM and use an external one. The gauga can support abotu 29Ah of passed charge, so you will have to scale the current and capacity to support 200A. 200Ah/29Ah = 6.9, so you will have to use at least 7x scaling.

    The device does not have a sense resistor register, but the gauge know the value of the sense resistor from the current calibration parameters.

    Your charging profile is not ideal for Impedance Track. The gauge will require periodic rest periods to update Qmax for learning.

    You will need to run an optimization cycle to fine tune the Ra table and Qmax for your pack you can use this data to create a "golden file" that can be programmed into production packs. Ideally, you should calibrate every pack, but you can get by with just calibrating the voltage. We have a document that provides information on calibration options. It is named "Going to Production..." and is available on the product website. Changes in resistance will affect performance, but learning cycles will update the Ra table and Qmax to compensate. I would recommend using the EVM and run some profiles to make sure the the gauge will support your operating profiles.

    Tom