• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 396 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

BQ27531-G1: ChemId matching

Nils Larsen
Prodigy 70 points
Part Number: BQ27531-G1
Other Parts Discussed in Thread: GPCCHEM, BQSTUDIO, EV2400, BQ24192,

I'm in the process of determining the BQ27531-chemId matching our battery model. With the GPCCHEM tool, I consecutively get the same chemId-results when I repeat the test. The "max. deviation, %" value is around 1,5%-2,0%. However, The textual description of the matched chemId does not fit the battery model very well.
For example, I get
chemId 2073 with 1,68% max. deviation. Model: "cylindrical (2490mAh) - NiCoMn/carbon"
chemId 1102 with 1,81% max. deviation. Model: "laminate (4380mAh) - LiCoO2/carbon 11"
chemId 2446 with 1,83% max. deviation. Model: "PGF515974-2 (3720mAh) - LiMn2O4 (Co,Ni)/carbon, 4,35V"

But the battery is a prismatic (65x40x6,0mm) 3,7V LiPo battery with a nominal capacity of 2000mAh. The two above results seem pretty far from the "facts" of our battery model? Does it make any sense to proceed (do learning curves) with one of these chemIDs? Our battery has a built-in protection circuit with a specified "protection board resistance" of 30-40 mOhm.  Unfortunately, we don't have details about the anode, cathode and chemistry.

Test setup details:
The battery is installed in the target application. The target application has been modified such than an external programmable power supply/load can charge and discharge the battery through the sense resistor of the BQ27531 of the target application (normally the battery would be charged using a BQ24192). Also, the BQ27531 is connected to TI EV2400 (normally it is connected to the application processor). Before running the discharge-charge-relax-discharge-relax-cycle, the BQ27531 was calibrated (CC Offset, Board Offset, Temperature, Current 1000mA, Voltage 4000mV) using bqStudio. The test is run as described in "Simple Guide to Chemical ID Selection Tool (GPC)" (SLVA725A).

The voltage drop of the battery "protection board resistance", the 5cm battery cable and battery connector is NOT being compensated for, exactly as it would be the case in the target application. The voltage drop may be around 25mV at max test current (400mA).

Thank you very much for any hints or assistance.

  • 0
    TI__Guru 57697 points

    Hello Nils,

    Series resistances can cause errros with chemistry matching. IVT logs for chemistry matching are expected to be taken with bare cells. We recommend that you run the test with bare cells.

  • 0
    Prodigy 70 points

    Thank you for your answer, Shirish. I could probably remove the integral protection circuit of the battery pack (it needs delicate tampering with the battery) and  repeat the recording of IVT logs.

    Assume that the GPCTOOL will then determine the "right" chemId for the bare cell. The protection circuit will surely be part of the final application, however. How would I tell BQ27531-G1 that there is a protection circuit to compensate for? Just using the found chemId number would "fool" the gauge, I believe, since there is another series resistance than during the IVT recording.

  • 0 in reply to Nils Larsen
    TI__Guru 57697 points

    Hello Nils,

    The log with bare cells is used to match curves that are the characteristics of the cell. Other factors will be accounted for during the learning cycle.

  • 0
    Prodigy 70 points

    (I was away due to holidays)
    Ok, that makes sense. But it rises more questions. If so small voltage measurement errors affect the chemistry matching so significantly, I would surely also need to take the wiring and contact resistance into account?
    What magnitude of measurement errors are acceptable?

    I just checked:
    At zero current, the bq27531 gauge measures the bare cell voltage with an error of less than 2 mV (I use a Fluke 175 multimeter as a reference).
    At 400 mA charging, the measurement delta of the gauge is +38mV in reference to the terminals of the bare cell. The pack protection circuit only accounts for 9mV of the error, the rest is from a fuse, contact and wireing resistance.
    At 200 mA discharging, the measurement delta of the gauge is -20mV in reference to the terminals of the bare cell. The pack protection circuit only accounts for 5mV error.

    Is it feasible to use the gauge to record the IVT logs at all?

  • 0 in reply to Nils Larsen
    TI__Guru 57697 points

    Hello Nils,

    We recommend using thick wires for connections. The gauge measures cell voltage. The pack current does not flow in this path. There should be no issue as long as the connections are made correctly.

  • 0
    Prodigy 70 points

    Thanks for the recommandation using thick wires. I have chosen to compensate the voltage measured by the gauge instead. At several current levels, I compared the voltage mesaured by the gauge with the bare cell voltage measured by a Fluke 175 multimeter. I determined a resistance of 103 mOhm causing the voltage difference. With this value, I compensated the voltage in the IVT logs and resent them to the GPCCHEM-Tool.
    I now do get sligthly different results from GPCCHEM, but the textual description of the found chemId still seem to deviate significantly from the cell we have (a prismatic LiPo 4,2V with 2020mAh nominal capacity) - in the "top five" there is even 600mAh cell:

    Best chemical ID : 242    Best chemical ID max. deviation, % : 2.5       
                   
    Summary of all IDs with max. DOD deviation below 3%        
            
    Chem ID    max DOD error, %    Max R deviation, ratio
    242    2.5    0.33      = Panasonic CGR-18650CH (2250mAH) LiNiO2 (Co, Mn doped)
    2127    2.61    0.58  = Sanyo UR18650P-H0MCA (2200mAh) NiCoMn/carbon
    2151    2.77    0.42  =Advanced AE18650C-26 (2600mAh) NiCoMn/carbon
    2135    2.94    0.76   = US14430VR2 (600mAh) NiCoMn/carbon
    257    2.96    0.5       = Coslight CA705462 (2600mAh) NiCoMn/carbon

    Does it make any sense to go on to the learning step with any of these chemIds? After all they are within the recommended tolerance of 3%.

  • 0 in reply to Nils Larsen
    TI__Guru 57697 points

    Hello Nils,

    The closest match is 242 and it is within 3%. Of course this is with compensated values. Any inaccuracy in the compensation will impact accuracy.

  • 0 in reply to Shirish
    TI__Guru 57697 points

    Hello Nils,

    The battery type does not match, i am not sure if that has an impact. Let me check with the team on that.

  • 0 in reply to Shirish
    TI__Guru 57697 points

    Hello Nils,

    I checked with our battery expert and learnt that Li-polymer does not define cathode chemistry, it just means cell is in polymer pouch package.

    So you can use any of the matching IDs. Using the best match is recommended.