• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 1516 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.

BQ76940: Coulomb Counter and Charging Current

Micah Martin
Prodigy 90 points
Part Number: BQ76940
Other Parts Discussed in Thread: ALLIGATOR

In my circuit, where the coulomb counter is implemented just like the reference design (tida00792), the charge current measurements go sour in a peculiar way.  While discharge current measurements are consistently accurate.  

As I ramp up charge current, the CC-readings begin quite accurate.  However, at around 0.6A, CC-readings begins to decrease, despite the current increasing. 

I have verified that the current is indeed increasing by measuring separately between the battery and the power supply. 

Voltage measured between BATT- and PACK- (both sides of Rsns) increases with current as expected

However, voltage measured between SRP and SRN climbs and falls with the CC-readings.  So it seams the problem is somewhere on the BQ76940 side of the 100 ohm resistors.  I even replaced them 100 ohm resistors with 0 ohm resistors leading to no difference in behavior.  I can't help think the problem is coming from the BQ76940.  

This behavior has been observed on 3 separate circuit boards and chips.  I'm stumped and reaching out to see if you have any ideas.  Any thoughts?

Data below shows CC-readings and calculations during testing.

Micah

Discharge: Slowly ramp up discharge current (just above 4A), then rapidly dial down.

adc:  -2 v:  -1.688e-05 a:  -0.01688

adc:  -15 v:  -0.0001266 a:  -0.1266

adc:  -43 v:  -0.00036292 a:  -0.36292

adc:  -90 v:  -0.0007595999 a:  -0.7595999

adc:  -112 v:  -0.00094528 a:  -0.9452799

adc:  -122 v:  -0.00102968 a:  -1.02968

adc:  -139 v:  -0.00117316 a:  -1.17316

adc:  -157 v:  -0.00132508 a:  -1.32508

adc:  -175 v:  -0.001477 a:  -1.477

adc:  -193 v:  -0.00162892 a:  -1.62892

adc:  -207 v:  -0.00174708 a:  -1.74708

adc:  -207 v:  -0.00174708 a:  -1.74708

adc:  -207 v:  -0.00174708 a:  -1.74708

adc:  -207 v:  -0.00174708 a:  -1.74708

adc:  -288 v:  -0.00243072 a:  -2.43072

adc:  -327 v:  -0.00275988 a:  -2.75988

adc:  -412 v:  -0.00347728 a:  -3.47728

adc:  -494 v:  -0.00416936 a:  -4.16936 (max, begin dial down)

adc:  42 v:  0.00035448 a:  0.35448

adc:  -1 v:  -8.44e-06 a:  -0.00844

Charge:  Slowly ramp up charge current (to ~1.5A), them quickly dial down

adc:  -1 v:  -8.44e-06 a:  -0.00844

adc:  -208 v:  -0.00175552 a:  -1.75552

adc:  -207 v:  -0.00174708 a:  -1.74708

adc:  -39 v:  -0.00032916 a:  -0.32916

adc:  -2 v:  -1.688e-05 a:  -0.01688

adc:  19 v:  0.00016036 a:  0.16036

adc:  25 v:  0.000211 a:  0.211

adc:  40 v:  0.0003376 a:  0.3376

adc:  50 v:  0.000422 a:  0.422

adc:  60 v:  0.0005064 a:  0.5064

adc:  69 v:  0.00058236 a:  0.5823599

adc:  69 v:  0.00058236 a:  0.5823599

adc:  59 v:  0.00049796 a:  0.49796

adc:  54 v:  0.00045576 a:  0.45576

adc:  37 v:  0.00031228 a:  0.31228

adc:  37 v:  0.00031228 a:  0.31228

adc:  25 v:  0.000211 a:  0.211

adc:  17 v:  0.00014348 a:  0.14348

adc:  17 v:  0.00014348 a:  0.14348

adc:  12 v:  0.00010128 a:  0.10128

adc:  8 v:  6.752e-05 a:  0.06751999

adc:  6 v:  5.064e-05 a:  0.05064

adc:  2 v:  1.688e-05 a:  0.01688

adc:  2 v:  1.688e-05 a:  0.01688 (max - begin dial down)

adc:  29 v:  0.00024476 a:  0.24476

adc:  -2 v:  -1.688e-05 a:  -0.01688

Other info:

10S li-ion pack. 

42V charge voltage

All other features seem to function properly

  • 0
    TI__Guru* 81034 points
    Hi Micah,

    This is very strange behavior. The BQ76940 should be able to measure the charge current accurately without any issue and the current values are well within the range of the ADC. Can you send me your schematic and describe how you are forcing the current? Do you have a battery connected?

    Regards,
    Matt
  • 0 in reply to Matt Sunna
    Prodigy 90 points

    Matt,

    Sure.  Schematic should be attached.  You'll find it very similar to the reference circuit.

    The BMS is connected to a 10s1p pack of LG MJ1 cells.  This test pack is typically kept at around 3.6V.  PACK+/- are connected to a bench power supply set at 42V, and initially 0A.  Of course the voltages drops (just above battery V) as soon I dial up the current (0.1A at a time). 

    As mentioned, I've measured the current going from the power supply to the PACK+/- terminals, and I've measured the current between BATT- and negative terminal of the actual cells.  Both measurements match the power supply output, whereas the BQ76940 reports the dropping current as it rises above 0.6A.

    This morning I assembled a new board with a new chip and the same peculiar behavior is exhibited.

    Micah

    AirworthyFlexBMS20190509.pdf

  • 0 in reply to Micah Martin
    TI__Guru* 81034 points
    Hi Micah,

    I looked over the schematic and I don't see anything that explains this. I did notice something else that is not likely related. Since you are only using 10 cells but have components for 15 cells, the VC pins for the unused cells should be shorted at the device and at the cell terminals. I see the note on the schematic to short at the cell terminals, but do not see the shorts at the device. Are you shorting these connections using the test points?

    When the charge current goes above 600mA, do you see anything unusual in the status bits for the BQ76940 (any fault condition)? Do the CHG or DSG pins toggle on the BQ76940?

    I don't believe this could cause the issue, but have you considered using a larger sense resistor if your charge discharge currents are not large? It is best to utilize the range of the coulomb counter ADC (+/-200mV) to get the best accuracy.

    Regards,
    Matt
  • 0 in reply to Matt Sunna
    Prodigy 90 points
    Matt,

    Thanks for taking a look.

    Yeah, my schematic assumes 15S packs, but in reality I've been making 10, 13, 14, and 15 cell packs. So I have some 0 ohm resistors that I use instead of the 0.22uF capacitors and I solder a shorts between the unused cell leads.

    The BQ76940 remains perfectly content regardless of current. It does not throw alerts, nor do the CHG/DSG pins change state.

    Some of my applications will use 130+ Amps. It's just that on my bench I don't have the equipment to test large loads.

    The problem must be somewhere in my test setup. I'm pretty sure I saw the BMS showing high charge currents (~5A) in my skateboard (a toy project to educate myself). Unfortunately it broke and I'm waiting for parts to repair it.

    If you have any ideas, please let me know. Otherwise I'll check the skateboard when possible to see what difference could cause the behavior.

    Micah
  • 0 in reply to Micah Martin
    TI__Guru* 81034 points
    Sounds good Micah. Let me know when you have data from the skateboard. Or if you have any more information about your test setup, please feel free to send it and we can see if there are any other clues.

    I will close this thread for now, but it will automatically re-open when you reply with additional data.

    Thanks,
    Matt
  • 0 in reply to Matt Sunna
    Prodigy 90 points

    Matt,

    The parts came in and I repaired the skateboard.  I promptly plugged in the charger and the coulomb counter worked perfectly.  When I saw this, the problem became obvious.  It was my test setup.  The alligator clip wires I used to connect my batteries to the BMS are very thin, maybe 24 awg.  They were getting hot as I increased current and the positive temperature coefficient of the copper wires caused an increase in resistance.  Ohms law V=IR; with increase in resistance either voltage had to drop or current.  Since my bench power supply was driving the voltage, the current dropped. I doubled up on the wires and the coulomb counter readings were much better. 

    I appreciate your help and I'm sorry to have wasted your time on this embarrassing mistake.  Silly me.

    Micah

  • 0 in reply to Micah Martin
    TI__Guru* 81034 points
    Thanks for sharing your findings Micah. This definitely helps us the next time we need to debug a similar issue, so I appreciate you sharing this! I'm glad you're getting good readings now.

    Matt