• State Resolved
  • Locked Locked
  • Replies 17 replies
  • Subscribers 65 subscribers
  • Views 4912 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.

BQ25890EVM-664: STAT output of BQ25890EVM does not match documentation?

John Seaber
Prodigy 130 points
Part Number: BQ25890EVM-664
Other Parts Discussed in Thread: BQ24075, BQ25890, BQ25892, BQSTUDIO, EV2400

I've used BQ24075 with success for years in many designs, and am currently testing BQ25890 using TI's BQ25890EVM-664 evaluation board. I planned to begin by testing at a typical USB rate, 500mA, before setting I2C parameters for faster charging rates, but even the default charging rate produced unexpected results.

Configuration:

  • All default BQ25890EVM parameters are set.
  • Resistance at ILIM measures 800 ohms, corresponding to a current limit of 487 mA. 
  • Load: Connected to J3 (SYS and GND)
  • Input: micro USB connected at USB IN
  • Battery: Sparkfun 2000mAH Li-Ion connected at TP5 (BAT) and PGND

Picture of Setup

From the EVM schematic and datasheet pg 30, I expected:

  • LED D3 (STAT) should be active while charging, and turn off when charging completes, or flash at 1Hz upon error such as exceeding the 12hr safety timer
  • LED D4 (PG) should be active while the micro USB cable is connected to a 5VDC USB charger or hub.

Results:

  • D4 (PG) is always off *WRONG*
  • D3 (stat) is always on, even after 12 hours *WRONG*
  • The battery charges at 4.1VDC *OKAY*
  • D3 later began flashing very rapidly. Please see video here. Note the camera is limited to 30fps and is unable to capture the high rate of toggling. My eyes estimate 30-60Hz, but I've not yet measured. Regardless, D3 should not blink!

Debugging Observations

  • Measured VUSB = 5.16VDC
  • Measured VBAT = 4.13VDC
  • Tried another USB charging source to rule out current limiting
  • Verified AGND is common to GND
  • D4: Checked EVM schematic, and noticed jumper JP3 is configured for USB D+/- detection, so D4 is left floating and will not work.
  • Battery protection datasheet lists 'overcharge protection discover voltage': 4.130V max, while BQ25890EVM lists max voltage applied as 4.208V

So, I think the battery protection board conflicts with the BQ25890's default charging parameters (table 5, pg 26 of datasheet). I will try other batteries next.

Regardless, I'm still confused by the unusual cycling behavior at the STAT pin. Can anyone explain what's happening at STAT? Shouldn't it cycle at 1Hz to indicate an error? I 

  • 0
    TI__Guru**** 166176 points
    John,

    bq25890 does not have /PG pin. The D- pin replaces /PG in the bq25890 compared to bq25892.

    Regarding STAT, can you send a scope shot of V(SYS), V(BAT), V(STAT) and I(BAT)? It is possible that the charger is oscillating because the battery OVP protector has opened and the charger thinks the battery is no longer connected.
  • 0 in reply to Jeff F
    Prodigy 130 points
    Thanks, Jeff. I agree this battery protection circuit is unsuitable for the BQ25890. The oscillation at VSTAT is still unexpected. Here are a few quick scope shots:

    V(STAT) is oscillating at about 20Hz:
    http://i.imgur.com/hsWRO4q.png

    Similar behavior is found at V(BAT):
    http://i.imgur.com/SCmr9MK.png

    as well as at V(SYS):
    http://i.imgur.com/wzCxrEl.png

    I(BAT) is oscillating from 132mA to 142mA.
  • 0 in reply to John Seaber
    TI__Guru**** 166176 points
    John,

    Do you have current meter in series between BAT pin and the battery? If so and you remove it and lower VBATREG below the 4.13V OVP of the battery, does the IC work as expected?
  • 0 in reply to Jeff F
    Prodigy 130 points

    Thanks, Jeff.

    I've tested with and without a current meter in series with the BAT pan and battery (same oscillation either way). I have not had a chance to lower VBATREG today, but this is a valid solution to escape the issue. This afternoon, I'm testing a complete charge cycle of a different LiPo cell, whose protection circuit can accommodate 4.2VDC charge voltage (okay so far).

    I'll report back after reducing VBAT with the original battery. Regardless, the BQ25890 has failed to detect an error, so safety of this circuit relies entirely on the cell's own protection circuit.

  • 0 in reply to John Seaber
    Prodigy 130 points

    Update:

    Using a TI EV2300 to monitor all registers, the BQ25890 eval board is indeed oscillating in and out of charge mode. I've tried reducing VBATREG, and have also tested two different batteries. I've compared our LiPo cell and protection circuitry datasheets to ensure all registers are properly configured, but the BQ25890 oscillates.

    Jeff - Are you able to recommend any LiPo cells known to operate well with BQ25890 and default parameters? I'd like to start with a known working configuration to make sure the eval board is okay, then proceed with further investigation of charger oscillation.

  • 0 in reply to John Seaber
    TI__Guru**** 166176 points

    John,

    We have never had issues with any LiPo battery that is rated for 4.2V maximum charge.  Does this battery have a new, cutting edge protector? 

  • 0 in reply to Jeff F
    Prodigy 130 points

    Jeff,

    Thanks for your quick replies! The batteries I've tried have been on the market for years, so they are by no means "cutting edge". Here are the exact parts under test; I tried two different LP803860's:

    PKCell LP803860 -  Lithium Ion 3.7V, 2000mAH - Datasheet 

    UnionFortune 585460 - 3.7V LiPo, 2000mAH - Cell Datasheet, Protection Circuit Spec

    We also stock 1300mAh Li-Ion cells, which I can try next if you're unable to recommend a cell for evaluation. I do find this strange.

  • 0 in reply to John Seaber
    Prodigy 70 points

    I facing the same oscillation (which also triggers the INT). Did you find any solution/WA for this behavior?

    /Mikael 

  • 0 in reply to Mikael Rogberg
    Prodigy 130 points

    Hi Mikael  - Unfortunately, this remains unresolved. I delegated the project to another engineer and he's observing the same behavior as well.

  • 0 in reply to John Seaber
    Prodigy 70 points

    I change the 'Charged Voltage Limit' in REG06 to better match my battery which first seamed to do the trick but when reaching the limit the oscillation was back. I then added a check on REG0b when getting the interrupt from the chip. I then noticed that CHARG_STAT is set to "Charge Termination Done". When reading this in REG0b I now disable charging and everything is fine. But I find this solution a bit strange, this means that the BQ25890 will not "work" as stand alone. Or is this a defined/wanted behavior?

    /Mikael 

  • 0 in reply to Mikael Rogberg
    Prodigy 130 points
    - Can you chime in?
  • 0 in reply to John Seaber
    TI__Guru**** 166176 points
    Mikeal,

    Is the IC on the EVM bq25890 or bq25892? JP3 should be configured properly for each EVM. Is not then I need to follow up with my EVM vendor.

    What is the input current and charge current that you are measuring? If you flip the S2A and S2B switches related to USB D+/D- lines and then remove and replace USB, does the current increase?

    When you say the charger charges to 4.1V, do you mean the voltage measured at the EVM board BAT test point TP5 (which is more or less Kelvin to the IC BAT pin) measures 4.1V and the charger stops charging at that voltage?

    Where is the TS voltage? Are you connecting to an external thermistor? If so did you remove JP10 and install JP9?

    Do you have an EV2400 and bqStudio software to read the register values?
  • 0 in reply to Jeff F
    Prodigy 70 points

    Hi,

    I have a BQ25890 and the JP3 is set to DSEL, which I think is correct(?).

    Currently my setup and battery is more ore less fully charged so I leave it like that and skip measuring the charge current right now since I don't think that is what we are looking for here.

    I have no external TS and JP9 and JP10 is as shipped == ON.

    But, when the battery is full and I insert the USB I get an interrupt. Ignoring this will result in "never ending" interrupts at approx. 40 ms interval.

    If I instead when getting the interrupt read REG0B and if it flags 'Charge Termination Done' disable charging by writing REG03 I don't get the repeated interrupts/oscillation: This is fine (disabling the charging on interrupt) but is this what I'm supposed to do?

    I don't have EV2400 but register readout is as below register address i first column after rest I also set the REG_RST to reset the chip

    00 48
    01 06
    02 DD
    03 10
    04 20
    05 13
    06 62
    07 8D
    08 03
    09 44
    0A 73
    0B 02
    0C 00
    0D 14
    0E 4E
    0F 4E
    10 51
    11 00
    12 00
    13 08
    14 1D

    irq_bms_cb                 // INT : USB inserted 
    B02                             // REG0B = 0x02
    irq_bms_cb
    B62                             // REG0B = 0x62 -> USB DCP input
    irq_bms_cb
    B66                             // REG0B = 0x66 -> USB DCP input | power good
    irq_bms_cb
    B7E                             // REG0B = 0x7E -> USB DCP input | power good | charge termination done
    charge disabled

    -------------------------------------

    If I disable the lines that disables the charging by clearing bit 4 in REG03 I get the log below

    00 48
    01 06
    02 DD
    03 10
    04 20
    05 13
    06 62
    07 8D
    08 03
    09 44
    0A 73
    0B 02
    0C 00
    0D 14
    0E 4F
    0F 4F
    10 51
    11 00
    12 00
    13 08
    14 1D

    irq_bms_cb
    B02
    irq_bms_cb
    B62
    irq_bms_cb
    B66
    irq_bms_cb
    B7E
    irq_bms_cb
    B7E
    irq_bms_cb
    B7E
    irq_bms_cb
    B7E
    irq_bms_cb
    B7E
    irq_bms_cb
    B7E
    irq_bms_cb
    B7E
    irq_bms_cb
    B7E
    ... etc ...

  • 0 in reply to Mikael Rogberg
    TI__Guru**** 166176 points
    Mikeal,

    To summarize, you attach a battery that is only at 4.1xV but the IC keeps toggling in and out of termination at 40 ms interval? With charge enabled, I expect the IC to do this if the battery at V(BAT) were much closer to your regulation voltage but at much longer intervals. The charger automatically starts a recharge cycle when the voltage at BAT drops below the recharge threshold (100mV in your case). Are you certain that your connection from the BAT pin to the battery is not highly resistive (for example through a current meter)? I have seen this type of oscillation in this case. Can you capture a scope plot of the V(BAT) showing this oscillation? Just curious if there a current pulse that pushes the battery voltage up to regulation? I can't explain why the voltage drops back down so quickly.
  • 0 in reply to Jeff F
    Prodigy 130 points

    FYI - My observations match Mikael's, and no current meters were used.

  • 0 in reply to John Seaber
    Prodigy 70 points

    Hi,

    I think I have found, if not "The" problem at least my problem.

    I shortened the wires to between the EVB and the battery to a minimum and also stepped up in wire area. Same problem. I measured the drop between battery + pole and BAT terminal on the EVB and got ~190 mV when charging. Hmmm... I then increased the Vreg_chg to 200 mV and the oscillation was gone. 

    Then I realized that we have a fuse in line with each battery so I measured the drop over the fuse and got almost all of the 190 mV. I then replaced the fuse with a wire on one board and the drop was gone. Measuring the fuse I get 0.60 ohm, which with a charge current of, say 250 mA, will result in 150 mV. 

    Since we plan to have 4-8 batteries in parallel we added the fuse as a protection so that any battery that for some reason gets really bad is to be disabled. I don't know if this is a good approach? Obviously we can't have the fuse, or if we have we will have to take care of this problem by disabling charging in the software or find a different fuse.

    John, does my findings spread any light on your side?

    /Mikael 

  • 0 in reply to Mikael Rogberg
    Prodigy 130 points

    Mikael - Interesting findings. I used a single cell in each experiment. Each battery evaluated was equipped with its own protection module (no fuse, but maybe a PTC).

    I will share these notes with our team and will report back when possible (BQ25890EVM no longer in my possession).