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.

BQ2004h charge cycle

Other Parts Discussed in Thread: BQ2004E, BQ2004, BQ2004H

I am trouble shooting an older design using bq2004E. The  problem is that fast charging never starts except when the powered is recycled, othewise the batteries are always discharged. The INH is disabled (pulled high) and Vsense is grounded. So according the data sheets, the only other way to start the fast charge cycle is for Vcell to fall below max cell voltage. But by design Vcell is always less than MCV. This condition is always met even when battery chargeing is terminated. In other words when charging is complete, Vbat is less than MCV. And yet the new charge cycle is expected to start when Vbat falls less than MVC !. What am I missing here ?

I

 

 

  • I would look at P.3 of the datasheet at "Starting A Charge Cycle".  The chip needs to see a voltage on Vbat between Vedv and Vmcv to start the fast charge.  Additionally it needs a Vtemp between Vhtf and Vltf.  Confirm that you don't have any issues there.  If you're saying that a power cycle does allow a fast charge to happen then these two conditions appear to be met.  My concern is that you're getting stuck somehow in the "Top-Off and Pulse-Trickle Charge" circle in the flow chart of P.8.

    When the batteries aren't being charged, is there any trickle current at all?  It could hint to what state the chip thinks the battery pack is in.

    The other gotcha I can see is if a falling edge is detected at DCMD.  It could sit and wait for Vcell to fall below Vedv.  If you see absolutely no attempt to charge / no charge current this could be what's locking you up.  Take a look at your connection to DCMD.  The pin has an internal pull-up but it could be possible that noise externally could cause a low-going glitch.

  • Mike,

    Vtemp and Vbat are within the range and DCMD is externaly pulled high. It seems to me that the original designer has design this by fallowing the exact procedures outlined in the data sheet (the designed looks very much the same as the eval board). But here is my confusion. If you follow the design procedure outline in data sheet given on page 3, you will select RB1 and RB2 such that mathematically it will not be possible for Vbat to ever fall below Vmcv (it is required for Vbat to be always less than Vmcv for everything to work as expected). Yet the only way to start a new charge cycle (other than recycling Vcc, and INH), is for Vbat to fall below Vmcv.  

    The problem with this design is that if you ever replace the battery without recycling the power to the machine, bq2004 does not start a new charge cycle. At this point it seems to me the only way to fix it is to add my own detection and trigger circuitry to the design.

    thanks for the feedback.

  • The intent of "falling below MCV" to start a charge is as follows:
    MCV should never be reached in normal charging, voltage or temperature termination occurs before MCV.
    IF MCV is reached the charger should stop, high cell impedance would allow higher voltage and indicate bad cells.

    When the battery is taken OUT of the charger, a pullup resistor from DC IN to OUT is used to pull the output high, placing the BAT input above MCV.
    This tells the charger not to switch, basically stay in the fault state since no battery is installed.

    Finally when a battery is placed back IN the charger it will pull the OUT down to its cell voltage, the BAT input will fall below MCV, and a charge cycle will start.

    With a battery in the charger, and it reaches full charge, a new cycle will normally not begin again because TRICKLE current keeps the battery charged.
    You can start a new charge in this mode by resetting the IC by cycling its input power.

    Many designs we see do not have the pullup mechanism on the OUT.
    Determine the resistor divider impedance and how much current is required to pull it above MCV with no battery installed.
    Based on the Minimum input voltage, calculate R from input to out such that it will provide the current to pull BAT above MCV.
    Then you must remember to add this R current into the trickle current amount.
    Usually this is not much because the divider impedance is high and the added R is high also.

    See SLUU004 EVM, R13 is the added resistor designed to bypass the buck switch and inductor and pull up BAT+ when no battery is installed.
    In some cases this resistor can be the ONLY trickle current method needed to account for self discharge.

  • Ed,

    Everything makes sense now.

    Thank you.

    Babak.

  • I'm having a similar problem with the bq2004h chip.

    built a prototype using bq2004 chip. worked good. built 10 production boards using the bq2004H chip, and none of them will go into fast charge mode.

    led blinks "fast charge pending", then goes solid for "fast charge", yet current draw and scope say is sending series of pulses, same as "top off" mode from prototype. i.e, not going into fast charge mode, but staying in top off mode

    state diagram from datasheet says Vbat or Vtemp need to be in range before it will go into fast charge. same values as prototype, but maybe I screwed something up. nope, all is in range...???

    Vbat is ~2.5V. (2 cell voltage) as described. 

    VMCV = 0.8 ∗ VCC ± 30mV = 4V so ok there.

    VEDV = 0.4 ∗ VCC ± 30mV = 2V so ok there

    TCO is 1.25V (1.34V measured)

    TS is 1.76V

    VLTF = 0.4 ∗ VCC ± 30mV = 2V so ok there

    VHTF = [(1/3 ∗ VLTF) + (2/3 ∗ VTCO)] ± 30mV = 0.66V + 0.92 = 1.58V so ok there

    so both Vbat TS are in range, but it still doesn't go into fast charge mode.

    thoughts? suggestions?

    here is the schematic of the prototype:

    http://www.richardspelling.com/temppics/smartcharger3.pdf

    changes for production board:

    r4 & r14 -> 1.8k

    r3 & r13 -> 3.9k

    the current limit resistors for the leds are 2.7k instead of 10k

    q1,2,3,4 are slightly different model fets and transistors. (shouldn't affect anything, it's a state problem, not a switching problem)

    chip is bq2004H instead of bq2004

    none of the changes above should cause it to not go into fast charge mode. I'm about at the point of pulling the bq2004H off and soldering on a bq2004 to test.

    suggestions?