BQ25790: Using TWO 25790s and controlling power sharing between them

Part Number: BQ25790

Tool/software:

I have a project that includes TWO power limited sources A and B, each capable of about 15 watts of power and each of variable voltage (one is not always higher than the other).

A single BQ25790 is really great at using ONE source of energy within a specified power limit to supply the load, and if it can't supply enough power then it will start to draw power from the battery.

For a DUAL BQ25790 design, I would tie the SYSTEM power and the BATTERY together for both chips. I then want the first chip to draw up to 15 watts from source A, if that isn't enough to supply the system load, then I want the second chip to draw energy from source B. If A+B do not have enough power for the load then I want the system to draw from the battery.

The only way I can think of doing this is by manipulating the BATTP voltage sense connection on the second chip to strip about 100mV off what it thinks the battery voltage is. In this way, Chip A will try to regulate VSYS at 200mV higher than the battery, while Chip B will think the battery is lower, and therefore would try to regulate VSYS at 100mV higher than the battery.

I think this will force A to power first, then B then Battery.

What I don't like about this solution is that the datasheet is not very definite about just how much higher VSYS is regulated above the battery voltage. Therefore, trying to play around inside that voltage zone seems risky, plus it depends on both chips having pretty tight voltage reading tolerance.

Can you suggest any other way to get two BQ25790 (or another TI chip) to play nice together? Making use of the power available to them, but in a well defined way of A then B then Battery...

NOTE: I know this part has a built in power source selector, but I want A plus B... not A or B.

I can't just use ideal diodes because I can't guarantee if A>B or A<B or A=B.

Thanks!

  • Hi Steve,

    If you have two separate batteries, this type of dual application with SYS pins connected is not recommended.   When either battery voltage > SYSMIN, the charger's internal BATFET turns on full.  So if one battery's voltage is below the other, the SYS of higher battery charger pushes unregulated current into the battery with lower voltage. 

    Even with a single battery tied to each charger's BAT/BATP pins, the internal reference variations cause VREG, ICHG and VSYS variations that result in one charger providing all of the power and/or one charger trying terminate before the other.  I wrote a load sharing app note many years ago that explains how to use ballast resistors to account for the variations https://www.ti.com/lit/an/slva250/slva250.pdf.  The relatively high SYS regulation accuracy and variation in BATFET resistance would require high value ballast resistors.

    BATP draws a few uA's of current in HiZ mode. Unfortunately, the max series BATP resistor is 100 ohm so the bottom resistor of resistor divider to ground would consume a lot of current. 

    Also many years ago, I wrote an appnote (https://www.ti.com/lit/pdf/slyt651) on paralleling 2 NVDC chargers by tying only their BAT pins together to one battery pack and using only one charger's SYS pin.  That means that the max supplement current to the system from the battery is limited to the max discharge current of a single charger.  For example, BQ25790 max discharge current, and therefore max system load current with no input power applied, is 6A.  If your max system load is <6A this might be an option.  

    You would also have to contend with both charger's having the same I2C address, which might require an I2C switch.

    All that said, one theoretical (meaning untested) design of 2 BQ25790's in parallel for up to 12A discharge current is 

    1. SYS pins connected by ideal diode OR plus series ballast resistors,

    2. BAT/BATP pins connected to single battery which VREG/ITERM of one charger set lower/higher than the other so they don't fight at termination

    3. I2C switch

    Regards,

    Jeff

  • A perfect response! Thanks so much.

    The second app note is really close to what I am trying to do, and indeed our SYS power is going to be well within the limits of a single part, so using one as charge-only is a good fit, with only a minor loss of efficiency when Source B is powering the output (since it has to pass through two BATTFET to reach the output.

    Cool!