• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 62 subscribers
  • Views 169 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.

BQ25185: BQ25185 (and BQ24210) current consumption at weak light

Gabriel
Prodigy 30 points
Part Number: BQ25185
Other Parts Discussed in Thread: BQ24210,

We're designing battery powered sensor devices for outdoor use for a client. These batteries are regular Lithium Ion batteries, and are recharged using solar panels. And to do that, we use battery chargers with VDPM support.

We have tested both the BQ24210 and the BQ25185, and have found that they both consume quite a lot of current from the batteries when the solar panel is outputting a high enough voltage to enable the charging, but cannot supply enough current. The BQ25185 consumes almost 200µA in this case. Through some design changes I have been able to get the BQ24210 down to about 40µA, but that is still more than we would like. This might sound like a transient occurrence when the sun is either rising or setting, but I'm confident that this can be triggered if the solar panel is lit by street lighting or similar. The winters are quite long and dark in the main market for these devices, and adding a drain of up to 200µA ends up being 144 mAh/month. If the device is installed north of the polar circle, it may not be able to charge at all for several months, both due to temperatures and lack of light.

Are there any ways to mitigate this, or any alternative charger IC:s we could look at?
I also have a quick tangential question regarding the BQ25185: How come there is no version of the IC with a lower fixed VSYS voltage? If the application can be powered directly from a single battery at 3.4-4.1V, it seems a little strange to have VSYS so much higher than the battery voltage. Many IC:s made specifically to be powered by lithium batteries have their maximum rated voltage at 4.3V, like the modem in our application. So if we use the BQ25185, we still have to connect the load directly to the battery.

I have attached images of the schematic diagrams. BAT_PG and BAT_CHG are externally pulled up to 3.3V, but the 3.3V rail was disabled during measurement.

image.png

image.png

  • 0
    TI__Genius 9235 points

    Hi Gabriel,

    Thanks for reaching out.

    What are the input voltage and battery voltage when you're measuring the 200uA from the battery? Could you share the I-V characteristics or curves of your solar panel? It sounds like you could be seeing a corner case where the device detects the input voltage as being "power good," but the input source is so current-limited that it cannot power many of the device's circuits that become active when the input is "power good," causing the battery to power them and resulting in the 200uA quiescent current from the battery that you're seeing.

    The VSYS regulation target is often higher than 4.2V in a linear power path topology to provide headroom for dropout due to the input current, charge current, and on-resistances of the input FET and BATFET.

    I recommend using a downstream DC-DC converter on SYS for this rail or connecting it directly to the battery, although I understand this wouldn't utilize the power path.

    Best regards,

    Alec

  • 0 in reply to Alec Lehman
    Prodigy 30 points

    Hi Alec!
    I've tried this with multiple battery voltages, but I think I used a lab PSU set to 3.8V when I did these specific tests. I don't actually have the curves for the specific solar panels readily available, but I can tell you that the open voltage is about 7V and the short circuit current is about 2A. This should put the MPP of the panel at about 5V or so at full illumination. I'll grant that it's a little higher than might be necessary, but still fairly normal. For my tests I simply had the solar panel exposed to regular indoor lighting. I modulated the available light by just placing my hand over the solar panel. If I raised the hand enough or removed it, it was able to charge a little. If I put it right over the panel, the voltage dropped enough to disable the circuit. But there was a fairly wide span where I saw the higher consumption from the battery.

    What you're suggesting is exactly what I assumed, that the open voltage is high enough to trigger power good, but that the panel can't deliver enough current to charge properly. But given that solar cells are inherently non-linear and that every implementation requires some voltage headroom, would this not more or less always be the case? Any time that the solar panel voltage is just above the threshold, adding any load to it will cause the voltage to drop. The hysteresis on the enable voltage doesn't seem high enough to account for this.

    Unless there's a workaround for this I think we're just going to keep using the BQ24210 as it has a lower consumption in this half-enabled state and doesn't have the unused power path functionality. I'll just add a jumper for the LEDs so that they can be disabled for devices out in the field, adding them as a load probably increases the illumination range where this can be an issue.

    Regarding the power path functionality: That does make sense, it's not a separate power path but tapping off the path before the battery charging stage. So there would need to be a headroom. In our case, adding a downstream DC-DC would add complexity, and an LDO would cost precious voltage headroom.

    Regards
    /Gabriel  

  • 0 in reply to Gabriel
    TI__Genius 9235 points

    Hi Gabriel,

    Solar panel irradiance is much lower with indoor lighting compared to direct sunlight. The I-V curve of a solar panel changes significantly with irradiance. Under weak irradiance, the I-V curve starts at a much lower short circuit current and is flatter.

    It seems likely that under the conditions you tested at, the I-V curve of the panel was very flat and at low currents, and the current supplied by the panel was less than the charger needs to power internal circuits when the input is considered "power good." I agree that this would be an issue in lower light conditions.

    Have you tested this outdoors when there is less sunlight, like in the evening?

    Also, have you measured the input voltage? Is the input voltage being regulated to VINDPM in indoor lighting?

    Best regards,

    Alec

  • 0 in reply to Alec Lehman
    Prodigy 30 points

    Hi again!
    I measured the voltage at the input, and it seems to track for the BQ25185. It's in battery tracking mode, and the input voltage stays about 200mV above the battery voltage even when it's consuming current from the battery. The BQ24210 on the other hand seems to basically keep the input even with the battery voltage, which seems a little odd. We could use a set VDPM instead of using battery tracking, but I think that would only move the issue rather than fixing it. And it might cause the device to charge less when light is available.

    Generally sunlight seems to keep it charging, even ambient light. But when the sun is rising or setting, or when it's dark and the panel is being illuminated by street lighting, this can still happen. This might be considered an edge case normally, but for this product it is normal operating conditions during winter.

    Regards
    /Gabriel

  • 0 in reply to Gabriel
    TI__Genius 9235 points

    Hi Gabriel,

    Thanks for checking the input voltage. That's what I expected.

    Battery tracking VINDPM is well suited for use with solar panel inputs, but there are also different approaches using a host MCU if more control is needed.

    I think this issue is more related to the solar panel's capabilities rather than VINDPM. Since the I-V curve is very flat under low irradiance conditions, the amount of current the panel can provide does not change much with panel voltage.

    I agree that low light is a normal operating condition for this application. What is the size of the solar panel you are using? Have you tried with a different panel to see if the same behavior still occurs? It sounds like this may be a very small panel if it is not able to provide at least a few mA.

    I recommend these app notes for more on using solar panels with battery chargers:

    Using Battery Tracking VINDPM to Make the Most of Solar and High-Impedance Power Sources

    Choosing the Correct Solar Battery Charger for Your Solar Application

    Best regards,

    Alec