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.

BQ25792EVM: Charging problems with solar panel, load and battery

Part Number: BQ25792EVM
Other Parts Discussed in Thread: BQ25792, BQ25798, BQ24650

Our problem is with the BQ25792EVM. We are using it with a solar panel (12V) (with a 12V regurator), a 6800mA batteries and a system that consumes 80mA.


We placed it last week and we show that the panel were giving about 200mA and the battery was being charged since the STATUS led was ON. But then, at the end of the day (at the night), the led was OFF (normal). The load consumes the 80 mA during the 24 hours, all the time, so the battery was consuming by the load. The next day, when the light from sun was again in the panel, the BQ25792EVM did not start to charge the battery.... and 4 days after, the battery is empty, the load is off (becuase it has not power from battery) and the panel was not giving any current even when it was in under the sun. Then, I have took out one of the solar wires in the BQ25792EVK and I have connected it again. Then, it has start to charge again. WHAT HAPPEND?


We need it as a solar charger, I mean, charge the batttery if it is not totally FULL during the day


We have install our of our devices that send voltage to our platform depending on the solar current and during 4 days it has sent that the solar has not given any current.

Should we change any configuration in th BQ25792EVM or what should we do?

  • Hi Javier,

    The BQ25792 has poor source detection circuit that latches the device off if the input source is too weak, for example in low light for a solar panel.  I suspect that is the issue.  Before the battery crashes, can you check the poor source detection flag and EN_HIZ bit?  Details from the datasheet are below. 

    If a poor source is detected (when pulling IPOORSRC, the VBUS voltage drops below VPOORSRC), the device repeats poor source qualification routine every 2 seconds. After 7 consecutive failures, the device sets EN_HIZ = 1 and goes to HIZ mode. The device will remain in HIZ until either the adapter is re-plugged or the EN_HIZ bit is toggled, which will restart poor source detection with another 7 attempts. The EN_HIZ bit is cleared automatically when the adapter is plugged in.

    We have made a new version of BQ25792, BQ25798, which has an improved poor source detection circuit that doesn't latch.  Unfortunately, the BQ25798 ICs that are currently sampling on the TI website does not have the improved circuit yet.  We should be replacing the ICs and the EVMs in the next few weeks.

    Regards,

    Jeff 

  • Hello Jeff. Thanks for your answer. But, we have designed the new PCB with the BQ25792 chip and we could change it to BQ25798 that if I am not wrong it has the same pins and layout. But as we cannot wait weeks to solve this since we have to manufacture the 45 first units now since we have just 1 month to manufacture then first 45 units and then 4 months for the next 2000 units. So, we need a solution now. 

    Our solar chargen does not use any MCU, just the BQ25792 and caps, resistors, regulator, etc. So, we are using a L7812 regulator to gives always 12V even when the solar panel gives 20V. But when the sun is going low, then it will get 10V, 9V, 8V etc... and then what you have told me is what happen. So, now, to have a fast solution for this since we cannot write or read the I2C, do you have any idea about how could we make the regulator voltage go 0 when it get a voltage below 12V? I mean something like a zenner but not to cut the voltage if it is bigger, but cut the voltage if it is lower... I am thinking about how do this and then give 0V to the BQ25792 when the solar panel gives less than 12V and give 12V when it gives more than 12V and then for the BQ25792 it will be as if we unconnect the wires and then the EN_HIZ = 0, can you help me?

    This is so important for us. This is our bigger project we have for the next 2 years and if we do not find a solution (we have tested 7 different solar chargers and any of them works ok, and that is why we have started to design our own charger... anyway if you have any info about any solar charger for 3S that works ok... maybe we can use it also). But we have to find a solution this week.

    Best regards

  • Hi Javier,

    I don't think it is poor source problem.  The problem is likely the VINDPM feature, which autosets to a fixed voltage (0.7V and 1.4V) below the sensed VOC at power up.  If you power up with 12V, the VINDPM autosets to12V-1.4V=10.6V.  When the input source drops below that threshold, the VINDPM loop stops charge.  You would need to remove and replace the input voltage to the charger as it drops lower and lower or use I2C to write a new VINDPM level.  If you don't have I2C available, then I do not recommend BQ25792 or BQ25798.  The BQ25798 requires I2C to initially enable MPP feature, which periodically resets the VINDPM thresholds.  Did you look at BQ24650 controller with external FETs?  It has external resistors that sets the MPP (VINDPM) threshold.

    Regards,

    Jeff    

  • Hello Jeff. In  the information  about the BQ2579x it was written that it is also for solar chargers.... and it is not!!!! So, now we should change it by the BQ24650? Come on!!!!! Maybe the best is what I have told you. Use any circuit to get 12V or 0V. 12V when the voltage got is over 12V and 0 when the voltage got is below 12V. With this, the BQ25792 will work, right? Maybe for the future we can use the BQ24650 but now we have to solve this problem with the BQ25792 for the first units we have made and that we have to ship to our customer next week.  Let me know if you have any idea about how make that. Best regards

    PD. I have spoken with our battery supplier about this problem and they will ship us a battery protector for 4s. Per "S" the battery let pass just voltage if it is bigger than 3V. So, for 4s, it will just close the circuit with more than 12V. We will install it in the solar panel before the 12V regulator. So, if the solar panel gives more than 12V, the circuit will be closed and if the voltage is lower than 12V the circuit will be opened. And then, every new day, the Vin will get 12V again, and it will be as if we take out the solar wire and we connect it again. This should solve the problem at least with the BQ25792, right?

  • BQ25792 resets its VINDPM threshold after VBUS drops below and then rises above UVLO (~3.5V).  So, assuming that the input to the charger can provide more current than the poor source detection circuit looks for (30mA), then yes, what you propose about removing and reapplying power to the charger should work.

    Regards,

    Jeff

  • Ok Jeff. Tomorrow we will get a chip that sends high level if voltage is over 12V and low level if is below 12V. We will connect it to a 12V regulator with EN and then with the solar panel will give more than 12V it will be ON and we will get 12V and when the voltage from the solar panel will be below 12V then we will OFF the regulator and we should get 0V in the BQ input. I will check if this work ok during the next days.

  • Hello Jeff. We have made a test with a voltage supervisor and a regulator with EN. So, when the voltage is over 12V it actiavate the EN and when it below 12V it disactivate the regulator output. But we have seen that when the voltage is from 11 to 14V the RESET of the supervisor gives 0 to 3V and the regulator is actiavate from 2V so it does not work. Then I have a different question. The BQ25792 has a pin called CHARNING ENABLE. My question is if we connect the RESET of the supervisor to that pin and then we place as high when the voltage we can give to the BQ25792 is 12V (from the regulator), and a low level when we cannot give it, and then reset the BQ25792 charging enable pin. Would this work? I mean, can we also reset that pin to rest the EN_HIZ? Thanks for your attention.

  • Hi Javier,

    /CE pin only disables battery charge current from SYS to BAT.  The converter itself is not disabled.  If you want to disable the converter, you will need to pull ILIM_HIZ pin low.  Disabling the converter does not reset the VINDPM threshold to a lower value, which is what really needs to happen.  

    Regards,

    Jeff

  • Thanks Jeff. So, the unique thing we can do to make it work ok is place less than 3.7V in the VIN1 during the night and 12V (from 10.X to 13.X) in the VIN1 during the day (I mean during the time we need to charge the battery) or reset the EN_HIZ by I2C, right?

  • Javier,

    If VBUS drops below VBUS-UVLO and then rises above, the charger automatically goes through its input source checks including locking VINDPM register to the (hopefully) final voltage at power up less an offset. Because the panel voltage is likely rising as the sun comes up and falling as the sun goes down, the locked VINDPM value is preventing the charger from operating the panel at its MPP. Ideally, you would use the host software and I2C to periodically force the charger into HiZ, read VBUS using ADC, reset VINDPM value to 80% measured value and then exit HiZ mode.

    So, yes, reducing the input below VBUS_UVLO turns off the charger completely for lowest quiescent current consumption or using I2C to force EN_HIZ will do the same thing.  

    Regards,

    Jeff

  • Hello Jeff. We have decided to work with the BQ24650 (we will change the PCB). but we have some questions about it. We have seen the information from your website, and it seems that it just have in input for the solar panel and an output for the battery. In a post and if I am not wrong, in the EVM schematic, it seems that we can connect our load (device we should power from battery), in the VSYS in the EVM. But the, my question is: What voltage will be in the SYS when the solar panel is giving 21V? I suppose that the BQ24650 will give 14,6V to the battery to charge it, but then, will we get 14,6V in SYS or 21V? And then, during the night I suppose that the load will get the power from the battery. Can you tell me something now please, this is super urgent!!! Thanks

  • Javier,

    On the BQ24650, the system = input voltage so, yes the solar panel is tied to system.  When no solar panel is present, the battery FET turns on to connect the battery to the system.  You could also connect your system load directly to the battery. 

    Regards,

    Jeff

  • Hello Jeff. Thanks for our answer. Understood, but I suppose you know that we cannot connect the system directly to the battery, since if we make that, the system will consume X mA and then the Iterm will not finish never and the battery will be always charged even with it does not need and then the battery will be broken. So, with Lithium batteries you never should connect your device to the battery if you want to change the battery. We have found a charger in USA and we will then test it to select if we finish want to design our own charger or using a comercial charger. 

  • Hi Javier,

    I am sorry that the BQ25792 didn't work for you.  For it to work with a solar panel, it needs host software control.

    Regards,

    Jeff