• State Resolved
  • Locked Locked
  • Replies 29 replies
  • Answers 12 answers
  • Subscribers 64 subscribers
  • Views 5954 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.

Original question:

BQ24650: Solar panel charger circuit needed for 6V, 12V and 24V solar to charge the lithium battery 3.7V

BQ24650: Power path design using BQ24650

Emilija Zukauskaite
Expert 1100 points
Part Number: BQ24650
Other Parts Discussed in Thread: BQ25895, BQ25703A, BQ25606, TINA-TI, TPS63070

Hi,

BQ24650 doesn't provide a power path but I would like to try and design a power path circuit using PFETs to be able to supply the power to the system even when the battery is not connected or low. Can you please suggest the design? Is it possible to prioritise the system load over a charger circuit? If there isn't enough current to supply the load and charge the battery I would like to use all available current to supply the load rather than charge the battery. 

Thanks!

Emilija

  • 0
    TI__Genius 10240 points

    Hi Emilija,

    The dynamic power management (DPM) of a charger priorities the system load over battery charging. It requires changing internal control loops. Implementation with external circuit would be extremely difficult. Adding a PFET for power path is possible but the timing control of the FET to prevent shoot through during the power path switching could be complex. TI has host controlled chargers with power path and DPM that can be used for solar charging. This is a blog talking about the application e2e.ti.com/.../optimizing-solar-power-with-battery-chargers.

  • 0 in reply to ezhao
    Expert 1100 points
    Hi,

    I'm struggling to find the IC that works without a microprocessor. Do you think bq25895 could work? It has a default charging current 2A. What if my solar panel can only provide a few hundred mA, will it still work? Will it adjust automatically without a power loss?

    Thanks!

    Regards,
    Emilija
  • 0 in reply to Emilija Zukauskaite
    Expert 1100 points
    Will bq25895 MPPT still work if I place a dc-dc converter in front of the charging circuit? Our solar panel might go up to 14.5V or even higher.

    Or can I use bq25703A without a microprocessor?

    Thanks!
  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points
    Hey Emilija,

    Neither of those devices will work to perform true MPPT without a host controller + algorithm. The closest device you will find will likely be the one this post was originally about, the BQ24650.

    That MPPT pin will in reality give you a fixed voltage operating point for your solar panel. So for low power panels or individual solar cells, this would be a fair compromise. You could set the operating point to fixed % of the VOC of the panel/cells.


    Regards,
    Joel H
  • 0 in reply to Joel Hernandez II
    Expert 1100 points
    Hi Joel,

    I cant use BQ24650 because there is no power path and I need it work without a battery. Unless I design a power path circuit myself which I'm not sure how. Do you have any ideas?

    I'm still thinking to use bq25895 because the datasheet says that it has a maximum power tracking and there is an application note for this part that its using a solar panel input. Do you think it will work? It doesn't have to be super efficient for our design. Is it okay to place a dc-dc in front of bq25895?

    Any other suggestions please?

    Thanks.

    Regards,
    Emilija
  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points
    Hey Emilija,

    Power path can easily be added if your system can handle two different voltage levels (battery and input). If would only take a P-Ch. FET to implement this between your system (connected to VBUS) and BAT. However, if you want a regulated system voltage that tracks with the battery voltage as it does in the BQ25895, then there is not.

    There is a reference design using a solar input to the BQ25895 however for MPPT tracking it requires the MCU host algorithm. There is also the VINDPM threshold which will automatically select and operating voltage either 600mV or 1.2V lower than your panel's VOC.

    Lastly, yes you can place a regulated DC/DC as the input of this charger.


    Regards,
    Joel H
  • 0 in reply to Joel Hernandez II
    Expert 1100 points
    Hi Joel,

    I have a dc-dc converter after the charging circuit so therefore the different voltage levels are acceptable.

    Can you please advice on the PFET power path design? I'm not sure how to design it but I would like to try using BQ24650 with an external power path design.

    If I use BQ25895 without a host controller, what would be an operating voltage? I would have a 5V input coming from the dc-dc before the charging circuit.

    Thank you!
    Regards,
    Emilija
  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points

    Hey Emilija,

    For the power-path, you could implement something like the following. You would need to tweak the threshold voltages and capacitor value to add some turn on delay as well as select the appropriate P-Ch FET. Remember here that the voltage will either be the battery voltage or a diode drop below the input voltage. 

    As far you question about the BQ25895, I am a little confused as it seems you are putting a regulated DC/DC in front of the charger? Will the front-end DC/DC provide power point tracking for you? As far as the charger, the default VINDPM voltage will be 5V - 600mV = 4.2V if the unloaded output voltage of the DC/DC is indeed 5V. This means the charger will allow it's input voltage to crash to 4.2V and then subsequently regulate at that threshold.

    Regards,

    Joel H

  • 0 in reply to Joel Hernandez II
    Expert 1100 points

    Hi Joel,

    Please see my schematic idea below. I would like to replace diodes by PFETs to reduce the power loss.

    Do you think that could work? 

    Why do I need to add a delay to switch to a battery power supply?

    As far you question about the BQ25895, I am a little confused as it seems you are putting a regulated DC/DC in front of the charger?

    -- I need to use a dc-dc because my input voltage can be up to 14.5V and the maximum input votlage of BQ25895 is 14V.

    Will the front-end DC/DC provide power point tracking for you? As far as the charger, the default VINDPM voltage will be 5V - 600mV = 4.2V if the unloaded output voltage of the DC/DC is indeed 5V. This means the charger will allow it's input voltage to crash to 4.2V and then subsequently regulate at that threshold.

    --- Can I increase the VINDPM voltage to 4.8V for example? Would I need a host controller?

    My prefer option is to use BQ24650 with a power path design. I just need help to design it.

    Thank you very much!

    Regards,

    Emilija

  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points
    Hey Emilija,

    That input PFET configuration will not work. You will effectively keep that FET always on as long as you have voltage present on either side. This means the battery will shorted to your solar panel. This could trip battery pack protection and/or feed current from your battery to the panel if it is in a poor condition/heavily shaded.

    I would recommend adding another PFET in the reverse direction with a small control NFET to drive the gates to ground when the solar panel input is present. And you are correct. I would actually add the delay circuit to this NFET to delay the turn-on of the input PFETS to give the "BATFET" PFET time to fully turn off.

    As far as the back-to-back FETs you drew up, I would just use a single PFET with its body diode in the direction of VSYS from VBAT (effectively Q1 only in your circuit). This will achieve power-path.


    In regards to the BQ25895, you will need a host controller to re-program the VINDPM threshold from the default value of VOC - 600mV.


    Regards,
    Joel H
  • 0 in reply to Joel Hernandez II
    Expert 1100 points

    Hi Joel,

    Sorry for the delay, I was off work.

    I used Q2 instead of the diode from BQ24650 D1 for a reverse protection.

    I added another PFET Q4 and NFET to turn Q4 and Q2 on. Is that what you meant? Can you please explain a bit more?

    How long I need a delay for? I dont want the system to loose power and turn off if the the power path switches between the solar panel input and the battery. I'm worried that the solar panel input will be unstable and will keep switching between the battery and the solar input.

    "As far as the back-to-back FETs you drew up, I would just use a single PFET with its body diode in the direction of VSYS from VBAT (effectively Q1 only in your circuit). This will achieve power-path. "

    Did you mean the to use a single FET with its internal diode? Will it give a reverse polarity protection?

    Is it okay to connect IN1 and IN2 to the points of the BQ24650 circuit as I drew it earlier?

    Please review my schematic. I'm not sure if it will work.

    Thank you!!

  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points

    Hey Emilija,

    You are close to what I was suggesting. For reference, this is a modification to the circuit you attached:

    The 1uF at the input and 40uF + 1A load at the output is an example of what is part of the required charging circuitry. The rest is similar to what you gave but adjusting the drive of the BATFET (Q1 in your circuit). 

    As far as the solar panel and battery interaction: We only want the delay long enough to allow the BATFET to be turned off before the input protection FETs turn fully on. You will have to play around with the capacitance and resistor value in a SPICE simulator to get a feel for how much voltage drop you will see based on a particular delay you implement. The voltage droop and any oscillation will depend on several factors including:

    1) System load current and output capacitance

    2) Charging current

    3) Delay based on RC and voltages applied to the control NFET and the P-Ch. BATFET. 

    4) FET threshold voltage

    5) The charger's programmable MPPT pin voltage setpoint

    6) Resistor divider ratios and lowest operating voltage of the panel

    This is not a complete list but it will get you started. Again, the main reason for the delay is so your input does not source as much current as possible into the battery.

    Regards,

    Joel H

  • 0 in reply to Joel Hernandez II
    Expert 1100 points

    Hi Joel,

    Thank you! I will try simulating and building a PCB. Is it correct to replace D1 from the typical applciation of BQ24650 by T3 and T2? 

    Basically connect T3 left to D1 anode and T2 right side to D1 cathode?

    I'm still not sure how two PFETS common source allows the power to pass through? How can power go from the source to drain of T2?

    Regards,

    Emilija

  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points
    Hey Emilija,

    I ran a quick SPICE sim on this circuit and replacing the diode with this is fine. Because you have that control NFET (T4).

    One thing that is missing in my circuit is a pull-up between VCMSRC and the shared gate connections of both T2 and T3. You will want to add a weak pull-up here.

    In this configuration, the body diode cathodes of the PFETs T2 and T3 face each other towards VCMSRC.

    As far as power transfer, when the input power is applied, the T4 NFET will be turned on and pull down on the gates of PFETs to ground, well below the common source voltage. With no input power, the common source voltage will be a diode drop below VSYS. With input applied, before the FETs turn fully on, the common source voltage will be a diode drop below the input voltage VIN.


    Regards,
    Joel H
  • 0 in reply to Joel Hernandez II
    Expert 1100 points
    Hi Joel,

    Do I need to add a resistor between VCMSRS and gate of T2 and T3?

    What is VCMSRC and where does it have to be connected?

    Sorry, I still dont understand how power goes from source to drain of T2 even though T2 is turned of? T2 diode will be reversed biased when power flows from VIN to VSYS?

    Thanks!

    Regards,
    Emilija
  • 0 in reply to Emilija Zukauskaite
    Expert 1100 points
    Also I forgot that I need to add a dc-dc converter from 12V to 5V from Vin to Vsys when the the system is powered up by the solar panel. Should I have it before T3 and T2 or after?

    Maybe there is another way to make BQ24650 to provide power output when the battery is not connected?

    Thanks and Regards,
    Emilija
  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points
    Hey Emilija,

    A resistor is the pull-up I was referring to in my previous reply.

    VCMSRC was just a simulation node in my SPICE simulator. It does not have to be physically connected anywhere else.

    When VIN is not valid and very low, both T2 and T3 are off. The body diode of T2 will conduct and show a diode drop below VSYS at the VCMSRC node.
    When VIN is valid, both T2 and T3 are turned on as their gates are pulled to ground. VSYS will see VIN. T1 should be turned off by this point.

    And the reason we are using the BQ24650 is so you don't have to use an extra DC/DC at the input. If you will do that, there is no point in using this charger and instead go with the BQ25895.


    Regards,
    Joel H
  • 0 in reply to Joel Hernandez II
    Expert 1100 points
    Hi Joel,

    Do I need a pull up resistor from VCMSRS to VIN and from both gates to VIN? Not between them? Why?

    I simulated two PMOS transistors connected the same way and the voltage doesn't pass through the second PMOS. I used LT spice. I will try again.

    I would still like to use BQ24650 because it has MPPT and doesnt require MCU. I actually printed and build the charger PCB using BQ25606
    and it works well, I get 75% efficiency. I have a step down dc-dc before the charger and a step-up after the charger. But a problem is when I limit the power supply current by half (simulating solar panel if it doesnt get enough light) and the efficiency drops to 57% because I guess the charger is trying to pull more current than its possible because it has a fixed current.

    I was thinking to use BQ24650 with a power path. A step down will only be active when the solar panel is supplying power to the system and a step up only be active when the battery is supplying the system. What do you think?

    Thanks.

    Regards,
    Emilija
  • 0 in reply to Emilija Zukauskaite
    Expert 1100 points

    This is what I'm trying to design. I also need to make sure that the system Vout is prioritised over the charging IC is there is not enough current.

    Please my schematic below. I don't think I need Q2.

    Thanks.

    Regards,

    Emilija

  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points

    Hey Emilijia,

    I just tested this on both TINA-TI and LTSpice and the result was the same as what I stated before. Make sure your spice model is utilizing a real PMOS instead of an ideal model without a body diode. Also check your simulation conditions. The voltage thresholds have to be set such that Q5 will turn on sufficiently and Q1 will turn off sooner than Q4 and Q2 turn on.

    The back-to-back input PFETs are needed so you do not short the solar panel and the battery. Right now, Q7 in your circuit has no function because it is always pulled to ground. Effectively, the battery can push current to the input with this always on PFET.

    I would also remove R18 or at least put it in series with the "Solar" and the Q1 gate.

    As far as your comments about using the BQ24650 for the MPPT and the extra DC/DC converter. Since you are looking for a fixed 5V rail, I would recommend using the BQ24650 plus a buck-boost DC/DC converter such as the TPS63070. In this case, you are looking at a model like this:

    Regards,
    Joel H

  • 0 in reply to Joel Hernandez II
    Expert 1100 points

    Hi Joel,

    I thought I replied but I cant see my message from today. 

    The design you suggested looks great! I will design a PCB and build it.

    I've just got a few questions. Where do connected points IN1 and IN2? I dont really need Q7 if I have two back to back FETs before. Do I just short IN1 and IN2? Or do I connect left side of the first FET to IN1 and right side of the second FET to IN2?

    Also I'm not sure what voltage should I set as MPPSET? I read that I should set to 80% of the VIN. But I have two solar panels I might be using either 6V or 12V. Do I set 80% of 6V?

    Thank you for your support!

    Regards,

    Emilija

  • 0 in reply to Joel Hernandez II
    Expert 1100 points
    Hi Joel,

    I was browsing other questions about this part and I found a few suggested answers that its possible to to remove battery detection function. by connecting schottky diode between SRN and the resistor feedback network. Or another answer suggested adding 5mF cap to trick the part that battery is connected. Maybe this way would be easier than designing a power path? What do you think?

    Thanks!

    Regards,
    Emilija
  • 0 in reply to Joel Hernandez II
    Expert 1100 points

    Hi Joel,

     

    Should I add a diode to prevent the leakage current going back to the charger as suggested in this document? http://www.tij.co.jp/jp/lit/an/slva829/slva829.pdf

    The document states that just having a diode and no extra capacitor will remove a battery detection function. Maybe that would be a good solution for my application? I could connect the load to the battery and I wouldn’t need a power path circuit? Would it work? Or will it have issues as the battery will be always in charging mode because the load will draw current and that means that life of the battery will be reduced?

     

    Also I found this power path design document:

    http://www.ti.com/lit/ug/sluu410/sluu410.pdf

    It looks more complicated than my design and it’s using a PG pin or STAT2. Do you think I need to take anything from this circuit?

    I would like to keep it simple, but I just wanted to check with you if anything’s missing.

    Please see my schematic below. Let me know if there is anything wrong.

    Thanks.

     

    Regards,

    Emilija

  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points
    Hey Emilija,

    If you are using the configuration I recommended in my previous reply, the left back to back FET will be connected to IN1 and IN2 will be connected to the right FET.

    As far as MPPSET, you should base this value on either the 6V or the 12V. Not both. 80% is typically a decent estimate of the voltage @ MPP of the solar panel. I would also suggest looking at the solar panel datasheet and look for the peak power point voltage (usually on an I-V curve given).


    Regards,
    Joel H
  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points
    This will not implement the power path you are looking for. Disabling battery detection just makes it look like the battery is always there. Realistically, you want to be able to switch to the input source even if you have a battery present. And when you don't have an input source, the battery is the power source.


    Regards,
    Joel H
  • 0 in reply to Emilija Zukauskaite
    TI__Mastermind 25800 points
    The document you attached was not a power path design document; it was the user's guide for this family of device's EVM.

    Several of the chargers in this family (BQ246xx) have a power selector, utilizing both a BATDRV and ACDRV pin to switch between input sources to SYS.

    From your schematic, I would recommend you check the resistor divider ratios and capacitance values to make sure it works across your voltage range. What I had given previous were example values based on a specific voltage. You will want to simulate across the full battery voltage range and the solar input voltage range, and see if there are any corner cases that could yield shoot-through. You will also want to make sure your simulation is utilizing real FET models that have their own specifications. But architecturally, it looks fine. Just tweaks will be needed.


    Regards,
    Joel H
  • 0 in reply to Joel Hernandez II
    Expert 1100 points

    Hi Joel,

    Thank you for your comments.

    I'm not sure about MPPT voltage setting. We require to use either 6V or 12V solar panel. If I set MPPT voltage to 5V will it still work with 12V solar panel? What will happen if a 12V panel voltage drops to like 10V. Will the charger adjust the charging current?

    Regards,

    Emilija

  • 0 in reply to Emilija Zukauskaite
    TI__Genius 10240 points
    Hi Emilija,

    You want to check the solar panel specification to get the quantified answer. General speaking, the MPPT is around the 80% of the solar panel open circuit voltage. The %V setting might be far away from the 12V panel MPP.
  • 0 in reply to ezhao
    Expert 1100 points
    Thank you for your help, really appreciate. I will try and test it.