• State TI Thinks Resolved
  • Locked Locked
  • Replies 15 replies
  • Answers 7 answers
  • Subscribers 63 subscribers
  • Views 1111 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.

BQ25792: Charge current breakdown at currents higher 4A

Oliver Bredow
Prodigy 20 points
Part Number: BQ25792

Good day sir or madam,

Im currently testing the BQ25792. I have build a 4x4cm testing PCB following the guidelines in the BQ25792 datasheet with the same Inductivity as the EVM. Power is supplied by a external 300W power adapter. When discharging, I discharge with around 2.7A from a 10Ah battery with Max-discharge-current 5A. I now run into two problems that I just do not understand. Im hoping you could help me:

1) The way I understand Vsysmin is that Vsys is kept at Vsysmin (default 3.5V) as long as possible. If that is not possible anymore, the BQ25792 switches off the System voltage. Do I understand that correctly? Because Vsys In my tests always follows Vbat. Even when Vbat is below 3.5V it is still discharging until Vbat=2.8V, where the batteries PCM stops the charging. Even the lifted Vsys (compared to Vbat) that is mentioned in the NVDC part of the datasheet I do not notice. VSys mostly equals Vbat.

2)  Our battery is supposed to be loaded with 5A. Setting the charge current up to 4A works fine. However setting it to 4.5A (with a supplied VBUS= 15V and IBUS_max=5A) makes the input current IBUS slowly break down. Looking into Register 33h confirms, that the charge current never reaches 4.5 A, but rather decrements slowly to around 3A. Register 03 still contains my input of 01C2h (for 450). Battery voltage according to 3Bh is around 3.7V, which I confirmed by multimeter. No fault is displayed. Even when decreasing the charge current now to something like 4A (03h->0190h) the current keeps dropping. I can not make sense of this behaviour. Do you have an idea what is going on and could help?

Thanks and regards,

Oliver BRedow

  • 0
    TI__Mastermind 40305 points

    Hi Oliver,

    Please make sure the RMS continuous discharge current does not exceed the maximum in the specification Section 8.3 (6A).

    1) The SYS and BAT are connected by the BATFET, so when the BATFET is fully on they will have similar voltages. This is best explained by Section 9.3.8.1 and Section 9.3.8.2. The charger will always regulate the SYS to SYSYMIN if the battery is lower than SYSMIN. When the battery is over SYSMIN, the battery voltage is slightly lower than system voltage. In the case of input overloading, If the SYS voltage drops below SYSMIN, charge current will be reduced to keep SYS at SYSMIN. If the charge current is reduced to zero and the input is still overloaded, SYS will drop. When SYS is less than BAT the battery will start discharging.

    2) What is your input voltage and your battery voltage at this time? Can you send me a Register Sweep with the ADC enabled? The maximum input current is actually 3.3 A and I am wondering if you are being power/thermally limited.

    Regards,

    Mike Emanuel

  • 0 in reply to Michael Emanuel35
    Prodigy 20 points

    Hi Michael,

    thanks for the quick response. As written above im discharging with a 1.3Ohm Resistance which leads to a discharge current of arround 2.7A,  far beneath 6A.

    1) Ok. Assuming I do not have a charger connected, is VSYSMIN then always = VBAT? Because, as i wrote above, the way I understand VSYSMIN is that (when the batterie is purely discharging and no adapter is connected) the discharging process should be stopped at Vsysmin. Meaning that, when Vsysmin can no longer be maintained, batterie discharging is switched off entirely. Do I understand that correctly? (It is not clear to me from the datasheet whether your description of VSYS behavior applies always or only when a power adapter is connected and IBUS is drawn).

    2) Input voltage has been set to 15V at this point. At a charging current IBAT of 4A, a current IBUS of around 1.3A is drawn according to my power adapter. Batterie voltage measured at batterie by multimeter shows a voltage of around 3.6V to 3.8V. VBAT measured by Reg 3Bh shows a VBAT of around 4V (By the way, is there supposed to be such a difference between the Register value and the actual Batterie-Voltage?). So, as far as i can tell, im still solidly in the constant-current area. I do not get any sort of error messaging (no LED blinking) and all the status regs(1Bh - 25h) are good too. No sensor temperature problem or such is being noted.

    I have attached the eagle files for my testing board. Maybe I have made a mistake here? Is it maybe a bad idea to have a shunt resistor in my batterie-path? BQ25792_klein.zip

    Thanks a lot and regards,

    Oliver Bredow

    BQ25792_klein.zip

  • 0 in reply to Oliver Bredow
    TI__Mastermind 40305 points

    Hi Oliver,

    1) I think there is a misunderstanding of VSYSMIN. When the battery voltage is less than VSYSMIN (with adapter present), the SYS voltage is regulated to VSYSMIN. As the battery voltage reaches VSYS, the VSYS voltage goes up with the battery voltage, albeit slightly higher due to the drop across the BATFET. When the battery is purely discharging, VSYS will always be less than the battery voltage due to the drop across the BATFET. There is no normal operation that will stop the battery from discharging if SYS is less than SYSMIN. It will continue to discharge until either the load is removed, the ShipFET intervenes, or a sufficient input is supplied.

    2) What does the actual BAT pin measure during charging? To fully diagnose the problem I will need to see the full registers. It seems there may be too much impedance from your BAT pin to your battery. Can you measure this impedance as well?

    Regards,

    Mike Emanuel

  • 0 in reply to Michael Emanuel35
    Prodigy 20 points

    Hi Mike,

    1) I see. Thanks for the clarification. As a suggestion for sometime in the future, could you maybe update the datasheet so that it more clearly states that the mentioned behavior applies only when charging happens? The way i understood it I thought it applied all the time.

    2) I have attached a sweep of all registers for a charge current of 2A and 4.5A each. Impedance from Bat-pin(Bat_1 in my eagle) to Batterie+(Bat+ in my eagle) is around 50mOhm. I have noticed that, as soon as 4.5A Charge current is set, the measured Batterie Voltage jumps up to around 4.1V, while the actual batterie voltage jumps up to 3.9V. Before switching on that charge current, batterie voltage was 3.5V. That would be a jump of 0.4V from just switching on the charge current. Is this kind of batterie voltage jump normal? I know that batterie voltages jump when charging starts but this seems a bit high?

    Additionally I have noticed that when I set the charge current to 4.5A while Bat-pin voltage (Reg 3B) is above 4V, I get all kinds of error messages like System-Overvoltage and Batterieovercurrent, with the supply current frequently breaking down.

    Otherwise, when Batterievoltage is below 4V (Reg 3B) there are no error messages. Its just that the set 4.5A charge current are never applied to the Batterie. Instead charge current starts with around 4A and then slowly drops to around 3A.

    Regards,

    Oliver Bredow

    BQ25792_Regs.xlsx

  • 0 in reply to Oliver Bredow
    TI__Mastermind 40305 points

    Hi Oliver,

    If you have 50 mOhm of impedance and you apply 4.5 A charge current, that is 0.225 V of drop across the connection. There must be more impedance in the path on the order of around 40 mOhm.

    Can you please provide some documentation of Sysem-Overvoltage etc.? Maybe some waveforms and register settings when this happens?

    I am running my EVM at 15 VBUS, 4 VBAT, and I have 4.50 A charge current. I am at 88C on the die temperature, so not being thermally limited.

    Regards,

    Mike Emanuel

  • 0 in reply to Michael Emanuel35
    Prodigy 20 points

    Hi Michael,

    you were correct, my plug that I attach the batterie to seems to add around 40mOhms for a total of around 90mOhms in the path. Additionally, i have updated the register overview for the  aforementioned error case. Maybe you can find something in there.

    5734.BQ25792_Regs.xlsx

    Regards,

    Oliver Bredow

  • 0 in reply to Oliver Bredow
    TI__Mastermind 40305 points

    Hi Oliver,

    Unfortunately the registers are not complete. It says you have 0x01 for REG0x01 which is not a valid register setting. Can you please double check your register settings? REG0x19 also reads 0x01, which is not valid.

    In the meantime we need to isolate what is happening with the battery terminal connection. Can you decrease the resistance to the battery connection and see if this problem persists? Where is the BATP pin sensing? If it is too close to the charger with this much resistance you could get a situation where the charger enables charge, realizes full charge because of the voltage increase, then shuts off in a cycle. Can you try placing the BATP resistor closer to the actual battery via wire?

    Regards,

    Mike Emanuel

  • 0 in reply to Michael Emanuel35
    Prodigy 20 points

    Hi Mike,

    Unfortunately during my last test, the BQ25792 got fried somehow. I will have to rebuild the board. Last test I did was that I discharged my batterie to 2.8V, set the BQ25792 to IBAT=4.5A and waited until the precharge phase was done. Precharge phase was done well but as soon as it switched to CC phase current broke down again. This time I got a 07h in Reg 1B (power not good). Since I might have to redo my board, could you please check my design and the components and tell me whether I made a mistake there somehow? I have already sent you the eagle files, attached you now find the BOM for full context.BQ25792_order.csv

    Thanks and Regards,

    Oliver Bredow

  • 0 in reply to Oliver Bredow
    TI__Mastermind 40305 points

    Hi Oliver,

    My schematic review is as follows:

    1. The STAT pin is tested with sink current of 5 mA. With 420 Ohm resistor, you are at roughly 12 mA when conducting. Please lower STAT resistor as this could damage open drain FET.

    2. Can you please explain the selection for ILIM_HIZ resistors? I calculate the max input current at 3.6 A which is unachievable.

    3. Does the TS pin connect to thermistor?

    4. Confirming 1s 1.5 MHz operation.

    5. R14 is not needed.

    Unfortunately the BOM is not formatted correctly so I cannot read it. Please separate out columns with designator, part number, value etc.

    My layout review is as follows:

    1. Typically we route SDA and SCL next to each other. Although not necessarily a problem, it wouldn't hurt to fix this.

    2. I repeat my comment about the BATP placement for battery sensing. Can you try placing the BATP resistor closer to the actual battery?

    3. Can you send me the gerbers for your project? When I import the layout it makes it look like a bunch of the nets are shorted together which is wrong.

    4. The REGN capacitor could be closer.

    5. Please refer to our layout for the power stage (VBUS, PMID, SW1, SW, BAT). There is only one via to GND in the region to C1 and C13. In our layout we have 3 vias and the ground plane is directly connected to the rest of the plane. The present state is a mostly isolated island. See Section 12.2 Layout Example.

    Regards,

    Mike Emanuel

  • 0 in reply to Michael Emanuel35
    Prodigy 20 points

    Hi Mike,

    Thanks for your suggestions. Here my answers:

    Schematic:

    1) According to the BQ25792 Datasheet, the REGN output supplies around 4.8V-5V with a max current of 30mA. The Stat Pin can sink 6mA. I am using the SMLMN2BCTT86C as my LED. It requires 2.9V at 5mA. This means around 2.1V have to drop at the resistor which leads to a to a value of 420Ohm (at 5mA). I do not quite understand where the 12mA you are suggesting are coming from considering there should always be a voltage drop (forward voltage) over the LED?

    2) It seems I made a mistake there.

    3) Yes, it does.

    4) What would be the best way of doing this?

    The BOM is separated by ;;;. Setting Excel csv delimiters to ; does the trick. Designators and such are already included.

    Layout:

    3) BQ25792_klein_gerber.zip

    Thanks alot and regards,

    Oliver Bredow

  • 0 in reply to Oliver Bredow
    TI__Mastermind 40305 points

    Hi Oliver,

    Schematic:

    1. As I had trouble reading the BOM, I did not know the forward drop of the diode. Please limit the STAT sink current to 5 mA. 

    4. I verified the PROG pin pulldown for 1s 1.5 MHz operation.

    BOM:

    1. On our EVM we recommend rating the 0.1 uF capacitors on PMID, VBUS, and SYS for 50 V. I see the design is for 25 V. These capacitors are very important as shown in the Layout Guidelines. Please test according to your application.

    2. We specifically specify that with no ShipFET 50 V, 0402, 1nF capacitor is required on SDRV. I see yours is rated for 25 V. This needs to be updated.

    3. On our EVM we recommend terminating PROG pin with 100 mW rated resistor. Please test according to your application.

    4. Please make sure your inductor is rated (specifically saturation) for your application as shown in 10.2.2.1 Inductor Selection. You only intend to use one of them correct? From what I can see the input voltage is 15 V and I will assume a battery voltage of 4 V with a charge current of 5 A. This gives a ripple of almost 2 A, meaning the saturation current rating is at a minimum of 6 A.

    5. No need for shunt R14.

    Layout:

    I only see three layers, I seem to be missing the bottom layer. Can you please provide this?

    Regards,

    Mike Emanuel

  • 0 in reply to Michael Emanuel35
    Prodigy 20 points

    Hi Mike,

    Schematic:

    1) I limited the current the way I described it above. However, i Am wondering why the resistor in the EVM is chosen to be this big? It is 2.2k. That seems to be a bit overkill to me.

    BOM:

    4) I originally set up two testboards, one with the smaller Inductance, one with the bigger one that is also used on the BQ25792EVM. The same faulty behavior showed up with both of them.

    Layout:

    -It seems something went wrong during my first gerber creation with eagle. I have generated the files again, now the bottom should be there as well. 2148.BQ25792_klein_gerber.zip

    Regards and thansk for your help,

    Oliver Bredow

  • 0 in reply to Oliver Bredow
    TI__Mastermind 40305 points

    Hi Oliver,

    The resistor is what we recommend and test with on the EVM.

    I apologize for any inconvenience, but I need the overlay and silkscreen as well in order to evaluate this layout. I need to see where the components are placed. Looking at the most recent I only received the layer plots.

    Regards,

    Mike Emanuel

  • 0 in reply to Michael Emanuel35
    Prodigy 20 points

    Hi Mike,

    I see that. However, I am wondering why such a big resistor was chosen. I would think that the current is reduced to around 1mA this way. That seems to be very little to me.

    No Problem, guess i should have send you everything to begin with. Anyway, this should be everything now: 4846.Gerber.zip

    Regards,

    Oliver Bredow

  • 0 in reply to Oliver Bredow
    TI__Mastermind 40305 points

    Hello Oliver,

    Here are my comments on the layout:

    1. Please route SDA and SCL differentially. They separate after the IC.

    2. Please route BATP closer to battery positive terminal. 

    3. In our EVM we have more vias underneath the pad for the SWx traces to minimize the resistance. Please consider this.

    4. I believe the PMID and SYS 0.1 uF could be closer to improve performance.

    Regards,

    Mike Emanuel