BQ24075: Questions regarding circuit design for 18650 battery charger

Hi Fernando,

Fernando Fontes said:

I'm using the AP63205WU buck to regulate the input voltage from the solar panel to 5V and then sourcing it to the BQ24075. I've noticed already that probably instead of connecting BQ_EN2 to +5V I should have connected it instead to VOUT_BAT. But besides of that, anything else that pop's out? 

I've reviewed your schematic and it looks good. I agree that it would be better to connect EN2 to OUT if you intend for EN2 to always be pulled high. Other than that, I don't see any issues.

Since you're using a solar panel input, I recommend considering an alternative approach: connect the solar panel directly to the charger and utilize the charger's input voltage dynamic power management (VIN-DPM) feature.

VIN-DPM is active when EN2=0, EN1=0 or EN2=0, EN1=1. This feature prevents the input voltage from falling below the VIN-DPM threshold (4.5V) by dynamically reducing the input current limit, which is particularly useful when using high-impedance input sources like solar panels.

Fernando Fontes said:

nCHG is going low level even without the battery connected. However, activating SYSOFF by swathing SW1 makes it high level again. Is this the expected behavior? I'm bypassing BAT pin with a 4.7uF capacitor.

Refer to Table 9-2 in the BQ2407x datasheet. The /CHG output is expected to be high-impedance when the battery is disconnected.

Could you provide more details about this behavior? Is /CHG flashing or staying low?

Fernando Fontes said:

One last question, the BQ24075 will not protect the battery right? I still need to use a OVP and UVP IC right?

The BQ24075 does not include battery OVP or UVP protections. I recommend using a battery protector IC to implement these protections.

I also recommend checking out the BQ25185. It's our newest standalone linear battery charger and includes battery-tracking VIN-DPM and offers several integrated protections, such as battery undervoltage protection (BUVLO).

Additionally, here's a useful article: Using Battery Tracking VINDPM to Make the Most of Solar and High-Impedance Power Sources

Best regards,

Alec

Hi Alec,

Thank you for the reply.

Alec Lehman said:

Since you're using a solar panel input, I recommend considering an alternative approach: connect the solar panel directly to the charger and utilize the charger's input voltage dynamic power management (VIN-DPM) feature.

VIN-DPM is active when EN2=0, EN1=0 or EN2=0, EN1=1. This feature prevents the input voltage from falling below the VIN-DPM threshold (4.5V) by dynamically reducing the input current limit, which is particularly useful when using high-impedance input sources like solar panels.

So, I thought that ENx pins were being used for limiting the input current. My idea was to have the maximum possible current on the output even without a battery connected, since I will also want to use this gateway indoors, connected directly to a 12V transformer. In this case, you are saying that the DPM feature will not work? And the battery will continue to be charged normally (at 1A) even if the system voltage is dropping? 

Also, related to your suggestion, the solar panel that I'm planning to use is rated to 12V but can go up to 18V. From the datasheet, the operating voltage range is mentioned to be from 4.35V to 6.4V. So, I guess that the battery would not be charged until the voltage of the panel dropped for such range right? That's why, I've used the 5V regulator, to be able to use 12V solar panels and still charge the battery all the time. Since I also want the gateway to work indoors without the battery, I've picked the '05 version to have the SYSOFF pin.

Any suggestion that could work better here? Is the BQ25185 a better candidate? The gateway itself will drive a 4G simcom module while also powering some external sensors.

Alec Lehman said:

Refer to Table 9-2 in the BQ2407x datasheet. The /CHG output is expected to be high-impedance when the battery is disconnected.

Could you provide more details about this behavior? Is /CHG flashing or staying low?

I will connect the pins to the oscilloscope today and share some more insights.

Alec Lehman said:

The BQ24075 does not include battery OVP or UVP protections. I recommend using a battery protector IC to implement these protections.

I also recommend checking out the BQ25185. It's our newest standalone linear battery charger and includes battery-tracking VIN-DPM and offers several integrated protections, such as battery undervoltage protection (BUVLO).

Ok, thanks. I already have them. Was just to be sure that I was not adding extra components without the need. 

Thank you for the ongoing support,

Best,

Fernando

Alec Lehman said:

Refer to Table 9-2 in the BQ2407x datasheet. The /CHG output is expected to be high-impedance when the battery is disconnected.

Could you provide more details about this behavior? Is /CHG flashing or staying low?

Ok, so I was saying it the wrong way. When SYSOFF is connected to 5V the /CHG goes low. At least the LED stays solid green. And with SYSOFF high impedance the /CHG is high impedance too (LED off).

However, I've switched my attention to another issue which is, with the battery in place and with external 12V connected, so 5V on the regulator and supplying the BQ24075 the /CHG goes low as intended and the battery starts to charge. /PGOOD is also low. Then, turning off the power supply, /PGOOD goes high impedance as intended but /CHG becomes a square wave, which makes the LED on. Check the following oscilloscope image:

What could be the issue? I didn't find anything on the datasheet. I saw something related to preventing the device from charing with 50Hz half-rectified signals, but not this. 

I've checked the input voltage and it seems that somehow the 5V line is not going to 0V, is it possible that BQ_EN2 is somehow pull-up internally and is making the 5V line go battery level? 

I only see that current path that could be making this happening... Unless the input capacitor of 220uF is not being discharged, but I guess BQ24075 would try to charge the battery and discharge it no?

I will continue to investigate more tomorrow. I only had little time today to investigate this.

Thank you,

Best regards,

Fernando Fontes

Hi Fernando,

Thanks for the updates.

Fernando Fontes said:

So, I thought that ENx pins were being used for limiting the input current. My idea was to have the maximum possible current on the output even without a battery connected, since I will also want to use this gateway indoors, connected directly to a 12V transformer. In this case, you are saying that the DPM feature will not work? And the battery will continue to be charged normally (at 1A) even if the system voltage is dropping? 

Yes, you're correct that the ENx pins are used to configure the input current limit mode. Just to clarify, VINDPM and DPPM are different features.

VINDPM (input voltage dynamic power management) prevents the input voltage from dropping below the VIN-DPM threshold (4.5V) by dynamically reducing the input current limit. This is useful when using weak USB sources or solar panels. On the BQ24075, this feature is only active for the previously mentioned ENx modes. See section 9.3.4.1.1 of the datasheet for more details.

DPPM (dynamic power path managament) reduces the charging current when the system voltage falls below the VDPPM threshold due to the sum of the charging and system currents exceeding the input current limit. See sections 9.3.4.1.2 and 9.3.4.1.3 of the datasheet for more details.

Fernando Fontes said:

Also, related to your suggestion, the solar panel that I'm planning to use is rated to 12V but can go up to 18V. From the datasheet, the operating voltage range is mentioned to be from 4.35V to 6.4V. So, I guess that the battery would not be charged until the voltage of the panel dropped for such range right? That's why, I've used the 5V regulator, to be able to use 12V solar panels and still charge the battery all the time. Since I also want the gateway to work indoors without the battery, I've picked the '05 version to have the SYSOFF pin.

Yes, you are correct. The BQ24075 input overvoltage protection is 6.6V (typical), so the battery won't charge when the input voltage is higher than this threshold. In that case, using the 5V regulator is a solid approach.

The BQ25185 is our newest standalone linear charger and supports input voltages up to 18V, so you wouldn't need the 5V regulator.

Fernando Fontes said:

Ok, so I was saying it the wrong way. When SYSOFF is connected to 5V the /CHG goes low. At least the LED stays solid green. And with SYSOFF high impedance the /CHG is high impedance too (LED off).

Thanks for clarifying the /CHG behavior. This sounds expected. Refer to section 9.3.5.5 of the datasheet: when SYSOFF is high, /CHG is low.

Fernando Fontes said:

I've checked the input voltage and it seems that somehow the 5V line is not going to 0V, is it possible that BQ_EN2 is somehow pull-up internally and is making the 5V line go battery level? 

EN2 is internally pulled down. See the functional block diagram of the charger in section 9.2 of the datasheet. Could you monitor the input voltage on one of your scope channels?

Best regards,

Alec

Alec Lehman said:

Yes, you're correct that the ENx pins are used to configure the input current limit mode. Just to clarify, VINDPM and DPPM are different features.

Ok, thank you.

Alec Lehman said:

The BQ25185 is our newest standalone linear charger and supports input voltages up to 18V, so you wouldn't need the 5V regulator.

Ok, I might change then to the BQ25185. Is there any equivalent SYSOFF option? Since I want the gateway to work both indoors and outdoors (charger / solar) it would be interesting to have such feature, so that I can have the device working without battery. But I guess that, SYSOFF is more actually to have the battery placed and not charging it right? So, BQ25185 without the battery placed should work just fine, I guess.

Alec Lehman said:

Thanks for clarifying the /CHG behavior. This sounds expected. Refer to section 9.3.5.5 of the datasheet: when SYSOFF is high, /CHG is low.

Ok, so that make sense.

Alec Lehman said:

EN2 is internally pulled down. See the functional block diagram of the charger in section 9.2 of the datasheet. Could you monitor the input voltage on one of your scope channels?

This is really strange. So, Channel 2 (blue) is connected to the 5V output pin of the regulator and Channel 1 (yellow) to the \CHG pin. Figure 1 is with external power connected and Figure 2 without external power.

Trough the ripple that I see on Figure 2, I really think is some sort of feedback that is happening from the battery voltage. I've placed a 10K resistor in parallel with 5V and GND and nothing happened. However, a 120Ohm resistor makes the 5V line go 0V and the \CHG high impedance. 

This is how the power supply looks like:

Initially I was thinking that the feedback was thought the 5V line but I think now, it's instead from the VOUT_BAT line. I've a configuration of the TPS55165QPWPRQ1 (which I found out later that IGN > 5V so I will replace it and use other part) for boosting the voltage when running solely on battery to be able to power external sensors. The feedback might be happening from the Q1. I don't see why, since D4 is there to protect reverse voltage. I will try to un-solder it and repeat the tests....

Thank you for the ongoing support,

Best regards,

Hi Fernando,

I'm happy to help.

Fernando Fontes said:

Ok, I might change then to the BQ25185. Is there any equivalent SYSOFF option? Since I want the gateway to work both indoors and outdoors (charger / solar) it would be interesting to have such feature, so that I can have the device working without battery. But I guess that, SYSOFF is more actually to have the battery placed and not charging it right? So, BQ25185 without the battery placed should work just fine, I guess.

Yes, you're correct. The input source will be able to supply the system load even when no battery is connected.

If you need to disable charging while the battery is connected, you can use the /CE pin.

The BQ25185 also has factory mode (similar to ship mode on our other chargers), which is an ultra-low-power mode where the battery is disconnected from the system. This mode can only be entered when no input source is present and is exited automatically once an input is connected.

Fernando Fontes said:

This is really strange. So, Channel 2 (blue) is connected to the 5V output pin of the regulator and Channel 1 (yellow) to the \CHG pin. Figure 1 is with external power connected and Figure 2 without external power.

Trough the ripple that I see on Figure 2, I really think is some sort of feedback that is happening from the battery voltage. I've placed a 10K resistor in parallel with 5V and GND and nothing happened. However, a 120Ohm resistor makes the 5V line go 0V and the \CHG high impedance. 

After taking another look at your charger schematic, I realized that SYSOFF is connected to +5V through a switch.

SYSOFF is internally pulled up to VBAT through a large resistor (~5MΩ), which could be related to the behavior you're seeing. Is the switch closed when you see this behavior? From the second waveform, it looks like +5V is around 4V, which I expect is your battery voltage.

I don't see any obvious issues with the circuitry around Q1 in your buck-boost schematic.

Let me know if you have any updates.

Best regards,

Alec

Alec Lehman said:

Yes, you're correct. The input source will be able to supply the system load even when no battery is connected.

If you need to disable charging while the battery is connected, you can use the /CE pin.

The BQ25185 also has factory mode (similar to ship mode on our other chargers), which is an ultra-low-power mode where the battery is disconnected from the system. This mode can only be entered when no input source is present and is exited automatically once an input is connected.

I will change the schematic and then I will ask your revision if that is ok.

Alec Lehman said:

SYSOFF is internally pulled up to VBAT through a large resistor (~5MΩ), which could be related to the behavior you're seeing. Is the switch closed when you see this behavior? From the second waveform, it looks like +5V is around 4V, which I expect is your battery voltage.

I don't see any obvious issues with the circuitry around Q1 in your buck-boost schematic.

Let me know if you have any updates.

The switch was open... Today the issue was happening but only for a few seconds. I've removed Q1 just to be sure and I've also connected BQ_EN2 to VOUT_BAT. R16 was also removed. The only connection to the 5V line is now the IN pin (number 13). 

I've attached to videos, one that shows the /CHG led and another one that shows the 5V line. As you can see, when I turn off the external power supply the 5V line immediately drops to the battery voltage ~4V. Then it takes ~3 seconds and drops to almost 0V. At that point the LED goes off. 

Is this intended behaviour? I guess that in the diagram block of the BQ2407x Q1 is probably being activated? 

Thank you for the ongoing support,

Best regards,

Hi Fernando,

Fernando Fontes said:

I will change the schematic and then I will ask your revision if that is ok.

Yes, feel free to share your updated schematic and I will review it.

Fernando Fontes said:

The switch was open... Today the issue was happening but only for a few seconds. I've removed Q1 just to be sure and I've also connected BQ_EN2 to VOUT_BAT. R16 was also removed. The only connection to the 5V line is now the IN pin (number 13). 

I've attached to videos, one that shows the /CHG led and another one that shows the 5V line. As you can see, when I turn off the external power supply the 5V line immediately drops to the battery voltage ~4V. Then it takes ~3 seconds and drops to almost 0V. At that point the LED goes off. 

Is this intended behaviour? I guess that in the diagram block of the BQ2407x Q1 is probably being activated? 

The input FET (Q1, simplified in the functional block diagram) blocks reverse current from the output to the input, so I wouldn't expect the battery to be charging the input capacitance.

That said, it seems likely that the input caps are discharging too slowly after the input source is removed, which could be causing the behavior you're seeing.

You could try placing a resistor from VIN to GND to help them discharge more quickly and see if that changes the behavior. This can also help rule out the possibility that the battery is somehow back-powering VIN.

It appears that your battery is not connected when this behavior occurs. Are you powering BAT using an external supply?

Best regards,

Alec

Hi Alec,

Alec Lehman said:

You could try placing a resistor from VIN to GND to help them discharge more quickly and see if that changes the behavior. This can also help rule out the possibility that the battery is somehow back-powering VIN.

Any recommended value? 1K?

Alec Lehman said:

It appears that your battery is not connected when this behavior occurs. Are you powering BAT using an external supply?

Battery was being connected using the orange and green cables that were soldered to the bottom of the PCB. I was testing like that to make the process of removing and plugging back the battery more easily. 

I will perform the resistor test tomorrow and I will let you know.

Best regards,

Fernando Fontes

Hi Fernando,

You could try using a 100k or any value high enough to avoid significant current draw but low enough to help drain the residual charge from the IN caps more quickly.

Let me know how the test goes.

Best regards,

Alec

Hi Alec,

I was able to run a few more tests with the suggested resistor. 

Alec Lehman said:

You could try using a 100k or any value high enough to avoid significant current draw but low enough to help drain the residual charge from the IN caps more quickly.

100K resistor will not help. I've attempted 1K and also 200Ohm and the problem still persists. Only a 100Ohm resistor solves the problem. However, having 100Ohm will drain to much current.

When this issue happens I see a lot of ripple on the VIN pin of the BQ (external power supply is off).

Yellow line is the 5V / Vin of BQ and blue line is the +12V line input. They have the same voltage so this means that the AP63205WU is letting the voltage to go in reverse way. To confirm this, I've un-soldered the DC-DC AP63205WU and connected directly the +12V input to the +5V output. I've adjusted external power supply to 7V (because I've a full bridge rectifier) and yes, I've confirmed that indeed the behaviour is the same. 

With the 100Ohm patch when I turn off the power supply the 5V line goes immediately to 0V. 

Without the 100Ohm and without the DC-DC buck the 5V line goes to +BATT level. 

Today, the tests behaved similarly to the initial ones, where the +5V line was maintaining permanently to +BATT voltage even when the external power supply was off. The only difference in the setup is that, unlike the other day, I didn’t enable sleep mode. Back then, the average current consumption was just a few milliamps (< 5mA). This time, I left the cellular modem active, which significantly increased current draw ( ~25mA) —apparently making the issue more noticeable. I'm still not sure what's causing this behavior.

Thank you,

Best regards,

Fernando Fontes

Hi Fernando, 

Thank you for your detailed response. We are reviewing the new information and will get back to you with comment. Please be aware tomorrow, 4/18, is a company holiday here is US so please expect a follow-up response from TI on Monday. Thank you in advance for your patience. 

Best Regards,

Garrett 

Hi Fernando,

Thanks for your patience. I'll provide a response by end of day.

Best regards,
Alec

Hi Fernando,

Thanks for all of the information.

I expect that this issue is specific to your board. The BQ24075 blocks reverse current flow from the battery to the input, so I wouldn't expect the input to be directly connected to the battery internally through the charger.

Could you confirm the repeatability of this issue? Does VIN go to battery voltage every time the input source is disconnected? Also, could you clarify the exact issue you're seeing? It sounds like the input now stays at battery voltage indefinitely, but you previously mentioned it would drop to 0V after a few seconds.

Are you able to completely disconnect everything on the input path that's unrelated to the charger IC and power it directly? I understand there are several upstream components on VIN, but isolating the charger would help narrow down what's causing this behavior.

Does this issue happen on multiple boards or just one? If you only have one, could you try replacing the charger IC to see if that resolves the issue?

Lastly, I'd recommend letting the battery discharge (safely) or testing with a battery at a lower voltage. This would help confirm whether VIN is truly tracking VBAT, or if it's just a coincidence that they're both around 4V.

Best regards,

Alec

Hi Alec,

Thanks for reply. 

Alec Lehman said:

Could you confirm the repeatability of this issue? Does VIN go to battery voltage every time the input source is disconnected? Also, could you clarify the exact issue you're seeing? It sounds like the input now stays at battery voltage indefinitely, but you previously mentioned it would drop to 0V after a few seconds

Yes, I can confirm the repeatability. I think the behaviour is different depending on the load conditions of the OUT pin. With some load it does seem that it stays like that indefinitely. Previously, I was doing the tests with the MCU in deep sleep mode which was making it behave differently. 

Alec Lehman said:

Are you able to completely disconnect everything on the input path that's unrelated to the charger IC and power it directly? I understand there are several upstream components on VIN, but isolating the charger would help narrow down what's causing this behavior.

Yes, I kinda did that already, but I'm sending you privatly the full schematic so that you can take a look and tell me what to disconnect. I can cut traces if needed, no problem with that. 

Alec Lehman said:

Does this issue happen on multiple boards or just one? If you only have one, could you try replacing the charger IC to see if that resolves the issue?

I have two boards and the issue happens on both.

Alec Lehman said:

Lastly, I'd recommend letting the battery discharge (safely) or testing with a battery at a lower voltage. This would help confirm whether VIN is truly tracking VBAT, or if it's just a coincidence that they're both around 4V.

I will do that.

I think I will switch for the version without the System off pin since I don't really need it. I though it was required for working without a battery attached and that's why I've picked the BQ24075. I also need the 1.5A VIN on the OUT pin when no battery is attached, so I think that at least in this revision I can't go with the BQ25185. 

Thank you for the ongoing support,

Best regards,

Fernando Fontes

Hi Fernando,

I'm happy to support.

Thanks for sharing your full schematic. I've reviewed it closely but still don't see any obvious paths from the battery to the input of the charger.

To help fully isolate the issue, could you try disconnecting all components that might be creating an unintended path from BAT or OUT back to VIN?

Please desolder the following components:

- TPS63020

- TPS55165

- AP63205

- Q1 (the PFET)

Then, please cut the trace or pour going to VIN (+5V) of the BQ24075, and solder a wire to the input cap (C4) so you can apply external power directly to VIN.

This setup should leave only the BQ24075, with BAT connected to the battery and VIN floating unless driven by your external supply.

If battery voltage still appears on VIN with nothing connected to it, that would suggest the conduction path is internal to the charger. If not, then the path is likely coming through one of the removed components or associated traces.

Also, have you been able to replace the charger IC to see if that resolves the issue?

Best regards,

Alec

Hi Alec, 

Alec Lehman said:

If battery voltage still appears on VIN with nothing connected to it, that would suggest the conduction path is internal to the charger. If not, then the path is likely coming through one of the removed components or associated traces.

Ok, Monday I will do the experiment.

Alec Lehman said:

Also, have you been able to replace the charger IC to see if that resolves the issue?

No, unfortunately I don't have more ICs in stock. But will order some and try again.

Best regards,

Fernando Fontes

Hi Fernando,

Sounds good, just let me know what you find.

Best regards,

Alec