BQ25713: Charger stops charging battery

Hello Mitchell,

When the failure occurs, are you able to resume charging if power is removed from both VBUS and SRN pins (POR reset of the device) and then power the device back up and reprogram the register settings?

Regards,
Steve

Hi Steve,

No, even if the board is completely powered down it doesn't function again. They are able to reprogram the settings, the BQ chip accepts settings, and reports voltages, but the Buck/Boost does not function.

Thanks,
Mitchell

Hi Mitchell,

I am afraid that this generally indicates physical damage either on the IC or elsewhere on the board. Any of the error conditions that can latch the device into HIZ mode (such as short circuit detected on SYS) are cleared when power is fully removed.

Perhaps the best way to track down the issue is to simplify the system first. Even with the battery removed, BQ25713 should switch with an adapter attached in order to regulate SYS at VSYSMIN. This should occur with default I2C register settings. The VSYSMIN will be determined by the CELL_BATPRESZ pin to 3.584 V for 1 S, 6.144 V for 2 S, 9.216 V for 3 S, or 12.288 V for 4 S.

So please remove the battery and, if possible, replace the solar cells with a bench power supply and apply power. Any voltage between 5V-20V should be fine. 12V or 20V is usually a good choice.

If you are still not seeing any switching and no voltage on SYS, please:
1. Check the voltage at the VBUS pin (pin 1) preferably with a scope to make sure it looks as expected
2. Check the voltage on REGN (will be 6V as long as VBUS > 6.7V, otherwise VBUS - 0.7V)
2. Check CHRG_OK -- it should go high 50 mSec after source voltage applied
3. Check ILIM_HIZ (PIN 6) -- if this pin is grounded it will put the device in HIZ mode which will keep it from switching
4. Take a scope capture of Gate Drive pins for the appropriate FET pair -- either HIDRV1/LODRV1 if operating in Buck (VIN > VSYSMIN) or HIDRV2/LODRV2 if VIN < VSYSMIN. -- perhaps FETS are damaged and gate drives are going but nothing is happening at the FETs.

If you seeing expected values for 1-3 and still nothing on the gate drives, then I would recommend removing the device from the failing board and replacing it to see if this fixes the issue. This will help to determine if the issue is in the device or the board.

Let me know the results of these tests and we will figure out the next steps accordingly.

Regards,
Steve

Hi Mitchell,

I haven't seen a response to the above in more than a week so I am assuming you were able to resolve your issue.  If not, you can reopen the thread and let me know the results of the above.

Regards,

Steve

We have seen similar problem on our product which uses the BQ25713RSNR.  The unit stops charging the battery after 3 days.  After checking the signals as recommended, we have found that the ILIM_HIZ pin is stuck low.  The resistance measured between ILIM_HIZ and ground is about 1.2K whilst it is measured, on a good working board, 100K+ (we have 348K//200K on our external potential dividing circuit on ILIM_HIZ).  When conducting a diode drop measurement on the ILIM_HIZ pin relative to VDDA, also relative to ground.  The pin conducts both ways (+ve on ILIM_HIZ and -ve on VDDA and vice versa; +ve on ILIM_HIZ and -ve on ground and vice versa).  Does this imply the ILIM_HIZ pin is damaged?  Any opinion on what would cause the damage?  The unit was left running for 3 days powering the load on and off  5 minutes every 10 hours (the load draws about 200-300mA).  After 3 days, Vbus was applied but not charging.

Hey Shing,

I would recommend you monitor the voltage on the ILIM_HIZ pin and the VDDA pin to make sure an over-voltage event above the absolute maximum rating is not occuring.

Based on your testing, you can trigger if the voltage on ILIM_HIZ gets close to or above 7V.


Regards,
Joel H

Hi Joel,

 

Since VDDA is derived from the BQ25713 internal regulator (REGN) output, does it imply the over-voltage is generated by BQ25713?  Is there no clamping on REGN output internally to the IC?

 

Regards

Shing

Hey Shing,

Do you have a schematic you can provide? Is ILIM_HIZ driven by anything else? Is it only tied through a resistor divider from the VDDA rail?

If you are constantly cycling power to the charger, is it possible there was an over-voltage transient on the input side? Have you measured any other points for shorts to GND?

I still recommend you monitor the input voltage at the charger's VBUS pin, output voltage on the SYS pin, ILIM_HIZ voltage at the pin, and even the REGN voltage at the pin to make sure there was no transient event to trigger damage.


Regards,
Joel H

Hi Steve,

Sorry for the delay. My customer was on travel and just got back to me with some answers to your questions--

1. Check the voltage at the VBUS pin (pin 1) preferably with a scope to make sure it looks as expected
18V as expected

2. Check the voltage on REGN (will be 6V as long as VBUS > 6.7V, otherwise VBUS - 0.7V)
6V

2. Check CHRG_OK -- it should go high 50 mSec after source voltage applied
Looks good (we have an signal FET driving an LED on this line)

3. Check ILIM_HIZ (PIN 6) -- if this pin is grounded it will put the device in HIZ mode which will keep it from switching
3.3V looks good

4. Take a scope capture of Gate Drive pins for the appropriate FET pair -- either HIDRV1/LODRV1 if operating in Buck (VIN > VSYSMIN) or HIDRV2/LODRV2 if VIN < VSYSMIN. -- perhaps FETS are damaged and gate drives are going but nothing is happening at the FETs.

HIDRV1 looks terrible.  Non-working board HIDRV1 in yellow (CH1):

Nominal working HIDRV1:

On the nominal working board we do see some ringing when there is no battery to charge and the circuit is lightly loaded, but this could just be my probe causing the ringing. I would not expect that to damage the chip though (right?).

For a second test:

If we enable PTM (Pass Through Mode) with the steps mentioned in the datasheet:

• Set REG0x32[7] = 0, to disable the EN_EXITILIM.
• Set REG0x31[0] = 1.
• Set REG0x30[2] = 1.
• Ground ILIM_HIZ pin.

Shouldn't both HIDRV1 and HIDRV2 be driven and held high if the BQ chip is not damaged?

Thanks,
Mitchell

Hi Mitchell,

The high-side gate drive is driven to (SW + 6V) in order to turn on the high-side FET.  This is done by charging the bootstrap capacitor to 6V using REGN when SW node is at ground and when the SW node rises, the bootstrap cap is referenced to the SW node, so it's positive voltage is then SW+6V.  (This is a pseudo-charge pump.)

My guess would be that the sinusoidal oscillation that you are seeing in the gate drive signal is actually a sinusoidal oscillation in the SW node.  If so, this would indicate that the high-side FET is trying to turn on, but isn't.  In order for the SW node to ring like that, it usually means that neither the high side nor the low side FET has been turned on and the inductor is ringing with a parasitic capacitance in the low-side switch.  You shouldn't be seeing ringing like this in normal operation, and I don't think it is just noise in your scope probes -- it's too long-lived to be probe noise. 

Can you also measure the SW node when you see this ringing (the node between the source of the HI-side FET and the drain of the LO-side FET, also called the Phase node in some of our documentation)?  And if you can either use scope math or just line up the HIDRV signal to the SW signal so that we can get an idea how far above the SW node the HIDRV is reaching, that will help as we can compare it to the Vth of the high-side FET.

I think we need to get to the bottom of this ringing.  To answer your question, I don't think that the ringing can damage the chip unless the voltage overshoots past the rating of one of the pins, which it doesn't appear to be doing.  I will need to run a no-battery test on the EVM to be sure, but I don't think that I will see ringing like what you show.  I think this might be symptomatic of or even a precursor to the issue that you are seeing where the charging stops. 

Also, if you can remove the persistence in your scope capture, that will help.  It is a little difficult for me to see what is happening with multiple captures overlaid. 

Regards,

Steve

Hi Steve,

After looking into the ringing more, it is periodic on the order of 10 seconds. 

CH1 is HIDRIV1, CH2 is SW1, M is Math difference between the two

Then it will start ringing:

Thanks,
Mitchell

Hi Steve,

To hopefully save you some time, we tested the EVM with our load and did not see the same ringing. So, the problem is definitely something caused by our board/design.

Thanks,
Mitchell

Hi Steve,

I shared the schematic with you offline yesterday. Would you be able to take a look and see if anything jumps out?

On thing that stands out to me is that they decided to use the CSD88539ND instead of CSD17551Q3A (this is the one on the EVM). The key difference to note is that their FETs have a V_GS of 3V instead of 1.6V. I'm wondering if this is the reason why it appears the high side FETs aren't fully turning on.

Thanks,
Mitchell