BQ24072: BQ24072

HI Bill,

Would you be able to measure the voltage at the ISET pin as well? We can use the current translator feature of the device as a method of measuring charge current. 

Just to clarify, I see in the measurement that the VOUT you mention is = VBAT. Can you confirm if this is exactly VBAT or is it higher or lower than VBAT by a small amount when the 5V is applied to the charger? This would help indicate which source (VIN or VBAT) is powering the  OUT pin since this device's OUT voltage is battery tracking + 225 mV. 

There looks to be an issue with the last two devices for sure as the TS voltage does not seem to be present when it should be with a 5V charge and the /PG and /CHG pins are HI-Z (third device is also HIZ but the VTS is active). 

Are you able to provide waveforms for some of these devices? It'd be good to see how the device operates when VIN is applied (0 to 5V) and have the OUT, ISET, and TS voltage waveforms.

Best Regards,

Anthony Pham

Hi Anthony,

Please see attached for the screenshots that you requested. I also updated the data table with battery and Vout values and placed in in the PP.

Thanks

 7658.BQ24072_SOF_417SCH.pptx

Bill

Hi Bill,

Thanks for doing this data collection!

There are definitely two devices that look like they're attempting to charge(SN3473 and SN3273). The CH3 waveform comes up in the first few 100s of us most likely to check for shorts as part of the startup routine. Are you able to zoom out so that the horizontal scale is maybe in 500 ms per division? This isn't enough for me to see. 

SN6887 and SN4931 don't look like they're attempting to charge or switching to power by VIN. Can you verify that ISET is not shorting to GND? Since it doesn't see a valid voltage here it may not be initiating charge. 

Hi Anthony,

No worries on data collection, anything to help understand the issue.

See attached for additional graphs w the 500ms/div time scale and updated table with the resistance value from the Iset pin to ground. (spoiler - no shorts)
FYI - spending some time today trying to break the charger by repeated insertion removal of system from the charger and then repeated insertion removal of the wall wort from the AC. Solid to date. 

Thanks

Bill

BQ24072_SOF_417SCH_9282022.pptx

Hey Bill,

Appreciate the updates. The ISET pin I wasn't too worried about, just had to sanity check before diving into other possibilities. 

For SN3473 and SN3273:

It's very interesting, the device does appear to go through the proper startup with the voltage checks on ISET so it definitely does appear to recognize VIN being present. They even have VTS powered which can only happen with VTS present. Would you be able to remove the battery and measure VOUT while is VIN plugged in (waveform would be great)? I'm curious to see if it's the same or where the device seems to be not enabling charge in the startup sequence. 

For SN6887 and SN4931:

The lack of behavior here is concerning. It'd be good to look at what happens if you remove VBAT here as well.

SN5101 

This one's a puzzle. It appears to keep the ISET at 500 mV which I wouldn't expect at all unless it actually was charging. 

Best Regards,

Anthony Pham

Hi Anthony,

 BQ24072_SOF_417SCH_9292022.pptx

Please see attached for the additional requested graphs.

Looking at the block diagram 9.2 (page 18 of datasheet) it seems like Q1 is not fully turning on.

 Bill

Hi Bill,

Just curious, what buck/boost converter are you using? By any chance would it have some sort of regulation loop that keeps the VIN from dropping?

There does look to be an issue with the Q1 FET here if there isn't.

Do you have information on your power supply? Is there a current limit at all? I understand the board only consumes 10mA but wanted to check and see if it matches the power supply reading.

Best Regards,

Anthony Pham

Hi Anthony,

Not 100% sure I understand your question so I added some words to the charger schematic.

Note: the 10mA I mentioned previously is the TPS63036 supplying power only to this one board, in the system there are 2 boards.

The wall charger is solid - the charger behaves the same regardless of wall wort or bench supply. This is the same charger used on the previous board that has the BQ25616 set to 2.0Amps. In office we do monitor the supplied voltage and current and even at the 2A the output voltage is 5.0V and clean.

Bill

BQ24072_SOF_417SCH_10042022.pptx

Hi Anthony,

Any ideas into why the Q1 might be degraded and more importantly how to prevent in the future?

Thanks

Bill

Hi Anthony,

Looked at datasheet again- Question re max conditions:

Assume a dead battery -   the charger is set to charge at the maximum allowed 1.5A. If the user charges while the system is on and drawing several hundred mA is that "exceeding" the current allowed through Q1 since the battery is dead and not suppling any power for some period of time?? Is that sufficient to cause damage??

I have abused systems w the BQ24072 and our battery chargers - Vin insertions / removals at room temp and at 50C. Vin(+) connected first, Vin(-) connected first. Over 1200 cycles over 2 systems. No issues.

Looking to borrow an ESD gun and inject that into the Vin pin and see if I can induce the failure.

Still several returns a week - 

Bill

Hey Bill,

The input pin is limited to an absolute max of 1.6 A. If the device is attempting to supply 1.5A to the battery and the power supply is able to supply the 1.5A, it will.

With the addition of a system load at the output that you say "draws several hundred mA", this most likely won't go through the Q1 FET. The device will go into supplement mode and the device will divert some of the charge current that would go into the battery to the OUT pin. In the range that you mention, this would be fine from my experience. 

If the battery is dead (battery voltage < VBAT(SC) where VBAT(SC) is typically 1.8V) we'll only charge it with a max of 11 mA (this is the IBAT current used for short-circuit detection). This would now only be 11 mA + #00s of mA which the device will be able to supply as your current limit is 1.5A.

I'm looking forward to hear back on the ESD tests. 

For any of the returns, is there any signs of visual damage on the device? Does replacing the IC allow normal charging? 

Best Regards,

Anthony Pham

Hi Anthony,

No signs of physical damage on any of the returns. I have replacement parts in bound to try.

Bill

Hey Bill,

Looking forward to the ESD tests and testing an IC swap. 

Best Regards,

Anthony Pham

Hi Anthony,

Dropped a new part on a "failed" board and everything works.

I think that pretty much confirms that something happened to the charger in the field.

Still trying to locate a ESD source (cheap ESD source).

Bill

Hi Bill,

That's good to hear that a new part doesn't exhibit the same behavior. On these returns, is there any info on when they started failing? Any info as well on the typical conditions that your product is operating in?

Looking forward to hearing back on the ESD tests.

Best Regards,

Anthony Pham

Hi Anthony,

Other than we started seeing returns several weeks after we changed chargers no real intel. From a user perspective the collar does not turn on - and they send it back.

Note: the charger is located in GPS based dog collar. The collar will see outdoor temps that a dog would see, expected worse case high temp would be a car dashboard (60Cmax).  The system is intended to be charged at room temperature, although I did charge a small sample of collars at 60C without issue.

Unable to borrow an ESD gun yet, but today am using available materials to abuse the pins. Replaced DC52 (1nF)  with a TVS diode (Littlefuse SESD0201X1UN-0020-090) on several test boards JIC ESD is the issue. See attached schematic. 

Do you see a downside to adding the TVS in place of CH52? BQ24072_SOF_417SCH_10172022.pptx

Bill

Hi Anthony, 

At this point we replace the system PCB with the charger and send back out to the customer.  We will need to have our CM replace parts on boards. I was able to replace one part but we are not really tooled up here for that rework.  Returns are running 3 or 4 a week. To date no loopers on the "repaired" collars.

Still unable to get a charger to fail by inserting and removing the charger power. In our system the charger power enters the system via 2 small pogo-pins on the charge cradle that make contact with targets on the exterior of the system as the collar is slid into a charging cradle. I am able make and break the positive or negative charge input independently - 2000 insertions removals across 4 systems - no issues. Even tried using a bench supply set at 6V and 8 A trying to induce a failure. Currently rubbing the assembly different materials that they will encounter in the field. (simulating ESD hits - maybe:))

The two changes discussed above are being rolled into the next PCB build, but we still need to clear the pipe line. Would really would like to cause the failure mode and then show it was corrected.

Bill