Tool/software:
I am trying to use the BQ25896 chip to charge a 4000 mAH 3.85V 4.4V max battery at 0.5C requiring 2A. We are sourcing 12V at VBUS. I have the EVM kit configured for this, and it works fine. The BQ25896 gets warm but not hot. I am running the same register setup on a PCBA we designed, and the chip oscillates between drawing current and shutting off at roughly 1/2Hz cycle time, as observed by monitoring the current drawn from the power supply. Here is how the registers are configured:
REG00=1C HiZ=0 ILIM=0 INLIM=1500mA REG01=06 BHOT=0 BCOLD=0 VINDPM=600mV REG02=51 REG03=1A CHGEN=1 REG04=1F PUMPX=0 ICHG=1984mA REG05=33 IPRE=256mA ITERM=256mA REG06=8F VREG=4400mV BATLOWV=1 VRECHG=1 REG07=9D REG08=03 REG09=44 REG0A=73 REG0B=46 VBUS_STAT=2 CHG_ST=None PG=1 VSYS=0 REG0C=20 WD=0 BST=0 CHG=2 BAT=0 NTC=0 (faults) REG0D=53 REG0E=56 THERM=0 BATV=4024mV REG0F=5F SYSV=4204mV REG10=51 TSPCT=58.665% Temp=25.383C REG11=DF VBUS=1 VBUSV=12100mV REG12=00 ICHGR=0mA REG13=10 REG14=06
You can see from register 0C that I'm getting a thermal shutdown charge fault. With the same configuration the IC runs at about 95F temperature on the EVM board, and over 120F on our board, so it definitely appears to be thermal related. There are some wiring differences I want to get input on, to see if they could cause this:
The following lines are floating. It doesn't seem like this should cause an issue, but maybe it does:
QON: Pretty sure this wouldn't cause this as the EVM has this on a switch and the docs say internally pulled high.
ILIM: Left floating. EN_ILIM set to 0 in REG00. Does this need to be tied to something?
PG: This seems like it's an open collector output, but data sheet mentions pulling up using 10K to rail. Is this needed?
INT: Same as above. seems like this shouldn't be needed.
OTG: Left floating. OTG_CONFIG set to 0 in register. Does this need to be grounded?
If it's not a wiring issue, I assume it might be a layout issue. We used the footprint from Digikey which has 5 thermal holes under the part. The EVM kit has 9 (of which 4 were added after the footprint was created. Was a thermal issue found)? We are also building this on a 2 layer board with a much smaller copper pour on the back side of the board. Is this what's causing this? I can go 4 layer if that's what's needed but no one specifies this...
Only other possibility is assembly error. We had this assembled by a trusted assembler. Desoldering the chip showed some signs of incomplete coverage, but is this enough to cause this?
Thanks for any help? At this point we need to build a second proto run, but I hate to do that and just guess at what the correct solution is...
