BQ25890: BQ25890 charging current rise and down frequently when powered with 9V supply.

Frank,

Are you disabling the watchdog?  At what voltages are D+/D-?  What is the profile of your system load current?  Do you have an oscillscope shot of ISYS, IBAT, VSYS, VBAT?

Hi Jeff,

The program is this:

while(1)
{
tempL=BQ27411ByteRead(0x02);
tempL=(tempL|0x40)|0x10;
BQ27411ByteWrite(0x02,tempL);//1 – Start 1s Continuous Conversion
printf("REG02 = %d,\n",tempL); 

tempL=BQ27411ByteRead(0x12);
printf(" %d mA \n",(tempL&0x7F)*50); 

tempL=BQ27411ByteRead(0x11);
printf(" %d mV \n",(tempL&0x7F)*100+2600); 

tempL=BQ27411ByteRead(0x00);
printf(" %d mA \n",(tempL&0x3F)*50+100); 

tempL=BQ27411ByteRead(0x0B);
tempL=(tempL&0xE0)>>5;
if(tempL==0)
printf(" No Input \n");
else if(tempL==1)
printf(" SDP \n");
else if(tempL==2)
printf(" CDP (1.5A)\n");
else if(tempL==3)
printf(" DCP (3.25A)\n");
else if(tempL==4)
printf(" Adjustable High Voltage DCP (1.5A) \n");
else if(tempL==5)
printf(" Unknown Adapter (500mA) \n");
else if(tempL==6)
printf(" Non-Standard Adapter (1A/2A/2.1A/2.4A) \n");
else if(tempL==7)
printf(" OTG \n");

tempL=BQ27411ByteRead(0x0C);
printf(" REG0C = 0x%x\n",tempL); 

tempL=BQ27411ByteRead(0x10);
printf(" %d \n",(tempL&0x7F)); 

tempL=BQ27411ByteRead(0x0E);
printf(" %d \n",(tempL&0x80)>>7); 

Delay(0x0FFFFF);
}

      1. watchdog not disabled, but the CONV_RATE bit is written frequently, so the bq25890 is always in host mode.

      2.  I use a desktop power supply, D+/D- is shorted.

      3. System load is about 200mA.

      4. It not convient to monitor Isys and Ibat on my board. I read the registers instead.

      And I take another test this morning.

	9V when normal	9V when drop down
Isource	1200mA	950mA
Ichgr	1900 mA	1250 mA
VBUSV	8000 mV	8400 mV
IINLIM	3250 mA	3250 mA
VBUS_STAT	DCP (3.25A)	DCP (3.25A)
REG0C	REG0C = 0x0	REG0C = 0x0
TSPCT	52	52
THERM_STAT	0	0

    The 1200mA last for about 5 seconds and 950mA last for about 1 second.

    The bq25890 is very hot.

    This time ,the THERM_STAT is also monitored. but I find it is always 0 ,means never get into In Thermal Regulation mode.

Frank

Frank,

I will need to see all of the register readings during the 5 second period and also the 1 second period. If the top of the IC is hot, then it is likely going into and out of thermal regulation. Your register read may not be capturing the THERM_STAT bit. Do you have the IC's power pad properly soldered to the ground plane?

Jeff,

     "  Your register read may not be capturing the THERM_STAT bit"

     I think you are right, I do more test and monitor the registers carefully. And find what I miss is not THERM_STAT, what I miss is REG0C.

     Because the read is not frequently enough.

     Here is the result:

	9.2V	9.2V	9.2V	5V
Isource	980	980	477	695
Ichgr	1450 mA	1450 mA	550 mA	400 mA
VBUSV	8500 mV	8500 mV	8500 mV	4500 mV
IINLIM	3250 mA	3250 mA	3250 mA	3250 mA
VBUS_STAT	DCP (3.25A)	DCP (3.25A)	DCP (3.25A)	DCP (3.25A)
REG0C	REG0C = 0x0	REG0C = 0x2	REG0C = 0x2	REG0C = 0x0
TSPCT	52	51	51	61
THERM_STAT	0	0	0	0
TS PCT		44.715%	44.715%	49.365%
REGN		6.16V	6.14V	4.32V
TS value		2.75V	2.74V	2.13V

 REG0C=0x02 meas : TS warm.

 The battery is warm, so the bq25890 lows down the charging current.

 And I find that, it's because the battery is under the board, the board is hot so the battery because warm. And I replace the battery, the current rise and down phenomenon never happens again.

 But I still have one more question:

 The REGN is 6.16V when powered with 9V and REGN is 4.32V when powered with 5V.

 That makes the reference of TS variable. So the TS value can not reveal the real temperature of the NTC resister. Why does bq25890  not use a constant reference voltage for the TS?

 Thank you.

Frank

Frank,

Great job debugging this on your own! As REGN varies, the TS pin voltage varies proportionately by the same amount. The TS pin voltage is then compared relative to the REGN voltage.