BQ24650:bq24650 sense resistor burning

Hello Michael,

1. The input NMOS is being used instead of the diode. We also cancelled input NMOS and put a diode instead. The results are same.

1.A Now I have measured a working board, Q1_D voltage is 16.05 Volts and Q2_D voltage is 13.27 Volts.

2. MPP is OK.

3. LABEL Thermistor is open circuited. Only on board R6 and R5 resistors are being used.

4. Battery voltage is OK.

As you see below, we cancelled input NMOS and used a diode as shown in the application diagram.

This is our design. Please note that input NMOS is removed and a diode used.

How is the performance with the diode now?

Thanks,

Mike Emanuel

The problem still persists and has no effect. 

What is the purpose of D6, R30, and R31?

Thanks,

Mike Emanuel

NMOS is used to keep the mosfet on. When we use diodes instead of NMOS mosfets, we don't use them.

Did you physically remove them when the NMOS is removed?

yes we removed it. There is a problem when we connect the battery to the output. We have no problems with login. There is no problem when we apply voltage only to the input. When we connect the battery, the si7288 mosfet, sense resistor and IC burning.

What voltage battery are you applying?

13V

Dear Emanuel, We have already tried removing NMOS and all of its related components. Problem still exists even though when we use pure diode. As you see in the real photo, we did not put any component related with the NMOS.

As you confirm, the programmed battery voltage is 13.55 Volts. The battery voltage floats betweem 12V - 13.5 Volts during normal operation.

It looks like we are having problem with driving si7288.

1. We checked out the layout; no obvious problems has been seen.

2. There is a small difference between Eval board schematics and datasheet. We already mentioned that.

3. We do not have a knowledge of detailed theory of operation, so we can not foresee if there is problem with lead acid charging, inrush current etc, programming resistors etc. 

4. We did our best to follow the instructions on the datasheet.

Thanks

İlhami

Correct me if I am wrong, but your feedback network doesn't looked like it is connected to the battery.

Indeed we preferred to connect it directly to the battery connection terminal.

BTW we already cancelled all NMOSes in the purpose of reverse protection. Problem still exists.

Have you tried another board and IC with just the diode at the input? Are you still using the same IC, it may be damaged.

We manufactured 20 units. All was fine.  But when we started field tests, we have seen that there is a burning issue when you turn on the switch. We have already burned 10 units. We never tried to use the IC again after burn issue.

We are already using LT3652 in our solar product group and we were planning to use 24650 as a replacement. But we can not move further if we can not find a reliable solution. 

UPDATE: It looks like problem is happening when we run the charger without solar input. System works well with solar panel connected. But when I remove solar panel connection and connect only the battery, sense resistor, mosfet and IC burns. It looks like there is problem with VCC connection.

There is no problem when you remove solar input on the running system. It happens only on the cold start.