I'm using the MP3401A Li-Ion IC in a project but the load LED is annoying me.

When a load is applied the LED turns on but when the load is disconnected the LED stays on for about 5 seconds before going off. During this time it is still receiving about 1.8v from the LED pin on the IC which doesn't tail off so it doesn't seem to be resulting from residual power left in the circuit.

Do I just have to put up with this annoyance or is there something I can try?

Datasheet circuit attached.

 

I suspect C3 has a discharge path with a long time constant. You can shorten this time by putting a bleeder resistor across C3. You can compute a value which will provide a load in parallel with the actual load. Too small and it will consume too much load current, too large and the LED will stay on for a long time. You decide.

1K in parallel with 10 μF gives a time constant of 10 msec. So five time constants or 50 msec should be sufficient to discharge it to a value that will blank the led. In operation the bleeder resistor will draw 5 millamps. I don't know if that is a big deal or a small deal for your application.

Not sure if C2, which supplies Vcc to your chip has a role in this problem or not.

 

I suspect C3 has a discharge path with a long time constant. You can shorten this time by putting a bleeder resistor across C3. You can compute a value which will provide a load in parallel with the actual load. Too small and it will consume too much load current, too large and the LED will stay on for a long time. You decide.

1K in parallel with 10 μF gives a time constant of 10 msec. So five time constants or 50 msec should be sufficient to discharge it to a value that will blank the led. In operation the bleeder resistor will draw 5 millamps. I don't know if that is a big deal or a small deal for your application.

Not sure if C2, which supplies Vcc to your chip has a role in this problem or not.

Unfortunately the bleed resistor doesn't work.

The chip is always outputting 5v but the LED only turns on when a minimum load of 30mA or so is connected. I presume this is some sort of internal load switch in the chip which also provides the voltage for the LED and there's a delay from the load being disconnected to the internal switch reverting back to low power state.

If I remove the cell after the load has been disconnected but while the LED is still on it goes off instantly leading me to believe it is an operation of the chip itself.

Edit: Found an actual datasheet (and not just the circuit diagram) although it doesn't really help. It says if the load is 20mA or less for 16 seconds the LED will go off, although this doesn't seem to match my real world test. Anyway, the LED delay does seem to be inherent of the chip's characteristics and not something which can be amended :/