RGB LED PL9823 doesn't completelly turn off in stand_by_power_down with P-Mosfet

Good night!

I’ve got a project where a Pro-Micro (Leonardo) when powers up enters stand-by mode, and only leaves it when a capacitive button is pressed. To turn off all modules, when in stand-by, (RGB LED, TOF sensor, voltage sensor and power led driver) I used (by exploring the forums) 2 P-Channel Mosfets (one for 5V and one for the 8.4V battery).

Everything works nice, the only thing that makes me curious is that when entering stand_by_power_down mode, the RGB led still has a faint glow. When it leaves stand_by_mode it gets back to normal states. It seems that it never gets fully disconnected.

Could there be any … energy leaks, or something not grounded correctly, that makes it, even with the mosfets off, still some energy comes to this RGB led and lights if faintly?

I’m attaching my full schematics. If someone could give me a hint, or advice, it would be great.

Thanks a lot!

The LED can be powered by the VCC pin, but also by the data pin.
Disconnect one of them to find out.
leo..

Wawa:
The LED can be powered by the VCC pin, but also by the data pin.
Disconnect one of them to find out.
leo..

Hi!

I thought about it also. When disconnecting the data pin, it floats and the led turns on random colors, mostly bright blue, but when entering stand_by mode, it sometimes turns off competely and sometimes has that faint dim. I tried to ground even the unused OUT pin of the led (I'm only using one led), to no effect.

The strange thing is that, the Pro Micro, with nothing connected to it, when in stand_by_power_down mode draws 3.3mA. I can't make it draw less, without removing phisically it's leds (Which I didn't to compare).

And in my circuit, with everything connected to it, in stand_by_power_down mode draws the same 3.3 or 3.4mA, even with this RGB led with a little current over through it.

Is there any different wiring I should try?

Rimbaldo:
but when entering stand_by mode, it sometimes turns off competely and sometimes has that faint dim. I tried to ground even the unused OUT pin of the led (I'm only using one led), to no effect.

A fet, even when turned fully off (gate==source), has some leakage.

Try a 100k resistor from LED VCC to ground in sleep mode.
If the LED turns fully off, then I wouldn't worry about it. It's just some leakeage.
You could fit it permanently if you realy want the LED fully off.
If the LED doesn't turn off, then the fet has a higher than average leakage.
Higher than the ~1uA in the datasheet. Maybe you should then replace it.
Did you treat it with ESP precautions, and used hot air or an oven to reflow.
Leo..

Wawa:
A fet, even when turned fully off (gate==source), has some leakage.

Try a 100k resistor from LED VCC to ground in sleep mode.
If the LED turns fully off, then I wouldn't worry about it. It's just some leakeage.
You could fit it permanently if you realy want the LED fully off.
If the LED doesn't turn off, then the fet has a higher than average leakage.
Higher than the ~1uA in the datasheet. Maybe you should then replace it.
Did you treat it with ESP precautions, and used hot air or an oven to reflow.
Leo..

Thanks! I did some tests. When using the 100K resistor from LED VCC to GND only attached in stand-by mode it really turned the led off. If I let it connected all the time in the circuit, the led in stand_by lit more faintly than before.

I decided to try a 10K resistor the same way. With the 10K resistor connected in the circuit, at all times when stand_by mode gets called the led gets completely off.

Could I have it (10k resistor) connected for good in the circuit? Would it draw more current when in stand_by?

I'm still on a protoboard, so nothing soldered yet.

I'm actually using for testing a Mosfet P IRF9540, but the final one will be the one from the schematics...

Thanks a lot for your help!

Rodrigo