Hello all,


I would like to use the following circuit to power on a Raspberry Pi and handle power-off from the software side.


The circuit currently provides multiple functions:

• Short press → turns the system ON
• Long press → forces shutdown
• Pulling P7 to GND → shutdown from the Raspberry Pi side

However, I need the following behavior:

• (1)Long press to power ON
• (2)No shutdown by long press (disable this feature)
• (3)Shutdown from the Raspberry Pi side should remain the same (pulling P7 to GND)

I assume I need to modify the left side of the IRF7319 circuit, but I couldn’t figure out how to implement the long-press ON functionality while removing long press shutdown (2) and keeping Shutdown from Raspberry by pulling P7 to GND (3).



Any advice would be appreciated.

 

swap places of 100k and 1uF at the left side...
that should eliminate short press and at the same time, long press will used to turn it on (capacitor need to charge to activate mosfets).
initially capacitor will be discharged as before. but it will keep the MOSFETs off. that is until button was on for long enough to charge the capacitor.
after that everything is the same...

 

• (1)Long press to power ON
• (2)No shutdown by long press (disable this feature)
• (3)Shutdown from the Raspberry Pi side should remain the same (pulling P7 to GND)

Alternate suggestion using a 555 for the delay circuit.
Long press duration determined by the C1-R1 time constant

 

or maybe another:

button need to be held down till PI boots and GPIO is able to stay high. turning off GPIO shuts the power off.

 

Welcome to AAC.

ICs like the TI TPS342x (and several others) offer single SMD chip solutions with extensive functionality that, I think, covers your use case.

 

Alternate suggestion using a 555 for the delay circuit.
Long press duration determined by the C1-R1 time constant
View attachment 355931

1. You can get positive feedback for snappy switching action without the 555. The two FETs (and one diode) are all that is needed.

2. When I die, my estate is sending you my entire collection of ***unused*** 555's.

ak

 

or maybe another:
View attachment 355941
button need to be held down till PI boots and GPIO is able to stay high. turning off GPIO shuts the power off.

Of course, use the software to provide the delay. Duh!

 

Of course, use the software to provide the delay. Duh!

exactly... software is used anyway to control the GPIO, and depending on PI version and overhead (use GUI or not), boot times can differ... if my memory serves the usual range is 0.5-2sec. so adding delay in software can do the trick and make hardware simpler.

 

swap places of 100k and 1uF at the left side...
that should eliminate short press and at the same time, long press will used to turn it on (capacitor need to charge to activate mosfets).
initially capacitor will be discharged as before. but it will keep the MOSFETs off. that is until button was on for long enough to charge the capacitor.
after that everything is the same...

View attachment 355926

Thank you ! Really appreciated!
I tried that, I used 1 Mohm instead of. 100k and 47 UF instead of 1UF. Then i get 2-4 sec of delay. Is that normal (so big resistor and big capacitor, in teory it should be possible with smaller res and capacitor.
And another issue, If you press shor short 10 time it charges the capacitor and starts that way to (not a very big issue but would be better without this)

 

Welcome to AAC.

ICs like the TI TPS342x (and several others) offer single SMD chip solutions with extensive functionality that, I think, covers your use case.

Thanks! I checked, it is quite costly, i need a cheaper solution

 

Thanks! I checked, it is quite costly, i need a cheaper solution

Post #4 is the answer.

 

Alternate suggestion using a 555 for the delay circuit.
Long press duration determined by the C1-R1 time constant
View attachment 355931

Very good idea, let me think.
ps: I dont want to change right side of the mosfet because it works really good! It gives couple of sec of time till the shutdown and then cuts the current of raspberry. So I shutdown and when the GPIO is negative system still have couple of sec to safely close the rest.

 

It gives couple of sec of time till the shutdown and then cuts the current of raspberry. So I shutdown and when the GPIO is negative system still have couple of sec to safely close the rest.

I'm not following this line, what is the "rest"?

 

Of course, use the software to provide the delay. Duh!

Post #4 is the answer.

Really? My read of the problem is to delay powering on the uC until after a long button press, so how can an unpowered uC produce a software timer to delay its own power?

ak

 

Really? My read of the problem is to delay powering on the uC after a button press, so how can an unpowered uC produce a software timer to enable its own power?
ak

By delaying the latch output from GPIO 4.
The program does not advance until GPIO 4 goes High.

 

I'm not following this line, what is the "rest"?

The rest is all devices/etc on raspberry. I mean gpio goes negative but real shutdown may take 1-2 sec more than that

 

Question to the TS:

Is the goal of the circuit to delay power to the uC, or to delay operation of the uC when it has power?

A corollary question: Are other circuits/things besides the RPi being powered through this switching circuit?

ak

 

The rest is all devices/etc on raspberry. I mean gpio goes negative but real shutdown may take 1-2 sec more than that

Wouldn't the program switch OFF these devices first then pause a couple of seconds before bring GPIO 4 Low.

 

By delaying the latch output from GPIO 4.
The program does not advance until GPIO 4 goes High.

In that case you could have the RPi continuously powered in a low power sleep mode waiting for a GPIO input toggle, and eliminate the entire circuit.

ak

 

In that case you could have the RPi continuously powered in a low power sleep mode waiting for a GPIO input toggle, and eliminate the entire circuit.

ak

I suppose it depends on the what the TS requires.