depending on RPI version and installed hardware, that draw may be anything between couple of mA to few tens of mA. for many uses that may not be suitable.

 

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 target is turn on with long press. The right side of the mosfet is already handling the safe shutdown with the GPIO.

 

Here's an alternate circuit.

R1 adjusts the delay. Time delay depends on NMOS VGS(th), but an assumption is the delay time is not critical.

 

Here's an alternate circuit.

R1 adjusts the delay. Time delay depends on NMOS VGS(th), but an assumption is the delay time is not critical.

One issue I see is the power OFF is also delayed when pin7 goes Low.
C1 has to discharge through R1 and R4.
Still think post #4 is the best approach.

 

One issue I see is the power OFF is also delayed when pin7 goes Low.
C1 has to discharge through R1 and R4.

Your right.
This should fix it (added diodes, lowered value R4).



Edit: Need to Add pull down resistor at M2 gate

 

or just remove D2

 

If you remove D2, the uC high output, necessary to keep M1 turned on after the button is released, now will be connected to the timing capacitor, so it will not discharge (reset) while the circuit is running.

ak

 

If you remove D2, the uC high output, necessary to keep M1 turned on after the button is released, now will be connected to the timing capacitor, so it will not discharge (reset) while the circuit is running.

ak

Huh?
Pin7 Connected through D3.
When pin7 goes Low C1 will discharge through R4 and D1, D3 will be reversed bias.

 

When pin7 goes Low C1 will discharge through R4 and D1, D3 will be reversed bias.

Yes, but that's not the condition I was describing. When the pushbutton is open *and the RPi circuit still is running*, pin 7 is high.

I interpreted "remove D2" as replacing it with a direct connection. With a wire connection there, a high at pin 7 will hold up C1 because holding up R4 will take around 1 mA, which the RPi GPIO pin can source easily. Thus, C1 cannot begin to discharge until the RPI shuts down itself. For yhe ON-Delay time to be repeatable, C1 must discharge for at least three time constants. That could be a problem for some restarts. D2 lets C1 discharge through R4 while the RPi circuit is running, so it can be fully discharged and ready for the next start cycle immediately after an RPi shutdown.

If he meant replacing it with an open circuit, then you no longer need D3.

ak

 

He means replacing D2 with a wire. In order to pull the gate down through R1 and R4 when power is OFF.
I see your point but C1 will discharge rapidly through R4 and D1 when pin7 is low.
If D2 is removed how would the mosfet initially get turned ON?

 

Without a wire link, it wouldn't. I think we are in agreement on how the circuit works.

ak