Raspberry Pi soft power controller – the circuit


James Lewis has been working on a Raspberry Pi soft power controller, that is available on github:

The RetroPie project enables retro-gaming with a Raspberry Pi. All of the Pi models have enough computing power to emulate the major 8-bit and 16-bit computers of the 80s and 90s. With the Pi 3 I have even been able to play PS1 games with no problem. My current project is to put my Raspberry Pi running RetroPie into an old Super Famicom (SFC), or SNES, case. The catch? I want the original SPST power switch to work. And by work, I mean allow the Raspberry Pi to shutdown properly when the switch goes into the off position.  To accomplish this task, I am building a Raspberry Pi soft power controller.

More details at baldengineer.com.

Join the Conversation


  1. LOL hilarious, this guy must be an Arduino guru. I mean, reading such posts are often entertaining for old skool dinosaurs.

    He seems to like quoting an approx 10nA measured on the LDO. Does he think his soft-switch will get him below 10uA or so? Yeah, okay, but what about the power brick’s no load energy usage? He wanted a soft-switch, that will leave the power brick on, right? That’s way more than the I(q) of either the buck or the LDO.

    Why does the LDO need to be off? Just get one with 2uA I(q) or less, a truckload of manufacturers can do that.

    Why does he even need an LDO? Oh yeah, the whole thing is totally crippled by the use of the Arduino. A big thank you to Mr Massimo et al for crippling everybody’s projects (and thinking) worldwide…

  2. Well, at least it’s a source of entertainment for some of us to study the habits of modern hobbyists… it’s *raises an eyebrow, deadpan* fascinating.

    Oh, I forgot, he used a 100nF C0G on the LDO input. Who buys C0G for 100nF caps? Heh.

    Well, he wasn’t happy about the 10uF ceramic SMD. One thing I recall about 25V 10uF SMDs, they are expensive… 25V ooooh. I hope he did actually measure the leakage of just the SMD cap, because the shutdown current of the TPS563200 is 3uA typical. So why worry, the whole exercise of getting it under 1uA seems a little extreme.

    For a power brick application, if the shutdown load is less than 10uA, who cares about getting it to 1uA… for a 9V line no less, ouch, such luxury. One have to get premium caps, check flux residues, consider use of conformal coatings, what else?

  3. Yeah, so I’ll make a cell-phone call near this circuit and it will go GRAZY due to unwanted induced currents. Or am I missing something?

  4. No, the fact that the circuit consumes only a few 100nA in power-off state doesn’t imply that it is very sensitive to induced current. The current that is necessary to activate the circuit is rather high.

    However, the same functionality can be implemented with a rather simple latching supply circuit, without an additional microcontroller.

Leave a comment

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.