Go Down

Topic: How to power off the ?C by itself (Read 6033 times) previous topic - next topic


Try using the switch to turn the gate on, instead of 'shorting out' the MOSFET (like I had in my crude drawing).

A design like ajofscott suggested would be  more resilient, though.


I'll try this circuit tomorrow, i'm going to bed now.
Thanks guys :)


My brain says it should work, MOSFETs require so little current it should stay on easily.

Try adding a capacitor from the gate of the FET to ground, just to kill the odds that something is draining current there for a split second, causing your mosfet to shut down.

My circuit is in the attachments underneath.

Yes, I tried to add a 100uF and a 100nF cap without results

That circuit can't work, an N-channel MOSFET's gate has to be more positive than its source to turn on, yet you are expecting both to be at +5.0V when its on... It would be normal to use an n-channel to switch the 0V line, but here that would give exactly the original problem with + and - reversed.  CMOS inverter or similar required.
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]


Here's a circuit I did a while back, I never built it though

There's probably stuff you don't need there but the general idea is to use an N-ch FET to switch the P-ch FET. For example D67 is only needed because I had 30v input, with a ~12v input you wouldn't need it.

Truth is there are high-side FET switches and controllable regulators that can do all this now so I doubt it's worth the effort to make a circuit up.

Rob Gray aka the GRAYnomad www.robgray.com


Greynomad and I both do the same thing with different transistor families. Build our circuit(s) and then go from there. Both can be adapted to be on/off pb controllers with addition of an additional pb. Which would not be a bad idea anyway as it would allow forced powerdown by a human operator.


I don't know a lot about the transistors and FET's.

But you could have a look at the work done at jeelabs.com. The searchbar can be found at the bottom of the screen.
If you want your application to be batterypowered take a look at his battery board which power his nodes on a single AA battery for a LONG time.

Good luck,
My best Arduino project: Bench. Interactive transformable furniture.


Hi, I'm back with good news

I've built the circuit according to your suggestions (and drawed it =>attachment number 1), and it works XD.
Hope it's more or less complete. As you can see I didn't use limiting resistors on the gates, I heard somewhere that they aren't necessary for MOSFETs, am I right?
I also measured the current the circuit draws in the shutdown state, it's about 58 microA.

I tried to improve the circuit and it occored to me that I could use my enable pin of my vreg, instead of the N channel MOSFET . (I think they do the same in this case) Luckily it worked as expected and uses even less current in the shutdown state (about 29 microA)

I hope everthing is correct, otherwise tell me :)

I'm grateful for your responses, you helped me a lot.


Use the pb referenced to ground not the mains. When you press your power button you are applying unregulated power to a micro output pint driving the 5V supply to whatever the high side is. The pb emulates the control transistor being on, not biasing the control transistor on. While your arrangement is sililar, it is completely different and will eventually fry a controller.


Sorry, I drawed the second one wrong. I just copy and paste the first and moved the parts around without changing the button.
However, it was late at night and I wanted to show you my results.
I built it up this way:


Just Like Graynomad, I have a circuit that I drew up but have still yet to test.  In theory it should work.


So many components, so much board space, makes my single coil latching relay solution look better all the time.  ;)


Yes I remember that seemed to be a simple solution. guest1102's last circuit only uses 3 components though.

Rob Gray aka the GRAYnomad www.robgray.com

Go Up