What part to act/simulate switch for PC power button??

I'm after making a little Attiny85 based IR sensor which allow me to fire up my PC via IR remote from complete power off. Think pressing the IR key is the same as pressing physical power button :)

I've already got all the IR stuff working as I need but I'm not sure how best to make up the switch part. I've read a lot on using transistors or relays but I've been left a little confused at which is more suited.

As the normal PC button just acts to complete the circuit or short the connection on the motherboard header, my vision for the program was once the IR button is detected, to put a PIN (in my case using for no other reason that I am, I will figure the Attiny side after my breadboard Arduino version works as I need) as OUTPUT to HIGH for 2sec or so. This would then 'flick' the switch allowing / completing the connection from the motherboard header so turning the PC on.

What part is most suited to being the switch? and if you can, would you explain how its used and why its suitable?

Hi,

I would measure the voltage accross the power switch connector. Then use your dmm to measure the current that flows when you short the connector (start with the dmm on the highest current range, then work down).

If the voltage is no higher than 5V and the current only a few mA, you may be able to wire the attiny pin directly to it.

If its 12V and no higher than a few hundred mA, maybe use an ordinary npn transistor like bc337 with a resistor between the base and the attiny pin. But lets have those measurements first.

Paul

Connecting to bare wires which are connected to the motherboard power header give me 5v. I've tried to take a current reading by having my dmm on each side of the switch. Without the switch pressed it says 1ma but pressed it goes to 0ma so I think I must be doing something wrong somewhere.

Would a simple 5v reed relay be an easy option? I presume these can connect from Arduino Pin to Relay, to Ground. Then on the other side my two wires from the PC motherboard header?

I have BC547 and TIP120 transistors here already if any of those could be made to suit?

An ATX power supply uses an active-low 5V power switch signal (the green wire). When this wire is connected to ground the power supply turns on. Disconnect it from ground and it turns off.

The motherboard will be controlling that signal for you, with the power button connection to the front panel acting as a control to its internal circuitry. Chances are it is also an active-low input - connect it to ground to trigger it.

The PSU also provides a low power 5V "standby" supply (the purple wire IIRC), which is always giving 5V regardless of the state of the power switch. This is used to power the control circuit.

As long as the front panel switch is active-low (and I'll bet it is), then a simple NPN transistor (or opto-isolator if you want galvanic isolation) connecting it to ground would activate the button. Control the transistor's base (or opto-isolator's LED) as you would any transistor (or LED) with your microcontroller.

Dynamis_dk: Without the switch pressed it says 1ma but pressed it goes to 0ma so I think I must be doing something wrong somewhere.

Hmm... how are you connecting the dmm? Just unplug the connector to the switch, then use your dmm probes to short the two header pins. Also use the resistance setting on the dmm as a continuity tester. Measure the switch when presses and released. Is it push to make or push to break?

The bc547 would be ok, as would the reed relay. The tip120 would be overkill. But simpler than any of those would be direct connection to the attiny pin, as long as the current is low. You could just test it with a 100R between the attiny pin and the motherboard pin for safety. Dont forget to connect the grounds together or nothing will happen!

PaulRB:

Dynamis_dk: Without the switch pressed it says 1ma but pressed it goes to 0ma so I think I must be doing something wrong somewhere.

Hmm... how are you connecting the dmm? Just unplug the connector to the switch, then use your dmm probes to short the two header pins. Also use the resistance setting on the dmm as a continuity tester. Measure the switch when presses and released. Is it push to make or push to break?

The bc547 would be ok, as would the reed relay. The tip120 would be overkill. But simpler than any of those would be direct connection to the attiny pin, as long as the current is low. You could just test it with a 100R between the attiny pin and the motherboard pin for safety. Dont forget to connect the grounds together or nothing will happen!

1ma would be the right current for an input with a 5K pull-up resistor (it's probably 4.7K in actuality). It's probably inputting direct to a power management chip rather than discrete components.

Thanks majenko for the ATX details, useful info to have at hand as most of my tinkering is PC based.

Dont forget to connect the grounds together or nothing will happen!

hmm, I wonder if in my initial playing this is where I've gone wrong. I have my daughter over the weekend so I might not get chance to play today but I'll try a few things out and see how I get on. Should there be any more advice or a working product come out of it I'll post back my progress :)

Thanks for the assistance everyone - I'm ashamed at my electrical knowledge. It strange than if you give me PCB design as an image I can print, etch and solder away without any problems but when it comes to the design stages I'm only a small notch away from useless! Wish I took electronics in school lol