Hacking a power-on button

[Danijel Gorupec]

I have one "stupid" problem... I ordered an embedded fanless computer that doesn't turn on automatically when power is supplied. Instead it waits for someone to press the power-on switch. I found no BIOS option to solve this problem (there is no "power on after power faliure" or similar options in BIOS, or at least I couldn't find it).

 

I only had a brief time to examine the computer (had to ship it to its final destination). I found that:

- as I said, no BIOS option can help me

- the ATX-like power supply is integrated on a single multi-layer high-density printed circuit board

- the power-on switch leads are the only accessible places where I can solder wires. I can also access +5V, +12V and GND from a SATA connector

- one power-on switch lead is at 3.3V, and the other is GND

- a momentary short-connection of power-on switch leads will turn the computer on. A permanent short-connection will turn it off, unfortunately.

- using a 1 kohm (or less) resistor between power-on switch will still turn the computer on (a perfect short-circuiting is not needed)

 

I don't know how exactly the power-on button works. I suppose that there are two posbilities:

A) it drives the power-on line to GND [look at drawing A) on the picture]

B) it drives the power-on line to +3.3V [look at drawing B) on the picture]

 

Does anybody know what option is more likely, A or B? (As I said, I don't have the computer any more to make the test, and I forgot to make this test when I had oportunity).

 

 

 

I will atempt to solve the problem by making a small circuit around the power-on switch. Solutions for A) and B) are also pictured. Do you think this will work? Any better idea?

[Xittenn]

- a momentary short-connection of power-on switch leads will turn the computer on. A permanent short-connection will turn it off, unfortunately.

 

Would this not be caused by power-on reset? The reset is being held high, or on regardless . . . . .

 

PoR generator

[Danijel Gorupec]

Would this not be caused by power-on reset? The reset is being held high, or on regardless . . . . .

 

PoR generator

 

I actually hope that once the power supply is powered on, the +5V will be generated and:

- in case A) it will pull-up the power-on line

- in case B) it will pull-down the power-on line by activating the transistor.

Therefore, the power-on signal will not be active any more.

 

But there might be some traps I don't see.

[Xittenn]

Motherboard model #??

[Danijel Gorupec]

Motherboard model #??

 

Yes, it was not very smart not to write down these data when I had my chance (together with BIOS brand and version). Sorry.

 

It seems that I will have to prepare several most likely solutions and apply the right one durring 'commissioning'.

[Xittenn]

Someone must have the purchase order, or you must at least know the manufacturer name.

 

I think you are oversimplifying your approach, and generally wasting your time by trying to preconceive a solution. All of the embedded systems I've worked with employed solutions that were significantly more complex than anything you have proposed or implied. This wasn't simply to exhaust all solutions, but was also necessary given the onboard controllers for power-on reset generation. Many of the embedded systems I've used have employed standard PCs that required additional hardware for this purpose. They often use a custom card that receives a signal and turns the PC on as necessary. The cards often include circuitry that will detect sudden drops in voltage and will additionally signal the PC and other controllers to act accordingly. Any oversimplified solution may be misconstrued as bad engineering.

 

Also, this isn't my profession and my opinions come from experience with a narrow selection of system types. My experience is mostly limited to CNC and similar types of equipment. You haven't given any details about your project, and these details might be relevant in forming an opinion. The general lack of details surrounding your project might prevent others from having any opinion at all.