0

I have a special miniature ATX power supply, and it has the PW output pins that are supposed to be connected to the motherboard so that the PSU can turn the system on or off. But I have a different device to power up, not a motherboard, and I need to understand what it is that the PSU does with the PW switch.

I expected it to simply short the 2 pins (emulate a button), but it doesn't happen. Then I notice the pins are marked + and -, which makes no sense for a button. Do I need to expect voltage on the PW switch rather than expecting the two pins to be shorted?

P. S. Since you're gonna ask anyway, over and over again: this is the device I'm talking about. Although the question was supposed to be generic, a-la "what sorts of manipulations with PW pins could possibly turn the motherboard on provided it is powered up properly already".

P. P. S. NO, this question is NOT about how to turn the ATX power supply on!

Violet Giraffe
  • 1,187
  • 6
  • 21
  • 36
  • 1
    You are using a power supply "designed to provide power and to control the ON/OFF switch of an ATX motherboard". Your device is **not** an ATX motherboard (and you don't even say **what it is**). Clearly you need to but the correct power supply for your **in car whatever it is**. – DavidPostill Nov 03 '15 at 11:39
  • @DavidPostill: what? First, your sentences are broken. Second, clearly you need to actually understand the question before replying to it. You also need to stop thinking you know what I'm doing better than myself ;) – Violet Giraffe Nov 03 '15 at 14:10
  • "Clearly you need to buy the correct power supply for the device in your car whatever it is" which you haven't told us. – DavidPostill Nov 03 '15 at 14:50
  • @DavidPostill: this is the correct PSU. In fact, this PSU is made _precisely_ for my use case and is advertised as such. – Violet Giraffe Nov 03 '15 at 19:46

2 Answers2

5

I read through the link provided on this little automotive power supply, and it's got the neat feature that it:

is an intelligent automotive ATX power supply for in-car applications with maximum 180 Watts output, designed to provide power and to control the ON/OFF switch of an ATX motherboard based on ignition status.

X5-ATX will send power off signal to motherboard 30 seconds after the ignition off, and will cut off the +5VSB rail after a further pre-defined amount of time (90 or 600 seconds depends on user setting), to prevent the battery from being run out. X5-ATX constantly monitors the battery voltage, and when battery level drops below 11.2V from more than one minute, X5-ATX will shut down and re-activate only when the input voltage is larger than 9V.

A nice feature for controlling a computer, or other device that has a similar momentary power switch. It even comes with a 2 "pin" wire to connect the MB power pins to itself. A regular power switch is just a temporary shorting of the 2 pins. Apparently, this will momentarily short the 2 pins itself whenever it thinks the "ignition is on" (whatever that means, power on the ACC connector is on?). So this PSU becomes the power switch, and "presses it" itself. When a motherboard has it's power swtich pressed (shorted) it should connect/short the sense and ground connectors, which should fully turn on the PSU.

You also say "I expected it to simply short the 2 pins (emulate a button), but it doesn't happen" but it should be happening. Is your Ignition (ACC) wire getting power properly? Or, maybe the DELAY jumper is delaying power on (it's vague in the description of just how it powers on) maybe 30, 90 or 600 seconds? Or, maybe it's just broken.

"But" you say "I have a different device to power up, not a motherboard" so you really don't care how a motherboard power switch works, do you. All that matters is how your "different device" powers up, and that's a different question. If it's got a momentary power switch, just connect the little PSU power wire to the device's switch, and let the PSU turn on the device.

Or, if there's no momentary switch on your mystery device, and if the power supply doesn't turn itself on automatically, then just jumper the sense and ground connectors (like Journeyman Geek mentions) and that should at least turn "on" the PSU.

Xen2050
  • 13,643
  • 4
  • 24
  • 42
  • You've managed to make far more sense of the question than I did the first time round! The documentation is sketchy at best but seems to imply the device expects a MB to power itself on, because it only mentions the ability to signal the MB to turn *off*. The High/Low specification also implies the device doesn't short the pins but pulls the positive pin to local ground, which also makes sense as it's connected to a logic output pin not a relay. – qasdfdsaq Nov 03 '15 at 11:43
  • Definitely vague brief documentation on the webpage, I wouldn't be surprised if it was translated through 2 or 3 different languages first. Maybe it does "turn on" by grounding a pin... If I recall, on the power supplies I've tested they'll turn on if the correct wire (green one I think) is jumpered to any of the black ground wires, I think... – Xen2050 Nov 03 '15 at 11:55
  • Yes, that's how a real switch would act, but I suspect the microcontroller is trying to simulate it by setting an I/O pin to LOW instead of using a transistor or relay to actually short + to -, and hoping the motherboard interprets that as the two pins being shorted. – qasdfdsaq Nov 03 '15 at 11:59
  • I don't think the OP even has a power swtich connector that needs emulation, so understanding all the gory details of how it all works is unnecessary, I suspect the real answer is "connect the sense and ground wires", but it didn't get noticed. Therefore, I'm +1'ing Journeyman Geek's answer ;-) – Xen2050 Nov 03 '15 at 12:03
  • +1'd yours anyway because it most closely addresses the question. If his device expects to receive a signal or prompt from the unit to power off, then how the device emulates (badly) a physical switch is relevant. Of course if he's designing his own embedded device I'm not sure why he couldn't just take the signal from the same source as the power supply itself. – qasdfdsaq Nov 03 '15 at 12:07
  • Yep, you got it. And actually, the description of this PSU is pretty clear and is written in good English _as for Aliexpress_. Now, I wonder: could the PSU manipulate the PW switch so briefly my multimeter simply didn't catch it? – Violet Giraffe Nov 03 '15 at 14:07
  • Also, there's no mechanical relay on the PSU board as far as I can tell, but there could well be an electronic MOSFET switch. – Violet Giraffe Nov 03 '15 at 14:27
  • I've got a digital oscilloscope / logic analyzer, and still can't detect intelligible signals on the PW pins. Is it safe to connect voltage (+5V or +3.3V) to the PW+ pin in order to look for short circuit between PW+ and PW- by observing voltage relative to GND at PW-? – Violet Giraffe Nov 30 '15 at 14:58
2

Well there's two aspects here.

Happily, I get to use the term "proper manly switch" as demonstrated here again. A motherboard power switch is not one of these. A motherboard has a low voltage supply at all times, and you are merely shorting a boring momentary switch that tells your motherboard "WAKE UP YOU SLEEPY HEAD".

A ATX switch will typically power on when the sense and ground connectors are shorted and sometimes you will need a load. Short together the green and black connectors on your ATX supply and try sorting those switches. Alternately its a power led and you're barking up the wrong tree.

Some information on what this "special" power supply is, and the exact model would help, but the above would get you started.

Community
  • 1
Journeyman Geek
  • 127,463
  • 52
  • 260
  • 430
  • Sorry, I understand everything you said but I don't see the answer to my Q. Firstly, I'm not asking about the ATX power supply switch (the green wire in the ATX connector), that one is clear. I'm asking what to expect from an electric circuit that is said to _emulate_ the MB PW switch and is supposed to be connected to the PW pins of the MB. I understand that one way of doing it is to short the 2 pins, but is it the only way? The reason I'm asking is that either my PSU doesn't work properly and I need a refund, or it works and I'm just expecting the wrong thing from it. – Violet Giraffe Nov 03 '15 at 09:09
  • Of little relevance, but here's the PSU in question. There's a decent description of it there - short of schematics, naturally. http://www.aliexpress.com/item/X5-ATX-180w-output-6v-to-24v-wide-input-Intelligent-Automotive-DC-DC-Car-PC-Power/1451084014.html – Violet Giraffe Nov 03 '15 at 09:09
  • That connector looks odd and generic. Connecting *those* pins to the motherboard makes no sense unless there's a relay that turns the motherboard on and off when there's power. Time to break out a multimeter and see what's the voltage between those pins. There should be none. If its 5v, it *emulates* a power switch – Journeyman Geek Nov 03 '15 at 09:13
  • Why does it not make sense? It makes perfect sense. I measured no voltage between the pins, and no drop in the resistance either :( – Violet Giraffe Nov 03 '15 at 09:21
  • No drop in resistance? What happens with and without power? – Journeyman Geek Nov 03 '15 at 09:23
  • And I'm missing something. Maybe its looking for a voltage on the 'ignition' pin. Oh, and short the 'home' headers. – Journeyman Geek Nov 03 '15 at 09:26
  • I have no problem starting up the PSU. I have a problem making it trigger the PW switch. I didn't manage to detect any activity on the PW pins, so it may be broken after all. And there's no point in setting the HOME jumper, that's not the mode I'll be using the PSU in. It's for a car PC, and I need it to turn the computer on automatically when the output voltage appears (triggered by ACC). It's also supposed to trigger PW once again to shut down the PC gracefully 30 seconds before cutting power when ACC goes off. – Violet Giraffe Nov 03 '15 at 09:33
  • @VioletGiraffe: You can't make a PSU trigger a PW switch. The PW switch triggers the PSU, not the other way round. You need to configure the MB to power on automatically, using the "State after power loss" option. – qasdfdsaq Nov 03 '15 at 10:51
  • @qasdfdsaq: no, you're not getting it. This discussion is NOT about the ATX PW_ON switch. It's about the button connected to the MB that you press to turn the system on. – Violet Giraffe Nov 03 '15 at 10:55
  • @VioletGiraffe: I'm getting it perfectly fine. You are doing it wrong. What you want to do is achieved by configuring the motherboard correctly, NOT shorting switches. – qasdfdsaq Nov 03 '15 at 10:56
  • @qasdfdsaq: now you're just ridiculous in your stubborness. You don't belong to StackExchange community with that attitude. – Violet Giraffe Nov 03 '15 at 10:58
  • @VioletGiraffe: You're the one asking a bad question and refusing to listen to the answer. – qasdfdsaq Nov 03 '15 at 10:59
  • 1
    Actuaaaaaly, I think the PSU MIGHT actually turn on the PC but putting it in another mode is a sensible troubleshooting step, especially when bench testing. And its something you might get asked for trouble shooting anyway. – Journeyman Geek Nov 03 '15 at 11:05
  • 2
    This is very confusing, the power supply alone doesn't turn on modern computers. Maybe the OP is looking for just a jumper wire on the power connector to "always on" the power supply? I used one of those on a malfunctioning motherboard, but the computer itself still wouldn't turn on unless the regular power swtich (attached only to the motherboard) was pressed. Hearing *"I need it to turn the computer on automatically"* I think you really do need the *"State after power loss option"*. PS OP - Not winning any friends being *"ridiculous in your stubborness"* – Xen2050 Nov 03 '15 at 11:09
  • 1
    I understand it a bit better now. The PSU supposedly has battery/ignition detection logic that artificially "presses" the power switch on the MB, which is why the OP is asking about it. However the documentation only gives information about switching the computer *off* not *on*. – qasdfdsaq Nov 03 '15 at 11:36