Windows 10: PSU or Power Button fault?

Hello,

I have a somewhat peculiar problem. For a while now, to boot up my PC I'd have to press the Power button several times until it succeeds; the PC then boots up normally as if nothing happened. However, the Power button is now not working anymore even if I press it a hundred times.
So I tried booting up the PC by touching the Power pins on the motherboard with a screwdriver, and that works perfectly every time at first try (I have tested this several times now).

From this I'd think that the Power button or its cord to the mobo is at fault here. However, the weird part is that when I put the PC into Hibernation or Sleep mode (instead of completely shutting it down) and waking it up with a key on my keyboard afterwards, the PC wakes up normally but the monitor sometimes says that there's no signal. I have checked the connections to the monitor and everything is fine (don't have another PC to check it). When there's no signal, I'm forced to abruptly cut the power to the PC, thus completely shutting it down. When trying to boot the PC up again (by touching the power pins), the PC and the monitor go on normally again.

What can I make out of this? Is it really the Power cord that's at fault here or the PSU? How can it be the PSU if I'm able to boot up my PC by touching the Power pins on the motherboard? And is the "no signal" problem of the monitor a seperate issue from all this or can this also be caused by a faulty PSU?

:)

 

The kernel power manager has initiated a shutdown transition. Shutdown Reason: Button or Lid

I suddenly started having this issue as well. Its possible it can be a fault on the PSU itself, or an issue with the power buttons on the case so check the buttons are not catching and are connected properly. Its also possible the fault may lie in the system
board itself

However I did stop the irritating shutdown by going into Power Options, Change Advanced Power Settings, scroll down to "Power buttons and lid" and change the setting for "Power button action" and "Sleep button action" to "Do nothing"

Obviously this will affect the operation of these button but it may help keep the PC running whilst you diagnose the root cause.

I haven't found it myself yet, my machine seems fine, the temperatures are all within range and performance is Ok so I'm suspecting my PSU as its 5 years old.

 

How to prevent Windows 10 from shutting down

I realise that you have tested the PSU but, it still stands as the main suspect in my mind.

some temporary suggestions, see what happens

click start and type power plan and select 'choose a power plan' from the list (slightly quicker than following all the links through settings)

next to the power plan that is selected click 'change plan settings'

click 'change advanced power settings'

under power buttons and lid set power button and sleep button (and lid if its there) to do nothing.

should prevent events from power button.

and/or

under processor power management, maximum processor state, knock it back a bit, say to 90%

if it is experiencing a power issue when the cpu is maxed, throttling it back might prevent it, if it does I would definitely look at the PSU as the cause of the issue.

 

I would agree with your instincts: the power button is bad, the PSU is OK.

The monitor problem I think is a separate issue. No signal means that there is no output from the graphics card. If the problem was related to the PSU it would mean the PSU was not delivering power to the card through the dedicated lines. I would think that this would also affect the graphics card's fan(s). Do they come on when you come out of sleep?

A long shot but try this: disconnect the power switch from the MB header and leave it off. Then try to come out of sleep. Let's see if that changes anything.

 

Thanks for the reply. I was able to take a power button switch with its cord from another older pc, and it works perfectly. So it must be the power button/cord.

The only problem is that this power button switch is too small to fit in the hole of the previous switch. So when I press the button of the chassis from the outside, it pushes the whole button out instead of merely pushing the switch in. I'm not sure if I can find any that would perfectly fit in my chassis.

 

Yeah, that is always a problem with the switches. They are not uniform in size.
Sometimes you can get a washer (plastic or rubber washers are good/better) behind the switch to get it outside the hole, then use some hot melt or sticky tack to hold it in place.

Any effect on the sleep problem?

 

That might be a good idea.

Regarding the sleep problem, when waking up the PC but the monitor says that there's no signal, the fans do normally spin yes.

So I've disconnected the power cord from the MB and when putting the PC in Hibernation mode, I can't wake it up anymore with the keyboard or the mouse. So they probably do use the power cord to wake the PC up from Hibernation mode.

 

No, the power header on the motherboard is a dumb switch. It does nothing more than you do when you connect the two pins with a screwdriver.

The fact that removing the power header plug makes the problem worse confuses me. If anything I thought a short circuit in the switch might be causing the initial problem and removing it would clear that up. But no response from mouse or keyboard without it connected makes no sense.

When you say "Power cord" you mean the 2-pin plug that goes to the power switch from the motherboard's case header, right?

 

Yes, I mean exactly that. When put in Hibernation Mode, how does a keyboard or mouse exactly turn the PC on then? Because from what I've read about hibernation, the PC basically shuts off just like when choosing "shut down', the only difference being that it copies the RAM data to the HDD.

 

In either Sleep mode or Hibernation mode the PC goes into a low power state, it is not OFF, but maintains power on the 5v line. All 12v power is switched off.
Because 5v power is still going to the motherboard the "logic circuit" can be activated from one of a few locations: the power button, the mouse, the keyboard, or the LAN. The settings in Windows and the BIOS/UEFI must be set to allow some of the methods.

You might try checking all your settings to be sure nothing changed without your knowledge.
For mouse and keyboard in Windows go to Device Manager > Mouse or Keyboard device > Properties > and the Power Management Tab, and make sure the check box "Allow This Device to Wake The Computer" is checked.

With modern UEFI BIOS there is not much to do, but older motherboards had Power State settings that needed to be correct (usually S3 state).
But you could check under Advanced > APM in your UEFI BIOS to see if anything obvious stands out.

But I am not aware of the need for the PWR header to be populated in order for the power logic circuit to work. I've never experimented with that one. Maybe someone else in the forum who reads this knows for sure.

 

Before removing the power header I was able to wake up the PC using the keyboard or mouse. After removing it I wasn't. I didn't change anything in the BIOS in between.

However, I now just bought a new power header with a switch (used duct tape to make it not move when pressing the chassis button) and plugged it in the mobo. Until now, waking the pc up 3 times using the keyboard or mouse didn't show any "no signal" problem on the monitor.

Could it be that the one time experience of the PC not waking up with the keyboard when there was no Power header plugged in, just be a mere coincidence? I don't dare to pull out the header to test this again because I'm then forced to boot the PC manually while it's in Sleep/Hibernation mode, in case it doesn't wake up using the keyboard.

 

Could be. But it sounds, based on your experience, that the power header needs to be populated to make the logic circuit work.
At least on your motherboard anyway.

So is all good now? (Other than the duct tape!)

 

Yep, it was a real hassle to stick the switch with duct tape at the right distance in order to let the chassis button press the switch but it eventually worked out. They really have to figure something out for this dilemma.

It was also a huge relief to know that the culprit was the power cord/switch instead of the PSU.

Thanks a lot for your time and help!

 

Glad you got it sorted!

They have figured it out. They make it difficult to repair so you will buy a new case!

 

Not on my watch, I'll let the switch hang out of the chassis and use that as a power button if I have to *Devil

Come to think of it, I don't think it's really coincidence that my PC didn't wake up without a power header plugged in and that my mobo does indeed really use it to wake the PC up in some way. The "no signal" problem was there when the old faulty power header was plugged in, and now that issue is all gone (I woke my PC up 5 times so far without issues) when I plugged a new one.
However, it's still weird that a faulty power header can wake the PC up but not the monitor.

Guess it'll stay a mystery until I test it out again.