Motherboard won't power on. Completely dead. Tried just about everything.

jonohtin

Distinguished
Nov 20, 2011
12
0
18,510
About a week ago my PC just completely shut off and died. One of my usb cables connected to the front panel of my PC somehow managed to get itself near my surge protector power strip and I believe touched one of the ports. There was a flash of light right before my PC dying and afterwards it would not turn on anymore. The end of my USB cable is entirely melted and charred with charred marks on the power strip.

So I opened up my PC to check inside, no burning smell. However when pressing the power button absolutely nothing would happen, no lights, no fans whirring, nothing. So I started trouble shooting.

First thing I did was check the PSU. PSU seemed fine, passed the paperclip test, had good voltage readings with a multimeter. But knowing that it might not perform well under load I used my GF's PSU (which works perfectly) and still no power to my PC.

I began breadboarding my motherboard with just the CPU, CPU fan, 8 pin connector + 24 pin connector. I attempted to short my mobo into starting using a flathead screwdriver + the two power start pins and still nothing. No lights no fans no beeping just dead quiet. So far not once has my motherboard given any indication that it was receiving any power.

So now I assumed that my motherboard was dead and sent it in to warranty/RMA to receive a replacement or get it repaired. A week later they sent a "replacement" motherboard and after breadboarding it and it not powering on again, I found out that they sent the same exact motherboard I had sent in. Same serial number and everything. Unfortunately I can't call them as they're closed to ask why they sent me the same one.

Have I missed anything? What else could be the problem? Or is it more likely that the warranty services just accidentally sent in the same motherboard?
 
Is there any chance you are plugging the PSU directly into the wall when your are testing the PSU, but then using the motherboard to try and power up the motherboard, or are you plugging the PSU directly into the wall socket?

Power strips are a bad idea to use for powering your PC power supply anyhow. It should always be plugged directly into the wall socket or into a UPS unit.

If that can't be the issue, then I'd contact them on Monday and say WTH is going on? You guys sent me the same unit back I sent you so what's the deal? If you log into your account with them, there may be an answer listed there that you simply didn't receive notification of. I've had this happen several times when dealing with RMA.
 


I'm not entirely sure I follow 100% with your PSU question. But I have plugged directly into both the wall + power strips for both testing + breadboarding.

I should also mention that I have tried resetting the CMOS battery as thats not included in the original post.

Is it possible that my CPU is just dead/fried? And thats why my mobo's showing no signs of life? I would assume that if my CPU was the culprit my mobo would atleast get some power .

But yeah i'll definitely be calling them first thing in the morning.
 
check the warranty return package. usually they give you a slip describing what they did and solution. maybe the board was returned to you because 1. it wasn't defective or 2. it cannot be fixed and youre out of warranty.

it can be possible that the electric circuit in your house is overloaded beyond capacity and your psu is not getting enough power from the wall.

what motherboard is this?

 
What I was trying to ask was whether you were testing the PSU plugged into the wall with the multimeter and paperclip, but then plugging into the power strip while trying to power the system up. Clearly, it would seem that you are not as you've tried both ways.

If the PSU powers on in testing when not connected to the motherboard, but when stripped down to the essentials the board shows no signs of life, then it can only be the motherboard or CPU. Usually, if the CPU is bad, and the motherboard is not, you will at least get something. Beep code, lights, fans, something. But if the motherboard is dead you MIGHT get something, and you might not.

I'm still not too sure how a USB cable could just find it's own way into an outlet, but I guess that's really not relevant at this point since apparently it did, somehow. LOL.

Have you tried connecting power to the motherboard WITHOUT the CPU installed? Do you have memory installed? It's not a sure shot, but I've seen boards that wouldn't do anything without any RAM installed. This could go either way with the MB or CPU. You may need to throw some parts at it depending on what the manufacturer tells you that they found or didn't find.
 


If that were the case, then it would not test out fine using the multimeter. If there is enough power to satisfy a paperclip test and voltage readings are good, then there should be enough power to get some kind of reaction from the board when attempting to power it up.

Might also want to triple check that you are jumping the correct pins.
 
well, if you think nothing wrong with the wall plug and your psu, then it could be the motherboard startup circuit. the modern boards have some type of "soft switch" that keeps the system on/off states, and maybe somehow its been stuck in the "off" state.


if you haven't already, try to do a full cmos reset by removing the cmos battery, wait a few hours or clear the power by pressing the power button (which seems not working). so that means remove the battery, wait a few hours for electric charges to clear, and then try turn it on again
 


well for example when you press the power button while your computer is up and running, which creates a short between the two pins, this signal is either ignored or picked up by the os/bios. for windows it picks up the signal and do a proper shutdown.

if your windows is frozen, bios/uefi typically ignores the signal but if you keep pressing the button for a few more seconds, bios/uefi will recognize the signal and shut the computer down. this is commonly referred to as "forced shutdown"


so in effect shorting of the power pins creates a signal, some component along the way picks up the signal and do what is programmed/instructed to do
 
Well I'll try the CMOS battery thing for a couple of hours. Couldn't hurt to try.

And yeah there was a slip of paper describing the RMA process. Literally one line: Model In (serial number). Model Out (same serial number) No mention of repairs, motherboard working fine, its not defective, etc.

The boards definitely still under warranty as the guy I called checked and its under warranty until next year. I'll call in a few hours and find out what's going on with this mobo.
 
what's the brand of the motherboard?

I'd expect RMA slips to atleast indicate what work was done or not done on the returned unit.


to sum up my point, the power button mechanism is handled by logic in CPU or a microcontroller on the motherboard (or a combination of both), so the easiest way to isolate the issue is to try a different CPU. If your CPU was fried then another CPU would power up with your existing mobo. if it wont then its probably the chip on your mobo that broke it, in that case you'll want to RMA or get a different mobo.

If the warranty folks returned your mobo because they tested it and it worked (with their cpu), that means your cpu may be a bit fried. There was one incidence I encountered similar to yours, and what I did was simply unplug the atx/cpu power cables on the mobo, plug them back in, that seemed to have "reset" something and woke up the power-on mechanism.





 
Yeah, I fully understood what you meant about the CMOS battery and the power button. The thing that hurts my brain is the fact that any residual energy still present in the motherboard or power supply, THAT CAN BE DISSIPATED by pushing the power button, is going to do so within about thirty seconds.

Removing the CMOS battery for a bazillion hours, isn't going to do anything that removing it for two minutes won't accomplish either.

So, remove CMOS battery for two minutes. During that time, press the power button for 30 seconds. Then put it all back together.


The more important thing is, this motherboard has already been to the OEM to be checked out, so it has already been initialized and had the latest bios installed if it didn't already have it. I've seen them NOT do this when somebody has sent an RMA in that they couldn't find anything else wrong with. Plus, since the system isn't installed in the case, it's kind of hard to push the power button. He'd need to simply connect the two pins by way of jumper or screwdriver, for a few seconds, which is unlikely to produce anything helpful in this case anyhow since it is not a configuration or compatibility issue, but a failure to power on or show signs of life. Clearing the CMOS isn't going to change that.

Knowing the model of motherboard however, might be helpful.

 
True, but remember pressing his power button isn't getting any circuits "jump started" on his motherboard, so I doubt press it will dissipate any electrical charges in his special circumstance. :^)



 
My motherboard is MSI z170A gaming m5.

I called the RMA department and they told me that they tested the board and it powered on just fine so they sent it back to me. Based on this the only thing I can conclude is that my CPU (i5 6600k) is fried.
 
If the CPU is fried, there must be more to the story than what you've explained. There's no way a stray USB cable could accidentally touch a power strip, and fry your CPU. Especially without also ruining the motherboard. In fact, it would be far more probable that the motherboard would be damaged and the CPU would still be fine.

I can't say it's not POSSIBLE, because practically anything is possible when it comes to electronics, but in maybe close to 30 years of working on and with computer systems and other electronics, I've never seen or heard of anything like that happening. There must be more going on than what we know about. I guess your option at this point is to find a used Skylake CPU and see if it works. Or upgrade the whole platform. I don't really see any other options.
 
Or test another CPU in that motherboard, which would probably be a much better option since the motherboard OEM already tested the motherboard and said it is ok, and putting a bad CPU in a good motherboard could result in a bad motherboard if something is directly shorted or otherwise haywire.
 
Well thats whats boggling my mind. Everything I've said is entirely what happened. I was fiddling with my headset wires (as they get tangled up a bit) and my iphone usb cable happened to somehow touch my power surge strip and there was a flash of light and PC just died. My iphone usb cable's entirely melted as well.

The part that really bugs me is like you said I can't safely test my CPU on another good mobo. I'll be sending the CPU in to try and RMA but I got tired of not having a PC and went ahead and just buy/upgrade a new mobo + cpu. i7 8700k + a msi z730.
 


well, well. you tested the cpu on your motherboard, right? and the motherboard was sent to service center, where it was tested to be functional. so appareantly your cpu did not damage your current motherboard, if what the service center said was true.

the service centers do place a cpu into your mobo when they do testing, as far as I know. otherwise, how could they have tested it?
 
They do not test that way. I've sent in I don't know how many customer boards that had glaring, specific issues, but would power on and POST, and practically all of them came back initially saying they were ok. I have had to send many boards back twice, in some cases escalating the RMA up the chain of command to higher tiered support channels.

Some manufacturers are better than others when it comes to the RMA process. Gigabyte and MSI are hard to deal with when it comes to RMA. ASUS, Biostar and ASRock are easier to deal with, and are more responsive to actual and specific issues. At least, in MY experience in having to deal with all four of these companies. Supermicro is pretty good. I've never had to RMA an EVGA, Sapphire, NZXT or Zotac board. Intel is the worst and they'll try everything they can to not have to replace your hardware. Since the majority of Intel boards are used in OEM builds or workstations, they will practically always try to shift the process over to those systems builders, which is right sometimes, but other times is not if the machine is out of warranty but the board is still within what the warranty would be if you had bought the board outright. They don't honor that on those kinds of systems though.

Even if you bought an Intel board in the retail sector, outright, they still suck on trying to get them to accept that a board is bad and not something else. I wouldn't buy an Intel board unless it was the only option still available, as it is on many older chipsets.

Anyhow, at least you have something to work with now, but it's disappointing since the newer 9th gen CPUs are likely to be released next month as well as a new series of motherboards and chipset to go with it. Still, those CPUs will likely work with your current board, supposedly, but it's doubtful you'll want to replace yours so soon. What sucks about that is that these new CPUs are supposed to go back to soldered TIM between the die and heat spreader, rather than using thermal compound like the last few generations have used. That should allow these CPUs to operate at MUCH cooler, more even temperatures, and increase any overclocking headroom substantially.