My PSU is making my computer restart

[bug]

The only thing to watch out for here is that a bad PSU will tend to take some other components with it when/if it decides to go belly-up.

 

[Bill_Bright]

The Link I posted still applies today because not everyone has a Shop Locally that even have load testers or O-Scopes, and most end users that ask these kind of questions are not going to buy equipment they may only use once or even try to interpret the readings.

I agree. That's why I said the next best option is to swap in a known good PSU.

I note most users don't have a multimeter either. And if they did, they most likely would not know that to test properly, the PSU MUST be under a variety of realistic loads, from idle all the way up to maxed out. And most multimeters don't test for ripple.

• That article makes no mention of testing under a variety of loads. A PSU can work fine at idle then flake out under load. In fact, that article says to short two pins on the main power connector which would mean the PSU is operating with no load at all. That NEVER was a proper way to measure ANY power supply - from a battery to switching supply to a car engine. All that does it tell you the PSU is capable of turning on. It provides nothing conclusive about the actual voltages.
• That article makes no mention of 3.3VDC. On many motherboards this voltage powers the chipsets and RAM. So it is critical, but not mentioned.
• It makes no mention of allowed tolerances either. So how is the user supposed to know that 11.6VDC or 5.14VDC are still acceptable, but 11.3V is not?
• And or course, excessive ripple can cause a PSU to be unstable, or cause instability across the motherboard too. But there's no mention of ripple.

So I am sorry, but no, it does not apply today.

BUT most do users have access to a second computer, either in the same household or with a trusted friend or relative they can swap out power supplies with. If not, then they really have no choice but to buy a spare PSU, or find a shop that will properly test it. Otherwise, they risk replacing another component that was not bad - it was only getting bad power.

 

[jsfitz54]

because not everyone has a Shop Locally that even have load testers or O-Scopes, and most end users that ask these kind of questions are not going to buy equipment they may only use once or even try to interpret the readings.

I have said the same thing in the past and I agree.

The last thing to do would be to pay a shop for O-scope diagnosis.
It's often not practical, not available and not cost effective. If you can get warranty coverage they are going to test it for you.
Buying a low cost Digital Multi Meter, can be a helpful tool to have for uses other than computer testing use.
Testing batteries of all sorts, easily comes to mind.

My approach in prioritizing these responses is:
1) check bios for V numbers if machine is bootable
2) try software, while in OS.
3) try different software, while in OS.
4) if numbers seem off = go to digital multi meter for more accurate numbers.
5) no meter or meter confirms bad numbers = swap out with known good PSU.
6) buy new PSU and / or RMA warranty covered PSU.
LAST
7) have shop diagnose bad PSU when other less economical options don't exist. Can they fix it or are they going to recommend a new one?
I will never do this, as I would put the fee towards a new PSU.
Anyone asking here is pretty much a do it yourself-er.

If you can buy a new PSU and RMA the bad one you can often avoid down time.
This leaves you with a new PSU and a spare PSU or you can sell the Warranty one, when it returns. I have a spare currently from just such an occasion.

 

[eidairaman1]

I have said the same thing in the past and I agree.

The last thing to do would be to pay a shop for O-scope diagnosis.
It's often not practical, not available and not cost effective. If you can get warranty coverage they are going to test it for you.
Buying a low cost Digital Multi Meter, can be a helpful tool to have for uses other than computer testing use.
Testing batteries of all sorts, easily comes to mind.

My approach in prioritizing these responses is:
1) check bios for V numbers if machine is bootable
2) try software, while in OS.
3) try different software, while in OS.
4) if numbers seem off = go to digital multi meter for more accurate numbers.
5) no meter or meter confirms bad numbers = swap out with known good PSU.
6) buy new PSU and / or RMA warranty covered PSU.
LAST
7) have shop diagnose bad PSU when other less economical options don't exist. Can they fix it or are they going to recommend a new one?
I will never do this, as I would put the fee towards a new PSU.
Anyone asking here is pretty much a do it yourself-er.

If you can buy a new PSU and RMA the bad one you can often avoid down time.
This leaves you with a new PSU and a spare PSU or you can sell the Warranty one, when it returns. I have a spare currently from just such an occasion.

Trust me I use a Multimeter, even trouble shot a capacitor using reaistance since it didn't have microfarads as an option.

@Bill_Bright bright that link is for the common joe, not the Techies.

 

[jsfitz54]

Trust me I use a Multimeter, even trouble shot a capacitor using reaistance since it didn't have microfarads as an option.

Do you have a favorite DMM one / brand?

 

[Ewro]

What motherboard?

Try this
https://answers.microsoft.com/en-us...8/0df6c858-d0a5-4f4d-8e13-88191b23bd1e?auth=1

My motherboard is a gigabyte ga-78lmt-usb3 6.0

That PSU may be a hunk of garbage. Never heard of the brand, not sure of the OEM and that concerns me.. It has a combined 40A on the 12V rail though so it should handle that PC without issue assuming it isn't malfunctioning and can handle its rating in the first place.

Updating drivers is not a very CPU intensive activity. The code though, the 0x124, is typically a low vcore thing (many see this while overclocking). So perhaps consider raising the voltage .05V and see if that stops it.

Are you at stock or overclocked?
Does it crash at any other time? Or only when updating drivers?
Does it crash when you stress test your CPU?
How about when you stress test the GPU? Both of those will use more power than updating drivers will. So if you can pass those tests, you know its something else.

Software readings can be inaccurate so its best to test with a multimeter where possible. That said, 5V and 3.3V are low according to that... check it out.
https://www.techpowerup.com/forums/threads/the-psu-rail-checking-guide.147571/

EDIT: System specs would be good as well.

I never overclocked my computer
The crash happens only when i update my video card drivers, i can use my computer the whole day without problems.
I will do a stress test for my CPU and GPU, and edit this post later

 

[jsfitz54]

I never overclocked my computer
The crash happens only when i update my video card drivers, i can use my computer the whole day without problems.
I will do a stress test for my CPU and GPU, and edit this post later

Those 3.3v and 5v numbers are bad.
I could overclock my CPU fine with a bad Corsair HX850 but it crashed in Steam. Slowly kept getting worse to where I could force it to repeat the error.
Thought it was Drivers or OS Update or AntiVir. NO, NO, NO, bad PSU.
Your DDR3 memory should be at 1.65v or (1.5v if low v type memory.) I assume if not overclocking and just running 2 mem sticks @ 2 x 4Gb or 1 x 8Gb.
Your CR2032 3.0v battery is boarderline at 2.9v and recommend replacing it.
Lastly, are you running the newest Bios F2: https://www.gigabyte.com/Motherboard/GA-78LMT-USB3-rev-60#support-dl

 

[Mr.Scott]

Pffft....that battery is fine. I've had less than 2.9 brand new out of the package before.

 

[jsfitz54]

Pffft....that battery is fine. I've had less than 2.9 brand new out of the package before.

Brand new OLD STOCK?

 

[Mr.Scott]

I don't stock the place.
I walk in, take a couple packs off the rack, pay and leave.
Sometimes you get one that's less than 3v. That's the way it goes...........anywhere.
They don't have expiration dates.

 

[EarthDog]

My motherboard is a gigabyte ga-78lmt-usb3 6.0

I never overclocked my computer
The crash happens only when i update my video card drivers, i can use my computer the whole day without problems.
I will do a stress test for my CPU and GPU, and edit this post later

Ok. So... the primary problem...

If you have a PSU to swap out, I would try that first. We can't confirm the software readings, so that is really the only easy way is to remove it and see.

Another quick question.. the same driver or different versions all cause this? Are you installing these via download of the driver package from AMD themselves or through another source (even mobo website, or their update software). If so have you tried another source and method of installation?
Does this happen in gaming?
The voltage readings, when idle, or at least not installing drivers, what is its minimum and maximum (Hwinfo 64).. if you can game and stress test, what is the minimum and max there?

EDIT: That wasn't a quick question... and its going to get longer.. ha!

If the CPU and GPU stress tests can pass and you tried multiple driver packages from different install methods, I'd look at software. Do you have a spare drive or can make a 50GB partition on a non OS drive and install windows there (unplug all drives except for drive to you installing windows on) and then see if you can update the drivers? That should tell you if its rooted in software or hardware...

 

[Bones]

2.9V's is fine as I said earlier.
I've seen them still hold settings and all worked fine with 2.8v's or even less before but I (Personally) use 2.8 as a rule to know when the battery is getting low, esp for my DD machine. The battery only has one job and that's to hold the BIOS info, nothing else.

The issues he's experiencing happen AFTER the machine has booted, not during bootup.

However.....
There is one exception to this, the Abit AN7 CAN be affected even after it's running by a weak CMOS battery but that's only because the UGuru chip made into the board eats them for breakfast and asks for more and directly affects the system while it's running, can cause all kinds of whacky things to happen when the battery gets weak with those....
But we're not talking about an AN7 here.

The battery as reported is OK.

 

[Mr.Scott]

However.....
There is one exception to this, the Abit AN7 CAN be affected even after it's running by a weak CMOS battery but that's only because the UGuru chip made into the board eats them for breakfast and asks for more and directly affects the system while it's running, can cause all kinds of whacky things to happen when the battery gets weak with those....
But we're not talking about an AN7 here.

The battery as reported is OK.

It's actually any Uguru board, but I know what you meant.

 

[eidairaman1]

Do you have a favorite DMM one / brand?

Fluke or Klein

Sequence, PSU, GPU, Software at this point

 

[Ewro]

One more thing i forget to say: im using the Microsoft basic videodisplay adapter so i cant use 3dmark or furmark to a GPU stress test.
I used prime95 for a CPU stress test but my computer just froze by 36 minutes until i made a hard reset, i let open HWINFO in the motherboard session so i got these numbers: results

 

[EarthDog]

So, is this something new? You never had drivers (please answer the questions in my last post )? Did you ever game on this? Its weird, this just 'happened'...

That said if it is failing at stock speeds, that isn't a good sign for anything. I would swap that PSU out with a known good and see if that is the culprit.

 

[Ewro]

So, is this something new? You never had drivers (please answer the questions in my last post )? Did you ever game on this? Its weird, this just 'happened'...

That said if it is failing at stock speeds, that isn't a good sign for anything. I would swap that PSU out with a known good and see if that is the culprit.

I used my computer to play games until I had some problems with my computer (probably for some update of Windows 10). After a few Windows installations, I was able to fix my computer , but this error appeared
edit: i tried download some older drivers from this site, but still getting the same error in the installation.
I tried with the AMD program, Windows Update, and even with DriverEasy in safe mode (he installs normaly, but when I restart my computer, he reach the Windows logo then restart, and starts to loop until I go to safe mode and disable the driver)

 

[theFOoL]

Shouldn't it been convinced that the PSU is a bad unit or brand? Windows should not interfere with the PSU going all crazy

 

[EarthDog]

Shouldn't it been convinced that the PSU is a bad unit or brand? Windows should not interfere with the PSU going all crazy

......

I used my computer to play games until I had some problems with my computer (probably for some update of Windows 10). After a few Windows installations, I was able to fix my computer , but this error appeared

But at the same time, it failed during the CPU test as well... so who knows. Im leaning towards power as is everyone else, but, I'm not quite convinced.

I still say swap the PSU or if its easier to install a temp copy of windows somewhere, do that as instructed (no other drives attached upon installation). Installing through USB, and installing chipset, ME, and GPU drivers shouldn't take but 30 mins on an SSD. My USB 3.0 image takes 22 minutes and I'm up with all my basic apps.

 

[Ewro]

Thanks for the answers, i'm gonna buy a new PSU, thinking about a Corsair CX 550 550W, looks good and the price is fine.

 

[EarthDog]

Thanks for the answers, i'm gonna buy a new PSU, thinking about a Corsair CX 550 550W, looks good and the price is fine.

its serviceable... as is your current psu.

I wouldnt cheap out on a power supply. My go to is evga 550w g2 or g3.

 

[jboydgolfer]

Thanks for the answers, i'm gonna buy a new PSU, thinking about a Corsair CX 550 550W, looks good and the price is fine.

What is your psu budget?

I wouldnt cheap out on a power supply

I cant stress enough how true this is