My PSU is making my computer restart

[Ewro]

When i try to update my video card drivers my computer restart. I use WhoCrashed and he shows the WHEA_UNCORRECTABLE_ERROR, bugcheck code 0x124, saying he was caused by a hardware problem

I used speedfan to get this specs from my PSU
Vcore: 1,10v
+3,3V: 2,49V
+12V: 12,02V
+5V: 4,71V
CPU VAXG: 2,03W
CPU VRIN: 2,22V
DRAM: 2,22V
Vsb: 3,31V
Vbat: 2,90V


My computer specs
AMD FX-6300 3,5 ghz
8gb ram
AMD R7 370 2G
Windows 10
PCYES ATX ELECTRO 500W (here is a image with the complete specs: https://www.pichau.com.br/media/catalog/product/e/l/ele... )

 

[jboydgolfer]

Do you happen to have another desktop that you can use to troubleshoot and rule out possible causes of the problem?

I never trust software to get voltage readings on anything ,because it's like using a thermometer to see What time it is. It would be great to have another pc to test components instead

 

[EarthDog]

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.

 

[eidairaman1]

When i try to update my video card drivers my computer restart. I use WhoCrashed and he shows the WHEA_UNCORRECTABLE_ERROR, bugcheck code 0x124, saying he was caused by a hardware problem

I used speedfan to get this specs from my PSU
Vcore: 1,10v
+3,3V: 2,49V
+12V: 12,02V
+5V: 4,71V
CPU VAXG: 2,03W
CPU VRIN: 2,22V
DRAM: 2,22V
Vsb: 3,31V
Vbat: 2,90V


My computer specs
AMD FX-6300 3,5 ghz
8gb ram
AMD R7 370 2G
Windows 10
PCYES ATX ELECTRO 500W (here is a image with the complete specs: https://www.pichau.com.br/media/catalog/product/e/l/ele... )

What motherboard?

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

 

[Bill_Bright]

That 2001 testing tutorial is really way out of date, and not conclusive anyway. It does not test under a variety of expected loads, does not measure ripple, and does not specify allowed tolerances.

The only way you can conclusively test your PSU is with an oscilloscope under a full range of loads or with a dedicated power supply analyzer - sophisticated and expensive test equipment typically found only in electronics repair facilities and used by properly trained and qualified electronic technicians.

This leaves most users with the next best option, swapping in a known good spare PSU and see if you have the same problem.

The ATX Form Factor standard allows for ±5% tolerances. Therefore, the acceptable tolerances are:

12VDC ±5% = 11.4 to 12.6VDC
5VDC ±5% = 4.75 to 5.25VDC
3.3VDC ±5% = 3.14 to 3.47VDC
​

So according to Speedfan, your 3.3V and 5V are too low. I would verify with another monitor. I recommend HWiNFO64 but note the shear amount of information it provides can be overwhelming so I recommend you check "Sensors only" when it first starts, then scroll down to your motherboard's section and check the voltages there.

But note it is important to understand software based monitors rely on very low-tech and often inaccurate sensors. So again, we are back to swapping in a known good supply to be sure. Since everything inside your computer case depends on good, clean stable power, this really needs to be done before you spend any money on anything else.

 

[John Naylor]

0124 is the most typical stop error code when ya have a bad overclock ... many other causes, but that's the most prominent.

And yes, if you are going to rely on software, Id suggest HWiNFO ... and follow up with meter readings. Check pinouts,org for where to use ya meter of PSU cables. For example

https://allpinouts.org/pinouts/connectors/power_supply/pci-express-pcie-6pin-power/

 

[Bones]

Is this the one you have?

Although I too have never seen or heard of one of these, doesn't mean it's "Bad" as in a poor quality - However you'd still want to try another PSU and see if the problem goes away or not. Using a meter with the above suggestion(s) is a good way to see what it's doing but access to the plugs while being ran is another deal.

Could be tested with the system itself outside of the case for accessibility to the plugs if it came down to it. However I do suggest at least being prepared to get another unit and don't cheap out on a PSU, it's at the very heart of the system and is probrably the most important piece of the entire build because all else depends on it to even function. Get a good namebrand unit, may cost more up front but will save you $$ and headaches like you're facing with this issue later.

I hope you don't have to replace it but if you do, get a good one and be done with it.

 

[Folterknecht]

Many years ago I had the same kind of problem and in my case it was a partly defective PSU. However as others have pointed out it could also have other reasons. For the regular enduser there sadly is no other option but to cross check with other hardware components, especially when it comes to PSUs.

 

[jsfitz54]

Not only are the 3.3v (way low at 2.49v) and 5v rails low but
the Ram at 2.22v seems way high.
CMOS battery needs replacement.

 

[EarthDog]

What does a CMOS battery have to do with voltage rails and readings? I don't understand the association there.

 

[theFOoL]

Could be just the PSU... Just because it's rated for Broze Sliver or gold doesn't mean it's a good brand

Like my rosewill... Users complain that the brand isn't good but I've had it for 5 years with no issues. All brands are different and use different parts. Since I've never heard of the brand I can only assume it uses bad parts

 

[jsfitz54]

What does a CMOS battery have to do with voltage rails and readings? I don't understand the association there.

Go back to 1st post and read the numbers.

 

[EarthDog]

I read them already. About the only thing I can see there is the battery sitting at 2.9V (its a 3V battery - saw that the first time and dismissed it). I don't see how a CMOS battery will affect voltages coming from a PSU. Will you explain that to me and others reading why changing that battery can resolve the OP's issue please?

 

[jsfitz54]

I read them already. About the only thing I can see there is the battery sitting at 2.9V (its a 3V battery). I don't see how a CMOS battery will affect voltages coming from a PSU. Will you explain that to me and others reading why changing that battery can resolve the OP's issue please?

I'll wait for the OP to reply, thanks.

 

[theFOoL]

Changing a CMOS Battery only effects bios? How that effects psu is beyond me

 

[jsfitz54]

Changing a CMOS Battery only effects bios? How that effects psu is beyond me

It's what you call an independent statement.

CMOS battery needs replacement.

 

[EarthDog]

Oh, there we go...so it has nothing to do with the two sentences directly above it. We appreciate the clarity.

I don't understand how replacing the CMOS battery helps the OP out though...

 

[Bones]

Could be just the PSU... Just because it's rated for Broze Sliver or gold doesn't mean it's a good brand

Like my rosewill... Users complain that the brand isn't good but I've had it for 5 years with no issues. All brands are different and use different parts. Since I've never heard of the brand I can only assume it uses bad parts

It's like anything else - Even the best have their lemons on occasion and the worst in fact do manage to churn out a gold turd at times..... But it's still a turd.

Example:
Imagine my shock when 15+ years later after these came and went here I actually saw a Yugo driving down the road - Not on a towtruck's hook or being pushed/pulled but running under it's own power. Either that or the wind was blowing really hard at the exact spot where it was......

I have examples of both, good ones by name that were crap and one noteable example that's defied all that's been said about it.
Imagine a PSU from way back in 2002 of what was called a craptasic brand (Arrow) that survived from then until now.... I have one that to this very day still works fine and those Arrow PSU's back then were notorious for going out.
I have no worries about leaning on it if I want because I've already done that in the past, namely it's a 500W unit I was running in a Socket A I was clocking up to the moon back then running XP-M chips in these systems.

 

[Bill_Bright]

What does a CMOS battery have to do with voltage rails and readings? I don't understand the association there.

On most (not all, but most) motherboards, when the computer is running or when in standby/sleep mode, the PSU is supplying the necessary 3.3V CMOS voltage to the CMOS circuit to keep the data saved in the CMOS device alive. This is also used to keep the RTC (real time clock) running). The battery only provides voltage when the computer is totally off (unplugged from the wall or when the PSU master power switch is set to "off") to keep that data alive and to keep the RTC ticking.

Even though the demand is low, if the CMOS battery was used 100% of the time to keep the CMOS data alive and the RTC clock ticking, these batteries would not last for the many years they typically do.

However, this still does not apply to the OP's issue.

Like my rosewill... Users complain that the brand isn't good but I've had it for 5 years with no issues.

Users who complain about Rosewill PSUs probably have not done their homework. Like most makers, Rosewill manufacturers several lines (tiers) of supplies. And if folks did their homework, they would see that Rosewill's upper tier supplies are top notch.

That said, as suggested above, until Man can create perfection 100% of the time, there will always be units that don't meet specs or fail prematurely.

 

[EarthDog]

Thank you, Bill, for the explanation. I know how they work as you explained, but simply did not understand how it applied to the thread and help the OP. Seems we were on the same page there.

However, this still does not apply to the OP's issue.

 

[jsfitz54]

However, this still does not apply to the OP's issue.

The OP asked about 1 issue, call it the primary issue.

I found 2 more issues with the data he posted. Ram v and Bios battery.

It's helpful to address all the issues when they are recognized.

 

[EarthDog]

Is it a problem that the CMOS battery is 2.9V? Its a 3V battery, and there are allowed tolerances there too...

I know we are drifting off the 'primary issue' here a bit (just following where led...), but, I personally do not know if that is bad. I am guessing its not as my system here shows 2.9V as well without issue (also through software so who knows if 2.9V is right, or the ram voltage for that matter for the OP or any of us). I am truly trying to get to the bottom of that statement in an effort to learn and share the right things with people.

 

[Bones]

There is no real issue with the CMOS battery, although it's a bit low it's not low enough to cause a problem based on the reading stated.

The battery is 100% independent of the PSU itself so if there is a problem with it then you certainly don't have the replace the PSU too if the battery gets too weak, namely around 2.8v's or less.

He's posted his readings from what Speedfan is showing and it will show the battery voltage alongside the rest but that's all it's doing, you can even see this in the board's BIOS if you want but does not mean they are one in the same because they are not.

I do agree with checking all the variables - In fact it's smart to do so but this is not a CMOS battery issue based on what was posted. The battery itself according to the reading is OK and since it doesn't affect the machine while it's running, clearly it can be ruled out and his issue happens while it's running according to his description of it.

 

[Kursah]

The OP made the thread and hasn't been back, so before anyone continues their disagreements and going too far off course, let's wait to see if they come back and provide any further information.

My two cents on this is simple, trusting software voltage readings can be a dangerous game that should be digested with a side of extreme caution. Also limiting to just one piece of software is equally dangerous, I'd use multiple such as HWInfo, HWMonitor, AIDA, etc., check BIOS/UEFI results, and also I suggest using a multi-meter as is suggested also in the post I have quoted here.

@Ewro please answer the questions below.

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.

 

[eidairaman1]

That 2001 testing tutorial is really way out of date, and not conclusive anyway. It does not test under a variety of expected loads, does not measure ripple, and does not specify allowed tolerances.

The only way you can conclusively test your PSU is with an oscilloscope under a full range of loads or with a dedicated power supply analyzer - sophisticated and expensive test equipment typically found only in electronics repair facilities and used by properly trained and qualified electronic technicians.

This leaves most users with the next best option, swapping in a known good spare PSU and see if you have the same problem.

The ATX Form Factor standard allows for ±5% tolerances. Therefore, the acceptable tolerances are:

​

12VDC ±5% = 11.4 to 12.6VDC​

5VDC ±5% = 4.75 to 5.25VDC​

3.3VDC ±5% = 3.14 to 3.47VDC​

​

So according to Speedfan, your 3.3V and 5V are too low. I would verify with another monitor. I recommend HWiNFO64 but note the shear amount of information it provides can be overwhelming so I recommend you check "Sensors only" when it first starts, then scroll down to your motherboard's section and check the voltages there.

But note it is important to understand software based monitors rely on very low-tech and often inaccurate sensors. So again, we are back to swapping in a known good supply to be sure. Since everything inside your computer case depends on good, clean stable power, this really needs to be done before you spend any money on anything else.

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'm the exception because I'd rather not have to drive in Houston Traffic @cdawall knows this lol. So once the op trys the stated above if he comes back the topic is cold.