Question CPU Voltage Overload - - - PC keeps freezing ?

[Dreamstreet]

My desktop through nearly 10 years now have started to be more or less broken. Every time I turn it on, it says CPU Voltage Overload. When I get to bypass it, my PC can freeze from the login screen to 2-3 hours later. So far I've tried removing dust, adding new cooling paste to the CPU and unplugging and replugging all the cables, but it persists. Add on top of this, sometimes the graphics card won't turn on when I turn on the PC, so while the monitor thinks it's connected, nothing is displayed.

My theory is that it's either because of the PSU, the motherboard or the cabinet which is pretty <bad language removed> at this point, might interfere with the current inside of it.

My setup goes like this:

CPU: Ryzen 7 5800x
CPU Cooler: Noctua NH-D15
Motherboard:ASUS TUF B450-PLUS GAMING
Ram: Corsair Vengeance LPX DDR4-3200
SSD/HDD: 1TB M.2 SSD, 1TB HDD, 256 GB SSD
GPU: MSI RTX 3070
PSU: Seasonic Focus+ 750W 85+ Gold
Chassis: Corsair Carbide 100R but it's nearly 10 years old
OS: Windows 10 currently
Monitor: HP X34 monitor

Hope anyone can help me with this.

 

[Lutfij]

PSU: Seasonic Focus+ 750W 85+ Gold
How old is the PSU in the build? Have you tried working(sourcing, borrow, not buy) with another PSU that's reliably built and has a little more power at the entire PC's disposal?

 

[TerryLaze]

Can you go into bios and look at the voltage settings?
If you can upload a pic or two to show it would be even better.

With the latest news about asus using higher voltage settings than others it might just have degraded your CPU.

If the PSU didn't have enough power it would always crash at the same time, either when getting into windows, because that's when there is high power draw, or when starting some game or workload, because then there is even higher power draw. The PSU would cause a crash when it couldn't provide the needed power not at random moments. It could still be the PSU if it's completely shoddy.

 

[Dreamstreet]

Can you go into bios and look at the voltage settings?
If you can upload a pic or two to show it would be even better.

With the latest news about asus using higher voltage settings than others it might just have degraded your CPU.

If the PSU didn't have enough power it would always crash at the same time, either when getting into windows, because that's when there is high power draw, or when starting some game or workload, because then there is even higher power draw. The PSU would cause a crash when it couldn't provide the needed power not at random moments. It could still be the PSU if it's completely shoddy.

I'll upload a picture when I'm at my PC again. Whenever the CPU voltage overload happens, I go into BIOS and see that it's at around 1.730v and CPU temp is at around 73 degrees celcius. Lately the voltage has been down to around 1.450v instead with 63 degrees celcius.

 

[Dreamstreet]

PSU: Seasonic Focus+ 750W 85+ Gold
How old is the PSU in the build? Have you tried working(sourcing, borrow, not buy) with another PSU that's reliably built and has a little more power at the entire PC's disposal?

The PSU is from March 2018 and has been in use pretty much daily since. I'll see if I can find a way to find another PSU to test it out, but it might be the PSU that is getting too old.

 

[zx128k]

My desktop through nearly 10 years now have started to be more or less broken. Every time I turn it on, it says CPU Voltage Overload. When I get to bypass it, my PC can freeze from the login screen to 2-3 hours later. So far I've tried removing dust, adding new cooling paste to the CPU and unplugging and replugging all the cables, but it pesists. Add on top of this, sometimes the graphics card won't turn on when I turn on the PC, so while the monitor thinks it's connected, nothing is displayed.

My theory is that it's either because of the PSU, the motherboard or the cabinet which is pretty <bad language removed> at this point, might interfere with the current inside of it.

My setup goes like this:

CPU: Ryzen 7 5800x
CPU Cooler: Noctua NH-D15
Motherboard:ASUS TUF B450-PLUS GAMING
Ram: Corsair Vengeance LPX DDR4-3200
SSD/HDD: 1TB M.2 SSD, 1TB HDD, 256 GB SSD
GPU: MSI RTX 3070
PSU: Seasonic Focus+ 750W 85+ Gold
Chassis: Corsair Carbide 100R but it's nearly 10 years old
OS: Windows 10 currently
Monitor: HP X34 monitor

Hope anyone can help me with this.

Enter BIOS and see what the voltages are, there should be an area for monitoring. Example of doing that here. Could be a sensor issue or RIP PSU.

 

[TerryLaze]

I'll upload a picture when I'm at my PC again. Whenever the CPU voltage overload happens, I go into BIOS and see that it's at around 1.730v and CPU temp is at around 73 degrees celcius. Lately the voltage has been down to around 1.450v instead with 63 degrees celcius.

If these are real numbers and not some kind of false measurement then it's a wonder that your CPU still works.
If voltage is set to auto then you should turn that to manual and dial it way down.

 

[Dreamstreet]

If these are real numbers and not some kind of false measurement then it's a wonder that your CPU still works.
If voltage is set to auto then you should turn that to manual and dial it way down.

I'll lower it, what should I set it to? Also do you think it's a case of the PSU being faulty and needs to be replaced? It is 5 years old at this point.

 

[zx128k]

Need to test the voltage with a multimeter to be sure but that voltage is not a good sign. You can buy a PSU tester on amazon for example, you need a good one. The sensor that reads the voltage could have failed.

Readings should not have variances larger than +/- five percent.

Maximum.........Minimum

12.6V.................11.4V
5.25V.................4.75V
3.47V.................3.14V

Also check and see if any caps on the motherboard have tops swelled, popped open or there is leakage.

1.730v and CPU temp is at around 73 degrees celcius. Lately the voltage has been down to around 1.450v instead with 63 degrees celcius.

You cant run the PC without finding out if there is a voltage issue. If you have a spare PSU that works, then swap to it and test. At best there is a fault with the sensor reading the voltage and you just need better thermal paste. At worst there is too high voltage going into the CPU and there is a hardware fault.

If you can't test yourself, take this PC to a repair shop for testing.

Be aware that you could replace the PSU and still get the high voltage readings. A decent repair shop will be able to read the real voltages from the motherboard with a multimeter and work out whats happening.

View: https://youtu.be/dTaSurfpldg

View: https://youtu.be/b3nyNS-PyDU


If you really had 1.7 volts into the cpu it would be already dead.

 

[Dreamstreet]

If these are real numbers and not some kind of false measurement then it's a wonder that your CPU still works.
If voltage is set to auto then you should turn that to manual and dial it way down.

No idea why it gets so hot while just being in bios.

 

[zx128k]

Good news is your Voltages from the psu look fine. That vcore of 1.744 volts looks bad.

The vcore voltage is just above all those settings in the bios. Also need a screen shot of vcore and the LLC setting.

Try resetting your bios to defaults and see if that changes the core voltage reading.

 

[Dreamstreet]

Good news is your Voltages from the psu look fine. That vcore of 1.744 volts looks bad.

The vcore voltage is just above all those settings in the bios. Also need a screen shot of vcore and the LLC setting.

Try resetting your bios to defaults and see if that changes the core voltage reading.

Thanks, I'll try that out. Do you perhaps have some ideas why this is happening or where the issue could come from?

 

[zx128k]

Thanks, I'll try that out. Do you perhaps have some ideas why this is happening or where the issue could come from?

Basically if the CPU is autoing 1.744 voltage at stock BIOS setting, its faulty. It could be reported as a high but wrong value. Even so you can't run a CPU that could be intentionally hitting 1.744 volts when boosting.

Most people have screenshots like this from reddit what using PBO2. Maximums in the 1.488-1.5 volt range. People post 0.2-1.5 as the normal range they see. You get the higher voltage on one core, with the all-core load voltage being a lot lower. 1.7 volts is not normal. You can open a ticket about it with AMD but I think they will think the cpu is RIP.

If we are lucky there is a bad setting in BIOS and the CPU is okay. Just don't run a heavy load because 1.7 volts will kill the cpu if it doesnt drop to safe levels.

 

[Dreamstreet]

So a bit of a weird update on this. Not sure if it's a case issue, but when the PC is standing up, the PC still freezes and over voltages the CPU (though less frequent now after some BIOS changes). However if I lie the PC down, it doesn't have a single issue. I got no idea why, if it's because the cabinet is nearly 10 years old or if it's something else.

 

[TerryLaze]

So a bit of a weird update on this. Not sure if it's a case issue, but when the PC is standing up, the PC still freezes and over voltages the CPU (though less frequent now after some BIOS changes). However if I lie the PC down, it doesn't have a single issue. I got no idea why, if it's because the cabinet is nearly 10 years old or if it's something else.

Have you checked the cooler? Basically if some of the screws holding it down are broken or not tight enough it will not "stick" to the CPU and only make partial contact if it's upright, but if it's laying down gravity keeps it connected enough.

 

[zx128k]

Sounds like bad mounting pressure when standing up.

 

[Dreamstreet]

It it a really heavy CPU cooler, but it seems to hold on tightly without moving. Maybe the fans move since they are detachable with clips on both sides, otherwise I'm not too sure.