Set VCore Voltage Too High During Overclock Did I Do Damage?

SquirrelMadness

SquirrelMadness

SquirrelMadness

Associate
Joined
30 Aug 2014
Posts
705
I accidentally set the vCore voltage to 1.8v @ 4.4GHz when it was supposed to be 1.28, hit enter-save-exit and the system crashed at the BIOS with a "Overvolt Error: Press F1 To Exit". I thought "What overvolt?" When I knew I was setting it lower (was originally 1.29v) so immediately went back into BIOS and changed it.

The whole process lasted maybe 10 seconds.

I set everything back to default... lowered clock speed to 3.5GHz and voltage on auto.

Booted into system and ran Aida64... now every time I get "Hardware Error".

So far system appears to be running normally but I'm too scared to try some games or intensive applications fearing crashes lol

What are the chances of actual damage? Is it something that will resolve with time... f.e. maybe I can turn the PC off for a while and it will calm down?

Here are some random screenshots taken during stress test:

QFuehnE_d.webp

JiLLTOL_d.webp

SKQgHx7.jpg

ModhJ2m_d.webp

Thanks in advance.
 
I've read accounts of people doing damage to 22nm CPUs with 1.8v doing suicide runs so I wouldn't be surprised if you've done some damage to it. I've done up to 1.6v on 22nm for brief benchmarks on suicide runs without seeming to damage them though (not on any CPU I'd want to keep).

If it isn't passing stability tests either the settings are wrong or you have done some damage.
 
OK I played 20 mins of Kingdom Come Deliverance (a game that my CPU bottlenecks the graphics card on) and CPU usage didn't hit more that 50%... which is weird because usually hits 100%.

I couldn't tab out of the game (had to ctrl+alt+delete out) which is unusual.

I took a screenshot and wanted to open Photoshop to crop it while the game was in the background... Photoshop gave an error and wouldn't launch.

I launched Chrome while game still in background and it crashed twice.

I quit the game and Chrome now launches fine.

It seems like the processor has trouble multi-tasking. As if the voltage spike burns some cores but not others?

Is there a way I can check every other component (by stress testing etc...) but NOT the CPU, so I can make sure all other components are fine?

Would it even be possible for a vCore overvolt to damage any other components?
 
Is vcore supposed to be that low during stress test, 1v seems like to little.

Which CPU is it btw and how long have you had it?
 
JohnG7 said:
One of the other voltages is the real killer if too high (cannot remember if it is VTT or CPU VRM or not).
Click to expand...

For 24x7 use high levels of VTT will kill a CPU quicker than high levels of VCore though that is truer of some generations more than others - Sandy Bridge and Ivy Bridge CPUs a lot more susceptible to it than others.
 
Yes that rings a bell from Haswell days) thanks, pity is it sometimes more VTT not vCore (you can actually lower this) to get stable.
 
JohnG7 said:
Yes that rings a bell from Haswell days) thanks, pity is it sometimes more VTT not vCore (you can actually lower this) to get stable.
Click to expand...

TBH not sure what actually kills them - this was based on people back in that era who were running 100s of these CPUs overclocked in render farms, etc. they found the ones that needed higher than normal VTT to get the same overclock were dying quite quickly while those that needed higher than normal VCore weren't dying any sooner than average. Whether that is actually the VTT that is killing them (and the same would be true of CPUs that didn't need as much VTT if you gave it to them anyway) or some factor of CPUs that manifest as requiring higher VTT I don't know.
 
OK things just got worse...

I realised the memory stress test in AIDA also has hardware failure.

I updated BIOS to lates version and set everything to factory except RAM (set that to XMP)

Tried a few RAM configs aka different chips, different slots, one chip at a time etc... AIDA stress test always fails.

So now we have FPU, cache, and memory tests failing.

Used Intel Diagnostics and got this:

BFlvvQG.jpg


Windows memory test also passed. I bet memtest86 will be a pass too so won't bother with it.

So now I'm having second thoughts on ordering a new (used) CPU?
 
It could have fried the memory controller in the CPU or similar. Try booting a Linux live drive and see if that's stable, if not I'd seriously recommend trying a different CPU
 
Hyburnate said:
It could have fried the memory controller in the CPU or similar. Try booting a Linux live drive and see if that's stable, if not I'd seriously recommend trying a different CPU
Click to expand...

Probably fried a voltage controller/regulator and/or components related to that on the CPU - possibly part of the uncore somewhere with unpredictable results.
 
I know you never got past the BIOS but am assuming you've been back and forth with voltages and (potentially) BSODs whilst looking for stability? If so, it's also worth scanning your Windows install for corrupted files. Not trying to give you false hope but it's worth a shot.
 
Rroff said:
Probably fried a voltage controller/regulator and/or components related to that on the CPU - possibly part of the uncore somewhere with unpredictable results.
Click to expand...

Are any of the components on the MB? I really hope that it is the CPU because the MB has the highest resale value (Asus Rampage V Extreme). I will be looking for a replacement cheap CPU on eBay but I can't replace the MB as the cheapest x99 is over £100.

sodaoat said:
I know you never got past the BIOS but am assuming you've been back and forth with voltages and (potentially) BSODs whilst looking for stability? If so, it's also worth scanning your Windows install for corrupted files. Not trying to give you false hope but it's worth a shot.
Click to expand...

Don't have to do that, I have a second Win 10 install on a separate SSD where I test my software so will try that and report back.
 
You must log in or register to reply here.
Back
Top Bottom