Can a overheating CPU shut down and corrupt Windows 7?

[MisterEd]

Has anyone had a overheating CPU that caused Windows 7 to become corrupted? Was the overheating caused by a stock cooler that was inadequate for the CPU it came with. Did upgrading to a better CPU cooler solve the CPU temperature problem?

I upgraded a computer. The hardware is discussed in this thread: Is anyone still using Windows 7? #457 .

I had a hard crash and had to reinstall Windows 7. I am not sure what caused the crash. I discussed whether updates or other software caused the crash in this thread: Can Simplix UpdatePack 7 conflict with Windows Updates?

Thinking back now I think the hard crash might have been caused by the "Power Plan" which caused the CPU to overheat.

The default power plan is "Balanced (recommended)".
With this plan the The AMD FX-8350 CPU is always at 4000MHz and sometimes boosts to 4200MHz. The idle temperature varies from 45-65C.
Note the AMD FX-8350 CPU has a maximum safe core (package) temperature limit of 61°C.

I did some benchmarks to use as a comparison with the old computer motherboard and CPU. I should have been watching closer but I think the CPU temp went as high as 70-85C or higher. When I came back the computer had shut down and wouldn't boot to Windows 7 anymore. I booted with the Windows 7 flash drive. System Repair failed. I had to reinstall Windows 7 Pro.

Through experimentation I found that I needed to use the power plan "Power Saver". It allows the CPU to idle at 1400MHz but boosts to 4000MHz if needed. The idle temperature is now about 35-45C. At 4000Hz the temp would still go up to 65C but came down as soon as the CPU speed dropped.

The temperature problem is probably caused by the AMD stock cooler. I have read that it is inadequate for the CPU I have. I want to replace it with a Cooler Master Hyper 212 EVO CPU cooler. That is if I can ever get one.

I bought a Cooler Master Hyper 212 EVO CPU cooler on Ebay on June 28. The seller estimated a delivery date of Thu July 2 - Wed July 8. Since it hasn't been shipped yet there is no way I will get it by July 8. The seller hasn't responded to any messages I sent him. If I don't get it by Thu July 9 I am entitled to contact Ebay for a refund. If it comes to that I will have to try to buy the cooler from another seller.

 

[townsbg]

Bench marks are going to push the processor to the limits to see what it can do no matter what the power setting is. You need to watch it carefully while doing that. I don't know about windows corrupting for sure but I suspect that it's possible.
I had a custom built computer once upon a time that kept overheating. Either dust would get in the heatsink or the heat sink fan went out. When it hit a certain temp the cpu turned off the computer to avoid damage. I'm surprised yours doesn't do the same. I've heard in general that stock heatsinks are bad news.

 

[MisterEd]

Bench marks are going to push the processor to the limits to see what it can do no matter what the power setting is. You need to watch it carefully while doing that. I don't know about windows corrupting for sure but I suspect that it's possible.
I had a custom built computer once upon a time that kept overheating. Either dust would get in the heatsink or the heat sink fan went out. When it hit a certain temp the cpu turned off the computer to avoid damage. I'm surprised yours doesn't do the same. I've heard in general that stock heatsinks are bad news.

I misspoke. Actually the computer just shut down without warning. I assume that was caused by the CPU temp getting to high.

The original motherboard failed a couple years ago. I bought a replacement on Ebay last week. With the former setup I had a Corsair H60 water cooler. It worked quite well with the AMD FX-8350 CPU. I bought it in early 2012. I could use it again but I am not sure I trust it because of its age. Would you trust an AIO water cooler that old?

BTW, I have another computer with an AMD Ryzen 3700X CPU. It has a Corsair H60 water cooler that I bought it in early 2017.

Note the stock CPU cooler came with the AMD FX-8350 CPU. I used to use it with an AMD Phenom II X4 960T CPU. It had no problem cooling it.

 

[townsbg]

I suspect it's programmed to shutoff at a certain tempt to avoid damage like mine was.

 

[Lenovouser72]

On my very old laptop I had that problem earlier - by some long time very high use of CPU -
INTEL Core Duo 2 T5200 caused sometimes Windows shut down.
I consulted with local computer service maintenance and was instructed to make some
changes > I then cleaned properly a cooler and copper cooling tubes from it +
made original pattern cooling vents/strips on the bottom of the laptop a little wider
and placed additional small adhesive pads ( 3 mm, for furniture) under the original plastic
ones under the bottom, lifting laptop higher, thus increasing air access to the cooler.
In result I succeeded to decrease CPU work temperature up to 5 degrees,
now 45 by idle and 60 - 65 degrees by 100 % CPU.
(Recommended by manufacturer 45 - 70 degrees, safe.)
No problems with Windows shut down since.

 

[MisterEd]

I finally found out why the computer was shutting down. I was blaming the cooler but it was not really at fault. In digging though some old computer parts I found an old CPU I thought that I discarded a long time ago. It was a Corsair H60 AIO water cooler. I connected the H60 pump and fan to the motherboard as follows:

I decided to stress the CPU using Prime95. I needed to confirm the CPU temperature was being kept under control.
I used the following app for monitoring:

• Core Temp/Gadget - monitor CPU temp
• AMD Overdrive - monitors Thermal Margin for AMD FX CPUs
• SpeedFan - monitor CPU fan & pump speeds

Setup #1: H60 pump - SYS_FAN1 | H60 fan - CPU_FAN

While at Idle the
CPU temp = 10C | Thermal Margin was 60C.
H60 Pump = 4383 RPM | H60 CPU fan = 1000 RPM

While running Prime95 for 15 minutes I checked the temps again.
CPU temp = 40C | Thermal Margin was 30C.
H60 Pump = 4383 RPM | H60 CPU fan = 1679 RPM

I went away for maybe 10 or 15 minutes. When I came back I was upset when I saw the following
CPU temp = 60C | Thermal Margin was 10C.
Pump fan = 4383 RPM | CPU fan = 393 RPM

It didn't make sense the CPU fan was running so slow. A little research said that speed was not accurate and meant the fan had failed.

Since I thought that the CPU had failed I replaced it with a spare. I also swapped power connector for the pump and fan.

Setup #2: H60 pump - CPU_FAN | H60 fan - SYS_FAN1

I ran Prime95 again after letting the computer cool off for awhile.

I went away for maybe 20 minutes. When I came back I was confused when I saw the following
CPU temp = 62C | Thermal Margin was 8C.
Pump fan = 0 RPM | CPU fan = 1679 RPM

Then it hit me. All the CPU coolers and pumps I had tried were probable good. The motherboard's CPU_FAN power connector was intermittent. When it failed either the fan or pump connected to it stopped turning. That caused the CPU to overheat. The computer was supposed to shut down if the CPU overheats very long but I shouldn't have to depend upon that.

Even though the motherboard was used from Ebay it is not acceptable to have such a major failure. I have sent it back for a refund. I bought another motherboard just like it from another seller. Hopefully it won't have any problems. When asked the seller told me there was nothing wrong with it but I will hold judgment until I check it out.

I really want to keep this computer going because it is the last one I have Windows 7 on. However, if the replacement motherboard doesn't work out I will probably have to retire the computer and Windows 7 also.

* Note: by comparing Core Temp & Overdrive I deduced that the temperature that Core Temp shows is 70C - Thermal Margin. According to what I read the CPU temperature is safe as long as the Thermal Margin is greater than 10C.