CPU over temperature / Network card not working: should I replace the whole motherboard?

Question

I have a custom built PC since 2011.

The configuration is:

• Asus P8P67-PRO
• Intel i7 2600k with stock fan
• 16GB DDR3 RAM 1600MHz
• AMD Radeon HD 6970
• Samsung EVO 750 / 250GB
• 2TB 5400rpm HDD

Although I'm primarily a Linux user this is my home PC and I need to run some windows packages (mostly Adobe software) so I'm running Windows 7.

Yesterday I experienced LAN issues (the card randomly disconnected without apparent reasons). I saw other answers which recommended to disable the network card and enable it again. It worked, but after a few minutes it stopped working again. It went like that for 4-5 times until the BSOD appeared. Since then I've not been able to enter Windows anymore, each time it would finish loading and then go to BSOD again.

I went to recovery mode and I got stuck at loading classpnp.sys, I couldn't enter recovery mode too.

I thought it was a Windows error, maybe some "hidden bug" which came out yesterday.

It's worth noting that after the network card error, the BIOS started telling "CPU over temperature error!".

I plugged in an Ubuntu 17.04 live distro, I managed to boot it up and opened firefox for browsing: if attached to the LAN wired connection, the connection drops in 4-5 minutes, as happened on windows. So it doesn't seem to be a windows-related issue, it seems that my hardware is broken somehow.

The LAN port is the default one integrated on the motherboard.

I also installed lm-sensors to investigate on the cpu temperature error, and this is the output:

ubuntu@ubuntu:~$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Package id 0:  +90.0°C  (high = +80.0°C, crit = +98.0°C)
Core 0:        +86.0°C  (high = +80.0°C, crit = +98.0°C)
Core 1:        +90.0°C  (high = +80.0°C, crit = +98.0°C)
Core 2:        +84.0°C  (high = +80.0°C, crit = +98.0°C)
Core 3:        +82.0°C  (high = +80.0°C, crit = +98.0°C)

These temperatures are not good. My CPU used to stay between 40-50°C when idling, the fact that the idle temperature here is between 80°C and 90°C tells me there is something really wrong.

Also, I regularly clean my PC to remove any excess dust, I can see the stock fan moving and it always worked pretty good, so my guess is that there is something wrong with the MB/CPU.

So, my question is: Are cpu temperature error and network card breaking related? Did my motherboard break? Should I replace both motherboard and CPU?

Answer 1

Temp is certainly high, but based on my experience it isn't enough to make other components fail, note that those are just the CPU temps, not the motherboard's overall temp, so it must be cooler, too. I'd be more concerned about dust or other conductive particles stuck in between connections that cause it to randomly short. Try cleaning it thoroughly (maybe this also causes high temps), using compressed air and an anti-static blanket or a hard brush, If this doesn't help then the mobo is indeed shorted and you might have to replace it before it can damage other components.

Answer 2

First things first. Those CPU temps really aren't good. If they are that hot idle they are almost certain to overheat under a load. When is the last time you took your CPU cooler out and cleaned it? Dust and grime builds up on the heatsink and the fan blades, both making the fan less effective at moving air over the heatsink and the heatsink less effective at transferring heat to that air.

It surprises people to find out this is part of regular maintenance for a computer, but it is. This goes quadruple for smokers, even if you don't smoke near the computer. For non-smokers, I suggest cleaning the CPU cooler every other year. For smokers, twice a year. If you are an avid gamer with a dedicated graphics card you should do the same for it... though not if it's under warranty (the warranty is void if you remove the EMI shield to clean the heatsink under it). And always clean the heatsink and CPU of old thermal paste and re-apply each time. 91% or better rubbing alcohol is good for both as it evaporates quickly, leaves negligible residue, is non-corrosive, but cuts through the oils in dust/cigarette residue. There are also cleaning solutions specially made for this.

Now for the other. Have you recently started using Bluetooth? I noticed your motherboard sports an Intel integrated NIC. It also has Bluetooth. Some Intel NICs actually have Bluetooth integrated into them... so it's possible they are one and the same. Some of those Intel NICs w/Bluetooth have a flaw that causes the NIC to malfunction when Bluetooth is active. It does not care what OS the system is running on as it's a problem in the NIC itself and not the driver/kernel module. My Intel WiFi NIC actually has this problem and behaves much the same as you describe when bluetooth is enabled. See if you can disable Bluetooth completely in the bios.

Failing that, you probably are suffering a failing NIC. Installing a discrete network adapter in one of those PCIe x1 slots would be your best bet.

If it _is failing this is probably not related to the CPU temps... but it's possible. Possible scenarios include the CPU overheating causing damage over time to the runs on the motherboard. ASUS makes good stuff generally so I would be surprised at that.

Another is that the real culprit is your PSU. If you PSU is not providing clean voltage or is producing too much/little voltage or too little amperage it could cause both heat issues with the CPU and malfunctions throughout the computer. PSUs really are the single most important component in a computer... but they get no love at all.

There are no real tests you can do for the latter without special equipment (PSU load testers... even most computer shops don't have them). I know you like your integrated sensors. I've currently got one that says the core voltage to my gpu is negative 3000000 or so volts. Obviously, if you're suffering issues of any kind the integrated sensors are not to be trusted. Test this one by trying a different PSU and seeing if the problem resolves.