Core i7 920 Sudden Overheating

[Dark54555]

So, was working at my desktop the other night, and it BSODed with a warning about CPU heat. I've been monitoring since.

To me, 60C idle and 100C max seems high...extremely high. This rig hasn't had a heat issue since it was built in 2008 (including before and after a case swap in 2010). No major changes recently, case is clean, house is a normal temp. Any ideas? Googling it, this seems like it happens with a lot of Core i7 920 chips, but no one ever seems to post a reason why, or a confirmed fix.

On a related note, the mobo is an Asus P6T, which does seem to show up in a lot of heat related reports, but usually in OC rigs (I'm running stock speeds). Voltage issues are often mentioned but never confirmed.

Thanks in advance for the help.

 

[fixbsod]

Dead fan? Bad bearrings? Did you move the computer recently -- perhaps an install/upgrade maybe the HSF isn't making as good contact lately. If you open the case are all the fans moving ? Any whining or screeching / scraping noises? 60c idle sounds way too hot, even for an i920.

 

[T_Yamamoto]

reapply thermal and heatsink

have you oced your cpu before?

 

[Dark54555]

Fans are working (CPU and case). No major moves or anything lately.

 

[dajeepster]

i just started having this problem with one of my systems... i'm going to reapply thermal paste... as long as i don't do anything cpu intensive, it's fine for now.

edit... and mines on water too..

 

[Zebo]

Some types of thermal pastes dry out leaving microscopic gaps. Dust is a killer when coating HSF. Fan itself my be going out. I'd just get a new HSF kit that includes paste and mount to see what happens. Mines a late stepping on water with three years old Arctic silver 5 tim joint and 4.2 Ghz and stays at 40-65C since build three years ago.

 

[VirtualLarry]

IDC, can you comment here? Can CPUs have their gate-oxide/dielectric degrade, such that leakage currents increase over the lifetime of the CPU, to manifest in sudden higher idle and load temps?

Or is this more likely just a cooling subsystem issue?

 

[PreferLinux]

IDC, can you comment here? Can CPUs have their gate-oxide/dielectric degrade, such that leakage currents increase over the lifetime of the CPU, to manifest in sudden higher idle and load temps?

Or is this more likely just a cooling subsystem issue?

I won't presume to try to answer for him, but I would suggest that any degradation would most likely cause gradual increases in power usage which would translate to gradual increases in temperature.

 

[SpeedTester]

Maybe one of your pushpins failed or popped out if you are using the stock heatsink. My 920to in the smart board had been running strong for about 2.5 years at 4.0 bit I am using an aftermarket hsf

 

[sm625]

Yeah check those push pins. Methinks intel designed them to fail as part of a planned obsolescence campaign. Watch your temps as you push down on each of those pins, one at a time. I bet that when you push on one of them, your temps will drop instantly by 20C or more.

 

[Dark54555]

Yeah check those push pins. Methinks intel designed them to fail as part of a planned obsolescence campaign. Watch your temps as you push down on each of those pins, one at a time. I bet that when you push on one of them, your temps will drop instantly by 20C or more.

Thanks for the idea (to all who mentioned clips), and I will test/report back once I'm home this evening.


Sent from my iPhone using Tapatalk which is probably the reason for the typos

 

[Makaveli]

wow ya that doesn't look right I idle in the low 40's with HT on.

Reseat HSF and what is the room temp?

Also what tower do you have ?

 

[Dark54555]

wow ya that doesn't look right I idle in the low 40's with HT on.

Reseat HSF and what is the room temp?

Also what tower do you have ?

Room is 75F. Case is an Antec 300.


Sent from my iPhone using Tapatalk which is probably the reason for the typos

 

[Dark54555]

Ordered a new cooler, clips are definitely not as secure as they were.

 

[Puppies04]

Ordered a new cooler, clips are definitely not as secure as they were.

Make sure you have some TIM remover to get rid of the old dried out TIM.

 

[tpin]

reapply thermal and heatsink

have you oced your cpu before?

Ditto,I too vouch for the same.

 

[Dark54555]

Put a Cooler Master 212 Evo on it with a new coat of AS5, which seems to have solved it. Idle is 30-40C, and it's still in the cure period.

 

[Makaveli]

Put a Cooler Master 212 Evo on it with a new coat of AS5, which seems to have solved it. Idle is 30-40C, and it's still in the cure period.

Glad you got it fixed.

So was the issue just a loose clip or something on your previous cooler or some other kind of contact issue?

 

[Zap]

Or is this more likely just a cooling subsystem issue?

99.9999999% of the time... this.

Put a Cooler Master 212 Evo on it with a new coat of AS5, which seems to have solved it. Idle is 30-40C, and it's still in the cure period.

That's like buying a new car just because the check engine light came on.

Glad it's solved, however.

 

[dajeepster]

99.9999999% of the time... this.



That's like buying a new car just because the check engine light came on.

Glad it's solved, however.

it's more akin to buying a new radiator and fan cause the check engine light came on..

 

[Dark54555]

Glad you got it fixed.

So was the issue just a loose clip or something on your previous cooler or some other kind of contact issue?

It looks like one of the clips had come loose and didnt want to re-attach. That is one of the disadvantages of the cheap plastic clips on the intel stock cooler.


Sent from my iPhone using Tapatalk which is probably the reason for the typos

 

[Makaveli]

It looks like one of the clips had come loose and didnt want to re-attach. That is one of the disadvantages of the cheap plastic clips on the intel stock cooler.


Sent from my iPhone using Tapatalk which is probably the reason for the typos

ahh good to know thanks for the update.