Abysmal thermals from 212 Evo?

[RadiatingLight]

My thermals have been mostly fine at stock speeds on my 4770K with a Hyper 212 Evo, around 70C or so at full load. however, I recently switched motherboards, and noticed that my new one had an auto-overclock feature. sure, why not?? it turns out that the motherboard did a stellar job at bringing my CPU up from 3.5Ghz to around 4.1Ghz with no user intervention, but when I decided to stress test that overclock, my temps soared up to 100 Frickin Degrees! I instantly shut off my computer and lowered the overclock, but it got me wondering: Is the 212 EVO simply not good enough to handle the heat, or did I do something wrong.

 

Note: CPU voltage was about 1.18V, and it was consuming a total of about 95 Watts, so it really shouldn't be this hot.

[Cereal5]

well the reason the 4790K came out is because the 4770K had internal thermal paste problems, so it might be possible that you got a chip with a more pronounced problem. 212 Evos aren't meant to cool that much CPU really either so it pronounces the problem even more

[RadiatingLight]

1 minute ago, Cereal5 said:

well the reason the 4790K came out is because the 4770K had internal thermal paste problems, so it might be possible that you got a chip with a more pronounced problem. 212 Evos aren't meant to cool that much CPU really either so it pronounces the problem even more

But with heat dissipation, more heat at the CPU should mean more rapid dissipation, should it not? how does this ever get close to 100C

 

Also, do you think that de-lidding my 4770K is worth the risk?

[Cereal5]

5 minutes ago, RadiatingLight said:

more heat at the CPU should mean more rapid dissipation

i dont quit understand. You mean that the higher the temps the better your cooler should dissipate it? no, the cooler fans would have to work harder and the fans can only do so much

 

6 minutes ago, RadiatingLight said:

Also, do you think that de-lidding my 4770K is worth the risk?

I've never delidded a CPU and I won't pretend to know much about it, but the short answer is no

[zMeul]

I've tried to tell people to stop using the 212EVO on OC CPUs for quite a while now

people don't listen

 

the 212EVO was good on older gen CPUs, not on the new behemoths that have 80-90W+ TDW at stock settings

[RadiatingLight]

18 minutes ago, zMeul said:

I've tried to tell people to stop using the 212EVO on OC CPUs for quite a while now

people don't listen

 

the 212EVO was good on older gen CPUs, not on the new behemoths that have 80-90W+ TDW at stock settings

So what's a better alternative?

[zMeul]

4 minutes ago, RadiatingLight said:

So what's a better alternative?

if you want to stick with air cooling, look for beefier ones; or go liquid cooling with at least 240mm rad

[Lays]

3 hours ago, RadiatingLight said:

But with heat dissipation, more heat at the CPU should mean more rapid dissipation, should it not? how does this ever get close to 100C

 

Also, do you think that de-lidding my 4770K is worth the risk?

Delidding and using CLU as your paste instead of the intel stock stuff will drop ~10-20c or so off your load temps, it's totally worth it.  It's really not hard, just watch a few tutorials or spend like $30 on a 3D printed delid tool

[WoodenMarker]

Try reseating the cooler.

[airdeano]

13 hours ago, RadiatingLight said:

... noticed that my new one had an auto-overclock feature. sure, why not?? it turns out that the motherboard did a stellar job at bringing my CPU up from 3.5Ghz to around 4.1Ghz with no user intervention, but when I decided to stress test that overclock, my temps soared up to 100 Frickin Degrees! I instantly shut off my computer and lowered the overclock, but it got me wondering: Is the 212 EVO simply not good enough to handle the heat, or did I do something wrong.

 

Note: CPU voltage was about 1.18V, and it was consuming a total of about 95 Watts, so it really shouldn't be this hot.

what application did you use to 'stress' the CPU? p95, aida, intel XTU?

 

yes, you prolly have done something wrong. more than likely the canned tune by the motherboard uses an adaptive voltage mode. you are not supposed to stress the CPU with synthetic stress tools. what does the temperatures do during normal use (game play, productivity, etc..).

[Gammagoat]

5 hours ago, WoodenMarker said:

Try reseating the cooler.

No brainer.

 

Doesn't take much to get the seat wrong on a 212.

[Maxxtraxx]

18 hours ago, zMeul said:

I've tried to tell people to stop using the 212EVO on OC CPUs for quite a while now

people don't listen

 

the 212EVO was good on older gen CPUs, not on the new behemoths that have 80-90W+ TDW at stock settings

Some of the Older gen cpus are 130w TDP see i7-950 Here

[zMeul]

8 minutes ago, Maxxtraxx said:

Some of the Older gen cpus are 130w TDP see i7-950 Here

at stock clock, not OCed

when you OC a 80-90W CPU you will end will end up will well above 130W

 

---

 

an example, couple of i7-5960X at stock and OCed: http://www.corsair.com/en-us/blog/2014/september/haswell_e_overclocking_and_power

[Maxxtraxx]

5 minutes ago, zMeul said:

at stock clock, not OCed

when you OC a 80-90W CPU you will end will end up will well above 130W

yep, and my overclocked 2500k runs great with my non evo hyper 212.

[DragonTamer1]

I'm assuming that you are using something other than Prime95 for stress testing. Use the trial version of Aida64 if you can get it.

 

Re-seat the CPU cooler with new thermal paste first.

 

Check the CPU voltage to make sure it is less than 1.4 in the BIOS. If necessary manually set it so that it does not reach any higher than that for verification.

 

Invest in a much more substantial cooler to help dissipate the heat (see below).

 

If none of this works, then and only then should you consider deliding the CPU. It is a last resort scenario, not a first resort. Even still I would not suggest it.

 

Possible CPU coolers.

 

[Maxxtraxx]

18 hours ago, RadiatingLight said:

So what's a better alternative?

Reseat the cooler with new thermal paste, Perform a manual overclock yourself and steadily increase the overclock while watching the temperatures. The self overclocking features seem to be very aggressive with core voltage (I'm thinking there is a possibility it could be much higher than what you quickly observed).

[Maxxtraxx]

19 minutes ago, zMeul said:

at stock clock, not OCed

when you OC a 80-90W CPU you will end will end up will well above 130W

 

---

 

an example, couple of i7-5960X at stock and OCed: http://www.corsair.com/en-us/blog/2014/september/haswell_e_overclocking_and_power

 

Yes overclocking causes power draw to increase significantly, that is known.

 

What my attempt to point out was:

 

that it is not the 212 evo that is the issue, it's a fine cooler that can work well with basic overclocks and it was working very well for him before the Mobo switch

and

That older processors tend to have higher power consumption because of the larger production node and higher voltages, so newer processors are not TDP behemoths but tend to draw less power.

 

I7-975 130w TDP can run 1.5v or higher when overclocked

i7-3770k 77w TDP tends to run under 1.4v when overclocked

 

 

[zMeul]

Just now, Maxxtraxx said:

Yes overclocking causes power draw to increase significantly, that is known.

 

What my attempt to point out was:

 

that it is not the 212 evo that is the issue, it's a fine cooler that can work well with basic overclocks and it was working very well for him before the Mobo switch

and

That older processors tend to have higher power consumption because of the larger production node and higher voltages, so newer processors are not TDP behemoths but tend to draw less power.

 

I7-975 130w TDP can run 1.5v or higher when overclocked

i7-3770k 77w TDP tends to run under 1.4v when overclocked

notice the temps with the 212EVO: https://www.techpowerup.com/reviews/MSI/Core_Frozr_L/6.html

 

that's a i7-6700K - at stock it's 91W TDP

[Maxxtraxx]

On ‎10‎/‎21‎/‎2016 at 8:26 PM, zMeul said:

notice the temps with the 212EVO: https://www.techpowerup.com/reviews/MSI/Core_Frozr_L/6.html

 

that's a i7-6700K - at stock it's 91W TDP

Sorry, but that graph specifically says OC(Overclocked) so it is not at its stock 91W TDP.

Even overclocked it has not thermal throttled the CPU and is still in Intel's safe operating range, and that is using an synthetic processor load that does not reflect normal use situations but reflects a worse case scenario.

 

You'll notice in the graph at Stock 91W TDP the 212evo shows a max temp of 63C.

[RadiatingLight]

3 hours ago, Maxxtraxx said:

Sorry, but that graph specifically says OC(Overclocked) so it is not at its stock 91W TDP.

Even overclocked it has not thermal throttled the CPU and is still in Intel's safe operating range, and that is using an synthetic processor load that does not reflect normal use situations but reflects a worse case scenario.

 

You'll notice in the graph at Stock 91W TDP the 212evo shows a max temp of 63C.

This is exactly why I was asking. with 91Watt TDP, why does my CPU get so much hotter than that graph, because not overclocked, it gets to 75C, and with only a 200Mhz increase in turbo boost clocks, it jumps to 100C (Never using more than 100W of power)

why??

[DragonTamer1]

2 hours ago, RadiatingLight said:

This is exactly why I was asking. with 91Watt TDP, why does my CPU get so much hotter than that graph, because not overclocked, it gets to 75C, and with only a 200Mhz increase in turbo boost clocks, it jumps to 100C (Never using more than 100W of power)

why??

Have you checked your core voltage yet. If you are at stock your voltage should be no higher than 1.2V.

 

It sounds to me like there is poor contact between the CPU and the cooler or there is more voltage being pushed through the processor than there should be.

[Maxxtraxx]

On ‎10‎/‎21‎/‎2016 at 1:55 PM, WoodenMarker said:

Try reseating the cooler.

 

2 hours ago, DragonTamer1 said:

Have you checked your core voltage yet. If you are at stock your voltage should be no higher than 1.2V.

 

It sounds to me like there is poor contact between the CPU and the cooler or there is more voltage being pushed through the processor than there should be.

 

4 hours ago, RadiatingLight said:

This is exactly why I was asking. with 91Watt TDP, why does my CPU get so much hotter than that graph, because not overclocked, it gets to 75C, and with only a 200Mhz increase in turbo boost clocks, it jumps to 100C (Never using more than 100W of power)

why??

With the problem that you are seeing, That being that your temps jumped drastically after only a motherboard swap.

 

I believe the best course of action is to:

 

Remove any overclock (which it seems that you have done already)

 

Reinstall the cooler:

1. remove cooler

2. clean off all old thermal paste from the cpu and cooler bottom using cotton balls and rubbing alcohol

3: apply a small pea sized dot of thermal paste on top of the CPU(about the size of the top of a capacitor on the motherboard)

4: install the cooler and tighten the cooler using a X shape pattern

5: Be sure to plug the Cooler fan into the CPU fan header on the motherboard

6: rerun your tests to see what happens

[zMeul]

9 hours ago, Maxxtraxx said:

Sorry, but that graph specifically says OC(Overclocked) so it is not at its stock 91W TDP.

Even overclocked it has not thermal throttled the CPU and is still in Intel's safe operating range, and that is using an synthetic processor load that does not reflect normal use situations but reflects a worse case scenario.

1. why would you put a 212 EVO on a stock CPU; other than K series don't come with a cooler
2. the scenario is about OCing, the OP OCed his CPU and had bad thermals
3. if it didn't thermal throttle it's OK?! safe?! absolutely not! running CPUs hot for extended periods of time will damage the CPU

[Maxxtraxx]

1 minute ago, zMeul said:

1. why would you put a 212 EVO on a stock CPU; other than K series don't come with a cooler
2. the scenario is about OCing, the OP OCed his CPU and had bad thermals
3. if it didn't thermal throttle it's OK?! safe?! absolutely not! running CPUs hot for extended periods of time will damage the CPU

1: Why not? Better thermals and quieter operation are always preferable.

 

2: The scenario is primarily about bad thermals not OCing, his thermals are off even when not overclocked. He before the swap had is system running perfectly, with an OC and thermals were fine. He changed ONLY the motherboard and thermals went very bad. Therefore, the 212 he is using was working perfectly before the Motherboard change therefore it should work for the same processor after a Motherboard swap. It's not the cooler that is at fault and not the OC that is at fault since even at stock it has poor thermals.

 

3: Correct, If it didn't thermal throttle that doesn't mean that the temps are preferable but are acceptable. How many Server farm CPU's with tiny heat sinks and low air flow, Clogged home computer heat sinks and terrible laptop thermal solutions lead to thermal throttling but those processors continue to work for years in the worst case scenarios. The processor thermal throttles itself in order to keep the thermals within Intel's safe operating range and thusly avoids CPU damage. Electromigration will have a greater effect upon Silicon with higher temps and more voltage but how many overclocked/overtaxed/clogged home pc CPU's are still running 5+ years down the road with little impact. I know my 2500k overclocked since got it and cooled with a 212 is still maintaining it's OC and working great 4 years later.

[zMeul]

3 minutes ago, Maxxtraxx said:

How many Server farm CPU's with tiny heat sinks and low air flow, Clogged home computer heat sinks and terrible laptop thermal solutions lead to thermal throttling but those processors continue to work for years in the worst case scenarios. The processor thermal throttles itself in order to keep the thermals within Intel's safe operating range and thusly avoids CPU damage. Electromigration will have a greater effect upon Silicon with higher temps and more voltage but how many overclocked/overtaxed/clogged home pc CPU's are still running 5+ years down the road with little impact. I know my 2500k overclocked since got it and cooled with a 212 is still maintaining it's OC and working great 4 years later.

BS

server farms actually go for liquid cooling if the thermals demand it

do you know how many mobos and PSUs I saw damaged because of dust accumulation in the CPU's heatsink? a lot