High temps on 6700k
- Thread starter Brok
- Start date
Why buy a new cooler if you have one that's working perfectly fine? that seems to be a waste of time and money. You said that your idle is around mid 30's and load beyond 90's, If I had to guess, I know that you have re-seated the pump block already, but did you make sure that you have a full contact between the coldplate and CPU? it's worth double checking.
You said the the pump is running around 2200rpm - that is the correct spec for the H100, but check the tubes also and see if one is warm and one cold. If they are both warm, that's an indication of no liquid circulation causing your CPU temp to be extremely high under load.
You said the the pump is running around 2200rpm - that is the correct spec for the H100, but check the tubes also and see if one is warm and one cold. If they are both warm, that's an indication of no liquid circulation causing your CPU temp to be extremely high under load.
Hi, thanks for the advice, what is the best method for checking the contact as you describe please? I'd be more than happy to just fix what I have
I did feel the pipes yesterday and one felt warmer than the other. I can't remember if that was under load though so will try again later (in the office today so less fiddle time.)The problem is not easily diagnosed. If pump is moving coolant, one pipe will be warmer than the other. But this does not tell us how much coolant is being moved, and if the pump is moving half as much as it needs to the CPU will run hot because there is not enough coolant flowing through to keep it cool. Feeling the pipes only tells us coolant is moving, not how much coolant is moving.
H100 pump supplies only just enough flow when new, so as impeller wears and flow gets lower the CPU get hotter.
Combine this with a mounting system that many have had problems with and you have a recipe for problems that is never ending.
I don't understand this 'Keep your CLC' mentality. It has been proven over and over that all the CLC is better hype has been proven to be nothing but marketing propaganda to the point they are often lies. CLCs are louder, don't last as well, cost more, are not as well built and designed, and when they start failing (and the all will) are hard to diagnose problems on. A good air cooler is well design, well built, and the only part that wears out is the fan .. which will last at least 4-5 years and usually much longer. Some have a MTBF of 150000 (17.1 years) while most are 50000 (5.7 years)
H100 pump supplies only just enough flow when new, so as impeller wears and flow gets lower the CPU get hotter.
Combine this with a mounting system that many have had problems with and you have a recipe for problems that is never ending.
I don't understand this 'Keep your CLC' mentality. It has been proven over and over that all the CLC is better hype has been proven to be nothing but marketing propaganda to the point they are often lies. CLCs are louder, don't last as well, cost more, are not as well built and designed, and when they start failing (and the all will) are hard to diagnose problems on. A good air cooler is well design, well built, and the only part that wears out is the fan .. which will last at least 4-5 years and usually much longer. Some have a MTBF of 150000 (17.1 years) while most are 50000 (5.7 years)
Hi, thanks for the advice, what is the best method for checking the contact as you describe please? I'd be more than happy to just fix what I have
When you've installed the pump on top of the CPU, make sure there's absolutely no play on all sides, the pump should be firm in place. Also, when you removed the pump after re-seating it, was the thermal paste evenly spread out? I've seen cases where some part of the coldplate that had TIM was untouched after securing the block, kind of like this, http://i.imgur.com/sGlmvNsl.jpg
Hi, I did check as you suggested, there was no movement at all on the block and the paste was evenly spread on the plate.
I fitted the Olymp this afternoon, mid-20s idle, high 50s under load at stock, where I was seeing up to 100c with the H100 on the same benchmarks. Much better!
I fitted the Olymp this afternoon, mid-20s idle, high 50s under load at stock, where I was seeing up to 100c with the H100 on the same benchmarks. Much better!
Yeah one thing that has stopped is an intermittent clicking, happened once every minute or so before so I guess it was dying.
It is quieter, but not silent. Its noticably quieter under load. Listening to it now I think the loudest sound is the gfx card fans, but tbh its quiet enough so I shan't spend too much more time on it for now.
I should probably remove the rear case fan as its only 1cm away from the 2nd fan on the Olymp so I don't suppose its adding much and is just blocking the rear vent. I had to mount the CPU fans with the second one to the rear as it didn't clear the ram when mounted on the front.
It is quieter, but not silent. Its noticably quieter under load. Listening to it now I think the loudest sound is the gfx card fans, but tbh its quiet enough so I shan't spend too much more time on it for now.
I should probably remove the rear case fan as its only 1cm away from the 2nd fan on the Olymp so I don't suppose its adding much and is just blocking the rear vent. I had to mount the CPU fans with the second one to the rear as it didn't clear the ram when mounted on the front.
Have you set the Olymps fan curve yet? I would set them to be about 30% at 30c and 100% at about 70c. Mine is set 20% @ 20c (right now that is 300rpm @ 26-28c) and 100% @ 80c, but at 100% load it is 900rpm @ 55c. This is to be my new test stations so is on top of my 'work bench' where I'm setting. You can't hear it in very quiet room.
Removing the rear case fan may help noise levels to. If you don't have a bottom intake put it down there
If you are going to dispose of your H100, I might be interested in it as a test radiator for airflow measurements. It's dense fin design makes it a good baseline resistance.
Removing the rear case fan may help noise levels to. If you don't have a bottom intake put it down there
If you are going to dispose of your H100, I might be interested in it as a test radiator for airflow measurements. It's dense fin design makes it a good baseline resistance.