Sign in with
Sign up | Sign in
Your question
Solved

Water Cooled CPU significantly above water temp

Tags:
  • Overclocking
  • Stress Test
  • CPUs
Last response: in Overclocking
Share
August 9, 2014 5:41:51 PM

Hi All
As title suggests. I have my CPU water cooled.

The water temp in the radiator is 18C and ambient temp is 16C.

The CPU is being reported as 28C at idle and 60C after 5mins of stress test.

Goes straight back to 30C after stress test but I feel like I am getting results roughly equivalent to air cooling and it has cost me a great deal of time and money.

Would anyone suggest that this is normal or should I revisit my TIM application?

For reference the following are the details.

Res -> Pump -> Core i7-920(@4ghz) -> Radiator(triplex80mm) -> Res

Thanks in advance.
Matt

More about : water cooled cpu significantly water temp

a b K Overclocking
a b à CPUs
August 9, 2014 6:14:35 PM

What is the model of your cooler and exactly what is the arrangement of case fans. Where is the cooler installed and in what configuration.
m
0
l
August 9, 2014 6:46:24 PM

Hi thanks for the reply. No case fans. The setup is spread across my desk at the moment.

The rad is an alphacool monster. Triple wide by 80mm deep.

The water block is a Phobya uc-2lt.



m
0
l
Related resources
a b K Overclocking
a b à CPUs
August 9, 2014 8:37:29 PM

Well, your lack of cooling ceraintly isn't due to components unless it's your radiator fans. Are you using a shroud with the radiator? Do you have a pic of the complete setup as it sits. I'd like to see it. Of course you know that some of the vastly differing opinions regarding the best cooling setups believe that for some particular setups not having the components in the case actually works to the negative due to the lack of a real directional airflow across components from good case fan configuration.

I don't know if that's true or not but I do know for a fact that in automotive cooling, cars with the radiator installed under and in between all that metal, with the proper fan shroud installed, run much cooler than test stand engines with complete but open air radiator configurations.

BTW, how much TIM did you use or which method?
m
0
l
a b K Overclocking
a b à CPUs
August 9, 2014 8:43:02 PM

Also, the cpu temp will always be higher than water temp. Components can conduct only so much heat away regardless of the precision with which they were installed.
m
0
l
August 9, 2014 8:57:45 PM

darkbreeze said:
Well, your lack of cooling ceraintly isn't due to components unless it's your radiator fans. Are you using a shroud with the radiator? Do you have a pic of the complete setup as it sits. I'd like to see it. Of course you know that some of the vastly differing opinions regarding the best cooling setups believe that for some particular setups not having the components in the case actually works to the negative due to the lack of a real directional airflow across components from good case fan configuration.

I don't know if that's true or not but I do know for a fact that in automotive cooling, cars with the radiator installed under and in between all that metal, with the proper fan shroud installed, run much cooler than test stand engines with complete but open air radiator configurations.

BTW, how much TIM did you use or which method?


I'll take some pics later on. But I'm the meantime. I used the pea/lentil method and tightened the screws clockwise sequentially.
m
0
l
a b K Overclocking
a b à CPUs
August 9, 2014 9:00:56 PM

Yeah, that should be about the right amount considering that's a 45nm die cpu. If it was smaller like the 4th gens I'd say it was too much but it should be fine for that cpu.
m
0
l
August 9, 2014 9:34:30 PM

So it is too
m
0
l
August 9, 2014 9:35:08 PM

So the temperature delta is too much? I should expect closer to the water temp?
m
0
l
a b K Overclocking
a b à CPUs
August 9, 2014 10:30:07 PM

No, I meant the amount of TIM you used. That processors stock clock is 2.66Ghz so you have it OC'd by a LOT. 1340mhz to be exact. Most guys are happy to get a 500-800mhz overclock. About the only thing I think that will change those temps is lowering your OC a bit or using much louder big nasty fans on the radiator like the high performance corsair sp120's or if you really want some fan power you can go with one of the Delta's and possibly adding some on the other side of it so you have a push pull setup (With some big nasties). Delta(Very loud but very, very effective.)

http://www.sidewindercomputers.com/de12tfexhisp.html

What fans are you using on that right now and did you say they are 80mm (x3) fans or that the radiator is somehow 80mm? If you are using three 80mm fans you could definitely improve on that by going to two 120mm or even two 140mm fans. Either way would be more efficient as three 80mm fans would create a lot of unnecessary turbulence that's counter productive. IMO.
m
0
l
August 9, 2014 10:53:21 PM

Thanks for the help mate.

I am using 2x140mm fans but the rad has room for one more. And 3 on the other side, i would prefer not to have 6 fans blowing though. The rad is 420mm long but 80mm thick.

My concern is that I know the water temp and the CPU temp seems to be too much higher. Either the water block is not making contact well or the IHS on the i7-920 wasn't top notch. Understandable considering it was never a high end unit to begin with.

Thanks again for taking the time. It helps to have someone say it appears normal. You can't check the TIM without ruining it so there is always a little doubt.
m
0
l
a b K Overclocking
a b à CPUs
August 9, 2014 11:03:24 PM

If you used a pea sized amount on that 45nm lid and you can't see any coming out on any of the four sides, and you're certain it can't safely be tightened any further, and nothing is obstructing the contact of the block then I'd say the paste job is fine. You might really want to add another fan on the end. It would probably provide a significant drop in temp.

What brand and model are those 140's you have on there. As you probably might know, but maybe not, the air pressure rating of the fans makes a bigger difference when trying to cool a radiator than strictly going off the RPM's. It might be that DIFFERENT fans could also improve your configuration.
m
0
l
a b K Overclocking
a b à CPUs
August 9, 2014 11:13:18 PM

Verification of needed pressure for that radiator as follows:

This picture shows that the Monster needs a lot of pressure for ventilation, with two fan layers in push/pull configuration if possible.

The available fans show in push mode that the Gentle Typhon D1225C (max. 1850 rpm; 25mm), the Noiseblocker B12-4 eLoops (max. 2400 rpm; 25mm) and the thick Yate Loon D12SM-12D (max. 1350 rpm; 38 mm) are superior ( > 1 K lower Delta) to the used standard fan YL D12SH-12 because they create more pressure.

Full testing here: http://www.xtremesystems.org/forums/showthread.php?2872...
m
0
l
August 10, 2014 5:40:44 PM

I am sure that more fans would lower the water temp. My issue is that the water temp is fine. Barely above ambient. 18c water and 16c ambient. Sorry I don't know Fahrenheit for those. But the delta remains the same. I have water that is at 18C running over a CPU that is staying at nearly 30C. 12C difference between the water and the CPU seemed like too much. I don't know how much cooler the water can get to be honest. There is only 2 degrees difference.
m
0
l

Best solution

a b K Overclocking
a b à CPUs
August 10, 2014 6:19:41 PM

Well, that really only leaves three possibilities that I can see.

1. You are using the wrong amount of TIM. Too much or too little creates the same problem, insufficient transfer of heat to the conduction mechanism.

2. Heatsink/Waterblock is NOT sufficiently clamped to the cpu lid. Once again, allowing a reduction in the ability to transfer heat from the cpu to the heatsink.

3. CPU voltage is higher than can be tolerated. This may be due to an overclock done using more voltage than multiplier. We've seen many stable overclocks where someone had, for example, use 1.3v or higher to OC an i5 or i7 cpu but relatively small multiplier. The same stable overclock was able to be achieved using 1.2 or 1.25V and a higher multiplier, which also resulted in a dramatic reduction in heat generation from the cpu.

Short of there being some kind of blockage in the water block, lines or radiator, which could in fact ALL account for a low water temp with a high cpu temp, but which I'm assuming you have already ruled out, those three are the only things I can imagine would cause an issue of heat not being transferred from the cpu into the heatsink and thusly to the radiator.
Share
!