H80i v2

Had a problem with my old replacement h80i which did absolutely nothing. Got a replacement h80i v2, but it just plain and simply sucks. I bought a evo 212+ as a temporary measure when my original h80i stopped working, and the evo beats the v2 hands down. Installed the h80i v2, had it on performance in Corsair link and it sounded like an airplane, also had prime 95 running and it hit 86 degree celsius at one point. Evo has now maxed at 78 degrees after 12 minutes of stress testing.

Well, is the h80i v2 this bad or did i yet again receive another bad product from Corsair?

Take a look at this prior post, Intel doesn't recommend Prime95 for stress testing. The reason is posted below.

one particularly punishing test (Prime95) the last core was still staying under 90C while the rest of its pals were already thermal throttling.

Intel doesn't recommend Prime95 for stress test since Ivy Bridge. Prime95 can lead to degradation and or smoking the CPU. I would question anyone's knowledge who recommends Prime95 for stress testing unless the system was built on a Sandy Bridge or prior platform. The system was throttling because it was overheating.

For best results use Aida64 or Intel Burn Test 'IBT' for stress testing. Some use OCCT for testing since it free to use. All three test will perform with similar results.

Had a problem with my old replacement h80i which did absolutely nothing. Got a replacement h80i v2, but it just plain and simply sucks. I bought a evo 212+ as a temporary measure when my original h80i stopped working, and the evo beats the v2 hands down. Installed the h80i v2, had it on performance in Corsair link and it sounded like an airplane, also had prime 95 running and it hit 86 degree celsius at one point. Evo has now maxed at 78 degrees after 12 minutes of stress testing.

 

Well, is the h80i v2 this bad or did i yet again receive another bad product from Corsair?

Is the pump running at an expected speed? (I think the pump should be run at maximum speed no matter what.)

Also, try remounting the CPU cooler. When removing the cooler's CPU block, see if the thermal paste was spread properly. Make sure you're using the correct standoffs for your socket type.

It'd be odd to see a 212 EVO outperforming the H80i v2 by a significant margin, as even a H60 should be about equal, if not somewhat better than it. If after remounting it (the unofficial guide here is a good place to start) and you're still getting unusually high temperatures, see if it does the same thing at stock settings.

If the temperatures seems persistently high, you might need to RMA a second time, I think. Sometimes duds just happen.

Also, try doing a comparison with a less stressful test than Prime95, as mentioned.

Take a look at this prior post, Intel doesn't recommend Prime95 for stress testing. The reason is posted below.

 

 

 

Intel doesn't recommend Prime95 for stress test since Ivy Bridge. Prime95 can lead to degradation and or smoking the CPU. I would question anyone's knowledge who recommends Prime95 for stress testing unless the system was built on a Sandy Bridge or prior platform. The system was throttling because it was overheating.

 

For best results use Aida64 or Intel Burn Test 'IBT' for stress testing. Some use OCCT for testing since it free to use. All three test will perform with similar results.

Not sure how rigid you'll be in your opinion, but what is known is that the later versions of Prime 95 is NOT recommended. Using an earlier version that doesn't generate that kind of heat (26.6 or something) is what's being recommended, which i also used.

The pump was at 2600 or so rpm when i checked on the link program's main screen. Temperature was fluctuating like crazy, 35-40-55-39-55 etcetcetc.

Ohh and i did remove all the thermal paste and reapply fresh with Noctua DT-H1, same result afterwards.

Not sure how rigid you'll be in your opinion, but what is known is that the later versions of Prime 95 is NOT recommended. Using an earlier version that doesn't generate that kind of heat (26.6 or something) is what's being recommended.

Unfortunately it's not our opinion, it's based on the fact that Intel doesn't recommend Prime95 for stress testing since Ivy Bridge. Additionally Asus, EVGA and MSI also do not recommend Prime95 regardless of what version is being run for testing. Prime95 and LinX will cook your CPU without the proper cooling capacity.

Regardless of whether you use Prime95, SuperPi, or a blowtorch for your stress testing, the Evo should not come in ahead of an H80i. It is unlikely you received two consecutive coolers with some basic defect that only slightly reduces its effectiveness. Normally if a cooler is defective, you would not be able to run Prime 95 for more than a few seconds before throttling kicked in.

The place to start with your cooler examination is the water temperature (H80i v2 Temp) in the last box on the Link main page. On cold boot it should be right about your room temperature and then settle in a few degrees above that as the system warms up. A large difference in H80i Temp compared to room temp at idle would suggest a problem. The water temperature should be relatively slow to rise and fall. Sudden large jumps in water temp are concerning and/or a water temperature that continually climbs. Also, tell us where the radiator is mounted and whether it is intake or exhaust. Most people rear mount it as exhaust, but GPU loads do affect water temperature as well, so we need to know in order to compensate.

The erratic behavior of your CPU core temps suggests another type of problem, but I would like to rule out pump or flow problems on the cooler before tackling that. When you get strange or unexpected results from one stress program, it's always a good idea to run another to compare differences. Not all 100% loads are created equal. Prime95 is a particularly tough program. When you are having cooling issues, it is too blunt of an instrument for diagnostic use. I like AIDA quite a bit and OCCT is very good for stability testing some aspects. However, plain basic Intel XTU is useful when you have temperature issues. It is a mild stress test by comparison. Unlike Aida, which has quite a bit of core temp variability, XTU produces very smooth sine wave like core temps across a line graph. This makes it easy to pick out contact issues or if one or more cores are behaving in an unexpected way. It will also keep you well short of the throttle point.

Unfortunately it's not our opinion, it's based on the fact that Intel doesn't recommend Prime95 for stress testing since Ivy Bridge. Additionally Asus, EVGA and MSI also do not recommend Prime95 regardless of what version is being run for testing. Prime95 and LinX will cook your CPU without the proper cooling capacity.

Well, i find it entirely obvious that they try to dissuade you from using it because the latest version "might" cook your cpu. 26.6 certainly won't cook your cpu, that's for sure. But you know, try it. Not just say what this and that told, because they have agendas, and in this case it's a good one, but still a half lie.

Regardless of whether you use Prime95, SuperPi, or a blowtorch for your stress testing, the Evo should not come in ahead of an H80i. It is unlikely you received two consecutive coolers with some basic defect that only slightly reduces its effectiveness. Normally if a cooler is defective, you would not be able to run Prime 95 for more than a few seconds before throttling kicked in.

 

The place to start with your cooler examination is the water temperature (H80i v2 Temp) in the last box on the Link main page. On cold boot it should be right about your room temperature and then settle in a few degrees above that as the system warms up. A large difference in H80i Temp compared to room temp at idle would suggest a problem. The water temperature should be relatively slow to rise and fall. Sudden large jumps in water temp are concerning and/or a water temperature that continually climbs. Also, tell us where the radiator is mounted and whether it is intake or exhaust. Most people rear mount it as exhaust, but GPU loads do affect water temperature as well, so we need to know in order to compensate.

 

The erratic behavior of your CPU core temps suggests another type of problem, but I would like to rule out pump or flow problems on the cooler before tackling that. When you get strange or unexpected results from one stress program, it's always a good idea to run another to compare differences. Not all 100% loads are created equal. Prime95 is a particularly tough program. When you are having cooling issues, it is too blunt of an instrument for diagnostic use. I like AIDA quite a bit and OCCT is very good for stability testing some aspects. However, plain basic Intel XTU is useful when you have temperature issues. It is a mild stress test by comparison. Unlike Aida, which has quite a bit of core temp variability, XTU produces very smooth sine wave like core temps across a line graph. This makes it easy to pick out contact issues or if one or more cores are behaving in an unexpected way. It will also keep you well short of the throttle point.

I mounted the Radiator in the back with both fans as an intake. And it's certainly not the cpu, if that's what you were thinking. If so i would see the same behaviour with the evo, thankfully i don't.

Well, i find it entirely obvious that they try to dissuade you from using it because the latest version "might" cook your cpu. 26.6 certainly won't cook your cpu, that's for sure. But you know, try it. Not just say what this and that told, because they have agendas, and in this case it's a good one, but still a half lie.

If you don't agree with Intel's recommendations, please feel free to contact them. Here is the link: http://www.intel.com/content/www/us/en/support/contact-support.html#@12

I mounted the Radiator in the back with both fans as an intake. And it's certainly not the cpu, if that's what you were thinking. If so i would see the same behaviour with the evo, thankfully i don't.

The radiator fans should be mounted as exhaust to remove the excess heat from the system and not dumped inside the case.

Second, take a look at our guide and make sure the waterblock is installed correctly.

http://forum.corsair.com/v3/showthread.php?t=139165

I mounted the Radiator in the back with both fans as an intake. And it's certainly not the cpu, if that's what you were thinking. If so i would see the same behaviour with the evo, thankfully i don't.

No, I don't think it's the CPU. That would be the absolute last thing after every other possibility is exhausted and the circumstances here do not suggest this is the issue. Examining your water temperature is easy. The step after is more tedious and it will be more obvious if you are convinced the cooler is functional. If it's not, then there is no reason to go further. Do the water temperature checks first.

I also agree you should turn the H80i around to exhaust. I know what the silly instructions say, but it is more of a disclaimer than bona fide directions for the best set-up. That is something different for every environment. Most people are better of as exhaust. There is a misconception that cool air somehow makes the CPU colder. It's not the temperature of the air passing through the radiator but the temperature of the water in the cooler that will determine results. The physical environment (the inside case temp) will play a factor in that, in addition to the obvious CPU waste heat. So when you start dumping hot air directly into the case, you will raise the internal case temp and everything else, including the water temperature. Add GPU waste heat and you will get start to pile on some degrees. It's almost always better to dump waste heat directly out of the case, unless something forces you to do otherwise.

However, for now you may want to leave it as intake so you can make accurate comparisons. In short CPU only tests, there shouldn't be much difference. The longer you run at any load, the more the temps will climb for the CPU, water temp, and every other component.

It seems that one of the tubes has duct tape sticking out from the fitting.

Can you post a photo?