H100i V2

Hello guys since some weeks i bought a Corsair h100i V2 to my I7 4770k (Not overclocked) but im starting thinking this temps are not normal in idle i get 38-42ºC and in full load i get 69-75ºC when is in full load the liquid does not get more hot that 35-36ºC.

When i was mounting the backplate for the first time I noticed the backplate is a little bit loose even with all four screws installed but i saw in another forums that is normal.

And i dismounted and mounted again and i get the same because i throught the pump doesnt not sit correctly on cpu.

I dont even know if this temps are good of if i have any problem with my cooling system

Can you help me guys? Thank you :)

It would be of great help to know your room temperature! Do live in a place with hot climate?

The room temperature is usually about 23º thanks to the air conditioning if i turn off the air I would have min 28º in my room because i live in Seville Spain and the summers here are very hot xD

Did you clean with alcohol and repasted when remounted? Also, what is the configuration of your case (intakes and exhausts). It would be helpful if you post a capture of your corsair link.

There are a lot of factors that influence idle and load temperatures. At idle C-states, Windows Power states, and the aforementioned room temperature all affect the base temp. Even at the stock frequency, load voltage usually varies for each unique CPU. It's hard to categorize some of this without that information.

 

However, I am a little concerned about the bracket. Yes, it should be loose when putting it together, but then feel secure once tightened down. Another thing to look at is your H100i v2 Temp in Link. This is the coolant temperature. If you are not making good contact, the coolant will show little change (2-3C) when at load. Even at stock frequencies, I would expect a +6C rise within 10 minutes or so. A final resort would be to take the block off and look at the thermal paste. If you didn't get it right, there will be a clear area of the cold plate with intact TIM.

Did you clean with alcohol and repasted when remounted? Also, what is the configuration of your case (intakes and exhausts). It would be helpful if you post a capture of your corsair link.

 

Yes everytime i remounted i cleaned with alcohol and repasted but i cant do more because i dont have more thermal paste

Here is the pic of the corsair link

 

There are a lot of factors that influence idle and load temperatures. At idle C-states, Windows Power states, and the aforementioned room temperature all affect the base temp. Even at the stock frequency, load voltage usually varies for each unique CPU. It's hard to categorize some of this without that information.

 

However, I am a little concerned about the bracket. Yes, it should be loose when putting it together, but then feel secure once tightened down. Another thing to look at is your H100i v2 Temp in Link. This is the coolant temperature. If you are not making good contact, the coolant will show little change (2-3C) when at load. Even at stock frequencies, I would expect a +6C rise within 10 minutes or so. A final resort would be to take the block off and look at the thermal paste. If you didn't get it right, there will be a clear area of the cold plate with intact TIM.

The idle c states i've never touched that so I guess it will be active and this is the load voltage (i was stressing only the cpu)

So when i get more thermal paste i would check if there a clear area of the cold plate

I will check the cooling system when i can thank you for answering guys :)

When you say the backplate is loose with screws installed, do you mean loose with the waterblock installed, or just with the standoffs before installing the waterblock?

Just with the standoffs before installing the waterblock

OK, well then the bracket install sounds normal -- loose then secure after full install.

 

In your screen shot, your idle temps looks fairly normal. If you had a genuine contact issue, you would not be able to maintain those low 30's core values.

 

If you don't have good contact, not all of the CPU heat will transfer across the cold plate and into the coolant stream. Your 27C baseline H100i v2 Temp (coolant) is what I would expect and is likely 4-7C over room temp. If you have poor contact, when you run a stress test like OCCT, XTU, etc., the coolant may only go up a few degrees. I would expect at least a +6C rise in 10 minutes or so when running those programs or even just general gaming over a longer period. If your coolant temp never goes beyond 30C, but with very high CPU temps, that might indicate an issue. Try to avoid using things like Intel Burn Test for this. Without a properly tuned BIOS, you are likely to hit core limits much too quickly.

When i putted the waterblock i made sure it was fine even pushes it down for if bad contact but the temps was the same

The only thing i can do now is the stress test here are the comparative

Before the test

http://i.imgur.com/WGovldm.png

 

 

And after 1 hour doing the stress test

http://i.imgur.com/lgAu25A.png

 

When i get the thermal paste i will disassemble it and see how it has expanded and i will change the thermal paste.

I don't think you need to change the TIM. You used Intel Burn Test anyway and are peaking in the low 80's in Summertime. That is a decent result. If you want to feel better, stop using IBT and run a normal stress test. You'll probably come back 7-10C lower.

But i think those temperatures are a problem since in the future I would like to overclock it

That's fine. Do you need to overclock and run the hottest stress around? Hopefully not. IBT does not equate to any normal task. It is exactly what it sounds like -- torture. Like any pressure tactic, it is going to find the weak spot and amplify it. For most most every one with a current 4 core Intel CPU, that is the material between the cores and the IHS. Heat is not able to transfer through the material fast enough to keep the cores from getting too hot when the voltage and/or load gets high enough. If you dig up some old 4770K reviews, you can read the sentiments at the time about the temperature in stock configuration.

 

My original suggestion still stands. Run AIDA64 blend, OCCT, Asus Real Bench or a more balanced stress test to give you a more realistic top line. Leave Prime95 and IBT alone for now. I am also unclear about your current Vcore. The second load Link shot does not make sense for me. Are you running on the AUTO voltage setting? What is the peak Vcore value you see when running at 100%? You like can tame this down for an immediate temperature reduction.

About the Vcore yes im running in Auto here the pics if you want check http://imgur.com/a/bgaHz

From the AIDA picture, peak Vcore is 1.15v, although that might not have been in synthetic stress test conditions where it can draw more. OCCT names stuff wrong. That 0.863v is not your Vcore and is some other value. On mine, it labels the Input Voltage as Vcore, another common glitch. Interesting that Link is grabbing the same value. Might be a quirk with your motherboard model, but that kind of thing is not my area.

 

What kind of temps did you get on OCCT? You can confirm the load Vcore from OCCT by looking a some of the other slides. It is probably going to be CPU VID. Look for a value close to 1.15.

Yeah the Vcore i think it really is the VID

http://i.imgur.com/obIJPno.png

Btw today i change the thermal paste and now if i run some stress test (Aida64 it was only 10 minutes because I did not have more time)

the avg temp is 58-60ºC and the maximum peak was 65ºC

Not bad at all

No, I think that is good and certainly in light of the season. When it finally cools off in October, your temps will drop even more. There still is head room and what really matters in your average and peak core temps when doing whatever it is you do. These stress tests are overvalued. Nobody's PC accidentally started encoding video one day or hunting for Mersenne Prime numbers. The best test of your system is to use it for its intended purpose.