Kraken X61 strange temperature

[einzele]

hello guys, i've recently just upgraded my cpu cooler with Kraken X61.
after i've installed the cooler, i checked my 4790k idle temp was around 26 and prime95 temp was around 50s both using silent mode, so i couldn't be happier with the result.

the story doesn't end here unfortunately, because after i've used the pc for a couple hours, my idle temp raised to 35C and prime95 to 60+C. i tried restart my pc, let it off for some time then turn it back on, the result still shows the same the 35C idle temp.

i was using CM Hyper 212 before, idle temp was 38-40 and prime95 temp was 65+C, it is still an upgrade but i'm just wondering if something went wrong?

i don't do overclock, only use the feature that force all cores to run at 4.4GHz. i bought this for aesthetic reason.
i use NZXT H440 case and Noctua NT-H1 thermal compound. i put the radiator on top of my case as exhaust in pull (fan on top of radiator pulling air out of the case).
i don't have tool to measure my room temp, but i live in tropical country and my aircon is set at 22C

 

[JackNaylorPE]

CLCs don't really bring a massive cooling increase to the table .... not sure why users expect otherwise. Yes, the X61 will give you better temps than a Hyper 212, but there are better air coolers that perform better than the Kraken for less money.

The Kraken, and every other CLC, still depends on air for cooling. The large area heat sink has just been moved from the CPU to the side of your case. When you start up a PC, the coolant is at room temperature. The coolant's temperature will rise slowly and continue to rise as the coolant acts as a thermal mass.... it will absorb energy until it reaches an equilibrium point where heat in = heat exhausted. For example.... when I run Furmark, GPU temps (Cards are WC'd) might be say 28C. Over the first 10 minutes or so, temps will rise to 39C and then remain stable as long as the test is running. In the mean time, the coolant temp is rising from say 23C to 33C.

Many folks, remembering 8th grade earth science, (hot air rises) go against most manufacturer's recommendation and install the fans backwards as exhaust ..... violating one of the primary rules of water cooling, two of which are:

a) Never mix metals (copper blocks / aluminum rads)
b) Always install fans as intakes

Sitting in a room, hot air does rise, slowly, but not in the presence of a ceiling fan. So no, hot air doesn't rise when a fan is present because the fan is way more powerful than the slight difference in air density. When running with an ambient room temp of say 23C, a coolant temp of 33C and an interior case temp of 28C...here's what happens.

Exhaust ... Your 33C coolant is being cooled by 28C air; Delta T = 5C
Intake ... Your 33C coolant is being cooled by 23C air, Delta T = 10C

So as intake, your X61 will be able to remove twice as much heat as when used as exhaust.

Also what oft happens is peeps wind up with two intakes on the front , a rear exhaust fan and two exhaust fans on top. With 3 unfiltered exhausts and two filtered fans as intakes, you have about twice the air out capacity and you do intake capacity since the filters are an obstruction and lose efficiency. So with more out then in, you are sucking in air from the rear grille bringing dust and any hot air exhaust from the GFX card(s) back in with it.

 

[GhislainG]

The CPU temperature difference really is minimal. If you were correct, then NZXT would certainly need to revise their installation procedure: https://www.nzxt.com/manuals/krakenx61/X61-115X.html Reviewers also mount it as per the instructions.

 

[einzele]

hi, thanks for your time answering my problem.
though i think you are missing my point. i'm not asking why my kraken doesn't perform well as it should, but rather why it performed better at first, then became worse in mere hours, until now.

as for my setup, i deliberately place it as exhaust because i want the fresh air cool my GPU more.
and it is easier to clean the radiator in pull since i only need to spray the radiator directly with air as opposed to have to remove the fans if it is in push setup.

 

[JackNaylorPE]

I have read the NZXT manual. It doesn't say anything about which way to install other than sticker side down and having never seen the X61 I am in no position to address what they are saying. I inferred from the image that they are exhausting but whether that is by intent or just use of stock art I can not say.

Buying a product from Asetek and reselling it doesn't put me in position where I can say that NZXT's expertise in this area is enough to throw out the majority of manufacturers who specifically state to install blowing into the case. Can't see the logic where two sellers buying from the same OEM have different instructions. One of them has to be wrong. And like climate change, when 97% say one one thing, and 3% say the other, gotta go with the majority, especially when science says that's how it should be.

As Corsair and most other CLC sellers state to install blowing in, the custom water loop industry all says blowing in and the laws of thermodynamics unequivocally show the superiority of blowing in, then I'm going to go with the majority and the laws of thermodynamics.

Maybe if they installed like the majority of the industry, the Kraken could fare better against cheaper air coolers.

Here's an example of what other CLC manufacturers say:

http://www.corsair.com/~/media/Corsair/download-files/manuals/49-000175_rev_AB_H100i_QSG_web.pdf

Interesting note:

Looking at the picture for Step 1, just like w/ NZXT, it would appear that the fans should be installed blowing out. Now read the text:

"For the best cooling performance, we recommend mounting the fans as air-intake into your PC case".

I hardly think that changing the logo on the side of the unit somehow makes one perform better one way and the other perform better the opposite way. The simple fact remains, 23C air cools the coolant and therefore the CPU better than 28C air.

A cooling systems performance can be reduced to a consistent Delta T.... if the Delta T is 10C, then it will remain 10C now matter what the cooling air temp is.

23C ambient will get you 33C coolant
28C ambient will get you 38C coolant.
33C ambient will get you 43C coolant.

Now, obviously, a cooler coolant will result in a cooler CPU. If you want to preheat the air before blowing it thru your rads, you simply must expect the coolant to be higher temperature than it would be if you didn't preheat it.

Our test rig uses six thermal sensors (accurate to 0.1C) and a fog machine to analyze air flow and cooling performance on radiators. There are 10 radiator fans (280 + 420 rads) and 6 case fans. In every scenario tested, intake had lower CPU temps, lower GPU temps and lower coolant temps.

 

[einzele]

maybe you should read my post again? because you are completely out of mark.
true, placing radiator as intake will give lower cpu temp. but the gpu will suffer as it is eating the warmer air that enters through the radiator.
and thats not even what i asked.
right after i installed the kraken, my idle cpu temp was 26, then couple hours later the idle temp raised to 35. sure, the liquid temp might have raised as well. but even after giving it hours and days to cool the liquid, my cpu temp is still 35.

 

[JackNaylorPE]

Not factual. Again, using a fog machine to monitor air flow, we see that air from the radiators never gets down to the GFX cards. Funny thing about air coming into the case, it has to also get out. So when you have the typical case fan arrangement with a two fan CLC....

You have 2 fans blowing in from the top ... so where does that air get out ?

Well obviously 1+ fans worth of intake air will get immediately exhausted thru the rear case fan. In reality it's much more than that as the rad fan had reduced flow due to the radiator's restriction. In addition, with the air as intake, you are creating the very condition you said you were trying to avoid. With those top fans blowing out, unless numerous case fans are installed to compensate for the dearth of intake air, the hot air exhausted by the GFX cards is sucked wight back in thru the rear grilles where it goes right back into the GFX cardcs !

Think about what you just said. You are worried about the impact of heat from the 84 watt CPU (say 130 w/ 4.8GHz OC) impacting the GFX cards but you have no concern about 1 or more 200 - 300 watts GFX cards impacting the CPU ? That makes no sense.

1. The GFX cards in this instance are not water cooled, thereby indicating that GFX card temps are not a concern.

2. An overclocked 970 pushing 200 watts will have much more of an effect on a 130 watt CPU, than visa versa.

I did answer your question ... but given what was posted, not sure if it applies. Read the part about "thermal mass". It always takes time for a system to "reach equilibrium". If your first test was a quickie, it might not have had a chance to rise to the temps you are seeing now.

Again you should run Furmark and see what is happening when under a constant load, it's hard to use idle as a reference point w/o a thermometer and definitive understanding as to what is running on the system when at idle "behind the scenes".