Just installed Corsair H50 Cooling system..how do I know its working?

[Obersturmfuhrer]

I have Rampage II Extreme MB with I7 930 processor (nothing special).

Looking to OC it, preferably to 3.8-4.0 ghz.

Already OC'd it with stock cooler, but I was reluctant to run any strain tests on it in fear it would fry it...So I Went out and bought Corsair H50 cpu water cooling system. (which was a real u-know-what to install, I didnt realize that the other side of my case had an opening to access the rear part of the CPU motherboard until AFTER I unscrewed, then removed everything, and installed the blackplate. lol)

Now I install cooler, and run x20 test with it @ 360 Mhz. As the test persists, it eerily eeks it's way closer and closer to 90C, going back and forth at times (when a load is breifly taken off of it) to 67C, 68C, then right back up to 90C... I dont remember exactly what it reached but it was very near TJ Max levels twoard the end.

What I'd like to know is, I thought with this great cooler I'd have temps well under control.. well under perhaps 80C on full load... Did I install it correctly? How do u know for sure if you've installed an aftermarket cooling system right (particularly this one)??

Any feedback is greatly appreciated.

 

[J_Sparrow]

"peaking of the radiator. Corsair need the H50 installed with the airflow effectively backwards. What is the standard exhaust port becomes the intake drawing cool air across the rad. "

so to confirm your pushing air out of your system at the back.... YOU SHOULD BE drawing air in from the back of the case and pushing into your case? COS THATS WHAT IT SAYS IN THE INSTRUCTIONS!!!! DO IT! I know it sounds crazy, but thats how they designed it.

and use fans of the same type and speed, cos 2 different fans cause 'dirty' air and the flow gets all fucked up

-i hope this is all it is, and that if you reverse fans we fix the fault!!!! :O) _

 

[J_Sparrow]

http://www.legitreviews.com/article/1025/6/

this is roughly what temps you should expect

-edit- u got your fans like this pic?

so i guess if you have a front panel fan that will need to be outtake rather then intake, the side can still be intake and top still out..

 

[Obersturmfuhrer]

http://www.legitreviews.com/article/1025/6/

this is roughly what temps you should expect

-edit- u got your fans like this pic?

so i guess if you have a front panel fan that will need to be outtake rather then intake, the side can still be intake and top still out..

Yes, the fans are EXACTLY like that... Like a sandwich, one attached to case is is sucking air in and fan on other side is blowing hot air OUT.. so the airflow is effectively traveling in the same direction through the whole system...

But yeah its exactly like that and same two ULTRA fans 120mm being used... The temps that are listed in the article u linked are SOMEWHAT similar to mine, but mine are obviously about 5-10C higher in some tests.. I wonder why? I noticed that the tubes are position upwards and mine are below, but ive read that makes no difference..

What the article doesnt say is how LONG of a cpu load did those temp listed result? I wish they speicified that, u know, like after 5 mins? 10 mins? etc....

Ive also done some research on compusa.com and it doesnt look like they have anything that would yeild better results.. maybee ill just exchange it for another one, maybe i just got a bad unit.. I just dont know..

 

[J_Sparrow]

well going from what i know from water cooling i've done and air cooling aswell, i'd say you should get a steady +/- 5C after about 10 miuntes on prime95. send it back mate, get some nice air cooling instead if it will fit. i use this:- http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/34658-cooler-master-v6-gt-cpu-cooler-review-2.html

 

[gerry410]

I ve already re-set it 4 times. I've purchased 3 more fans, one of which I implemented using the push/pull configuration for the radiator.. So I dunno whats up with this thing... It's idiling at mid-40's now.. Running 21x @191 = 3.99 mhz, at 1.19v. Does this sound any better or is it still high?

I think before you start worrying about over-clocking you need to get the thing running below 40c and stable. If you getting that its getting better.

The thermal paste can be a real bitch. NO more than the size of a pea. And when you squeeze it out have the tip right against the cpu. You dont want any air bubbles to form in there. Some say to much don't matter. I think thats bs. the heat gathers in the excess with no where to go and thats how your temps go high.

Did you utube your cooler installation? Thats what i did for mine just to make sure it was correctly installed.

if your getting a steady 35/40 and u run prime95 it will go to 65/70 and thats OK.
Prime really maxes out the cpu and memory. Go to weki-pedia and read the article about it.
Most of all take your time and be sure of what yer gonna do. don't be in a big rush to oc to 5.0. get the thing running stable first.

40c sounds good. keep in touch.

 

[J_Sparrow]

a peas big!, I used an amount the same size as a grain of rice or maybe a tiny bit more. its only there to fill the micro gaps in the metal i think.

-edit-

"Thermally conductive paste improves the efficiency of a heatsink by filling air gaps that occur when the irregular surface of a heat generating component is pressed against the irregular surface of a heatsink, air being approximately 8000 times less efficient at conducting heat (see thermal conductivity) than, for example, aluminum, a common heatsink material.[5] Surface imperfections inherently arise from limitations in manufacturing technology and range in size from visible and tactile flaws such as machining marks or casting irregularities to sub-microscopic ones not visible to the naked eye."

check the results of paste on this page lol:-

http://www.dansdata.com/goop.htm

 

[gerry410]

If I did any stress test on it at all I know it would hit 90-100C in no time. I m beginng to care less about these stress tests and more concerned with the stability and temp rage during gameplay. As long as its stable and temps are acceptable during Mafia II, Gears of War, etc.. Im happy. No game out there puts a 100% load on any cpu, 100% of the time. So what is the test for, really?

You say your beginning to worry less about stress test and more about stability? Thats exactly what the stress tests test. Go to wiki-pedia and read the article about prime95.

good luck

 

[Obersturmfuhrer]

You say your beginning to worry less about stress test and more about stability? Thats exactly what the stress tests test. Go to wiki-pedia and read the article about prime95.

good luck

Your missing my point, Gerry. Lets say for instance I do a 4GHz OC. After an 30 mins to an hour of frying it with LinX, temps stablilize around 90, however POSSIBLY could reach higher temps if one were to leave it sit for longer testing.

To benchers, that temp is completely unacceptable. Nevertheless, in a "real world" gaming scenario (lets say 1 hr per day) temps of which do not exceed 75C at any one time during gameplay.

Is the system stable to run 10 hrs. at 100% load, probably not.
Can u enjoy ur games without worrying about it crashing, over heating, etc., probably.

But yes I realize stress tests are for stability. Im talking about how far do you take it..

 

[wolfram23]

^75C is still "too hot" for the CPU according to Intel specs, but if you're ok with it... then that's fine.

IMO, LinX isn't needed as a long term burn test. I like to run it (I use Intel Burn Test, actually) using 4gigs RAM for 15 passes. If it's stable, it's fine. Prime95 usually has lower temps but also needs to be ran for longer. At least 4 hours Small FFT gives a good indication of stability.

Not too long ago I changed my base clock from 175 to 177. Running 10 passes of IBT passed at 2.5gigs RAM so I thought it was fine. Then one day I experienced several crashes while gaming. Thought that was odd (worked fine for the last several days) but decided to run some stress tests. P95 had an error after about 1.5hrs, so I upped my VTT/IMC a pinch and reran with IBT at 4 gigs RAM usage and it showed an error after about 6 passes (more RAM is a more accurate stability test, if using less RAM you'll want to do 20-25 passes I'd say). Eventually with a pinch more Vcore and VTT it passed no problems and since then I haven't had any crashes. So IMO stability is very important.

 

[Obersturmfuhrer]

^75C is still "too hot" for the CPU according to Intel specs, but if you're ok with it... then that's fine.

IMO, LinX isn't needed as a long term burn test. I like to run it (I use Intel Burn Test, actually) using 4gigs RAM for 15 passes. If it's stable, it's fine. Prime95 usually has lower temps but also needs to be ran for longer. At least 4 hours Small FFT gives a good indication of stability.

Not too long ago I changed my base clock from 175 to 177. Running 10 passes of IBT passed at 2.5gigs RAM so I thought it was fine. Then one day I experienced several crashes while gaming. Thought that was odd (worked fine for the last several days) but decided to run some stress tests. P95 had an error after about 1.5hrs, so I upped my VTT/IMC a pinch and reran with IBT at 4 gigs RAM usage and it showed an error after about 6 passes (more RAM is a more accurate stability test, if using less RAM you'll want to do 20-25 passes I'd say). Eventually with a pinch more Vcore and VTT it passed no problems and since then I haven't had any crashes. So IMO stability is very important.

Yeah, I understand wolf... totally... Im just about to take the case apart and take out the H50, box it up, and take back to compusa and exchange for another one. Once installed, that'll at least elimiate one more troubleshooting problem..

Quick question, aparently the H50 mount comes with sin susi paste or whatever its called, considered suprerior to even AS5.... What do u think, wipe CPU block clean then apply, or remove stock paste and apply my own Ultra goop I got (meant to cover entire cPU block with very thin layer via a nail polish type applicator)?

Or perhaps ill try both in sequential order and monitor temps, if stock TIM works then Im good, if not I always have my own to try..

man I hope this works..

 

[wolfram23]

Well maybe someone else can reply as to which TIM might be better. As far as I know, there's been a few improvements over AS5 but I've never heard of the two you're mentioning.

As for application, I personally would advise against the pre-spreading. I found after multiple applications I got the best result with two very thin lines of TIM (like having 3 or 4 uncooked rice grains in a row) running parrallel to eachother and to the cores of the CPU, approx 1/4" appart or a bit less than that (and the end before the edge, otherwise it may flow over). Then just apply even pressure, and it will automatically fill any large gaps and get squished out of areas that are well-mated together.

 

[Obersturmfuhrer]

Well maybe someone else can reply as to which TIM might be better. As far as I know, there's been a few improvements over AS5 but I've never heard of the two you're mentioning.

As for application, I personally would advise against the pre-spreading. I found after multiple applications I got the best result with two very thin lines of TIM (like having 3 or 4 uncooked rice grains in a row) running parrallel to eachother and to the cores of the CPU, approx 1/4" appart or a bit less than that (and the end before the edge, otherwise it may flow over). Then just apply even pressure, and it will automatically fill any large gaps and get squished out of areas that are well-mated together.

I saw somewhere on the 'net a guy with an H50 mounted the block upside down.... doing htat right now, sort of a last desperation attempt, to see if it yeilds any better results.... just about to hook it up for a trial run

 

[Obersturmfuhrer]

So, bottom line is, 50C idle (cpu at 0 load) even at a 4.0 OC, is till WAY too high for the H50, right?

 

[gerry410]

Your missing my point, Gerry. Lets say for instance I do a 4GHz OC. After an 30 mins to an hour of frying it with LinX, temps stablilize around 90, however POSSIBLY could reach higher temps if one were to leave it sit for longer testing.

To benchers, that temp is completely unacceptable. Nevertheless, in a "real world" gaming scenario (lets say 1 hr per day) temps of which do not exceed 75C at any one time during gameplay.

Is the system stable to run 10 hrs. at 100% load, probably not.
Can u enjoy ur games without worrying about it crashing, over heating, etc., probably.

But yes I realize stress tests are for stability. Im talking about how far do you take it..

i was just replying to your statement, " I m beginng to care less about these stress tests and more concerned with the stability and temp rage during gameplay". Your temps at 90c are way high. Did you check out any videos on you tube to make sure its installed correctly? NCIS or NCIX on the tube has a lot of good video how to stuff. good luck bro. I feel your pain its very frustrating but be patient and read alot yoy will figure it out.

 

[Obersturmfuhrer]

Guys:

Have some good news... just came back from retailer, returned H50 for a brand new one... got it home, installed it right away.. fired up the comp, and was very pleased with a quick LinX run, OC'd at 4Ghz @ 1.283v... Ill show a comparison of BEFORE and AFTER..

Here's the latest run with new H50:



Now yes I know, 82C is a bit high, but remember Im at 4 GHz (21x 191) , and this is compared to THIS (old H50):



If you'll notice in that test I couldnt even complete it, as temps reached to 94C before I stopped it around 3:00... Notice the same exact CPU-Z specs between the two..

I have HT ON BTW, and for some reason in my CPU Config BIOS menu Turbo mode Enable/Disable has disappeared for some reason, it does that on occasion, sometimes I see it , someitmes I dont.. It's usually right above Intel SpeedStep(TM) option... Any idea what thats all about?

Anyway, currently idling at about 45C, not exactly overjoyed about that, but compared to the old system idling at 50-51C, it's a welcomed improvement..

Here's a Core Temp pic I've just taken this second of the current temps of my cores with the new H50 installed..



So, I guess Im pleased with the latest results, I'd like to do a bit more testing with prime95, espcially a longer-duration test to see what my max temps reach.. Low-80's with a 4GHz Oc'd id say Id be happy.. what do u guys think??? thanks for all urhelp BTW =) U think the old unit was a bit defective??

 

[gerry410]

Maybe tonight you can get a good nights sleep. Good job, much better

 

[wolfram23]

Definitely better, but I'd keep working on that OC to try and bring voltages down and hopefully temps. I understand that hyperthreading is going to push the voltage and temps up, and maybe your ambient is higher too, but at 4ghz I was only hitting around 72C on my i5 750 (200x20) and that's with a single fan air cooler (Zalman CNPS10X Extreme). My ambient is usually around 20-24C.

I've found Turbo option dissapears if I take EIST off auto. With 21x you can probably still force Turbo on by enabling CStates and EIST but that'll need more voltage and higher temps.

I'm just curious, what do you mostly use this PC for? Because, and this is just my own feelings, I find a lower OC using Turbo and the power saving features to be far more sustainable. 4ghz on 8 logical cores is overkill if we're talking gaming and day to day stuff, only for a hardcore amount of video encoding or file compression would I consider pushing it that hard all the time. Like my sig shows, I'm currently at 177 base clock with turbo on, so it reaches 4.25ghz at single thread stuff... max temps are under 70. 1.36V LLC off (so load voltage is less, which also brough down temps by 3C compared to LLC on with less voltage). The RAM is also nicely OCed at this base clock, as I have 1600mhz CL8 RAM running at 1416mhz CL7 which required a slight voltage boost to be stable (lower speed isn't slower if the latency is lower too)

But anyway, that's just my opinion. If you're happy now, well congratulations 😀