Corsair revisits CPU cooling: introducing the Hydro Series H50

[Cyber-Mav]

strange, i would have thought on xtreme systems people would be pushing everything to the extreme along with xtreme testing methodologies. as for regarding the health of my cpu if i was that scared of it blowing up i wouldnt even overclock it. to me if its not linpak stable its unstable. and i just hate using unstable machines to do my work on.

back on topic now, since this is about cpu cooler reviews the more heat used the better it is. most people here wanted to see if the corsair rad would reach a saturation point. using prime wont show any of that especially on a dual core cpu.

[JPQY]

Sorry,i'am also not agree..i have here more then 100 chess engines and many off them stress the cpu more then Prime and these are my daily programs!!
My system has to be stable with my chess programs,no matter what other bench or stress program you use!
So,the only result that matters me is to see how good the Hydro H50 can handle my chess program!
And if it can give me much better temps (like -7°C from my max. now) then i will buy it.

JP.

[Smartidiot89]

you guys are idiots.

not only does running such stress tests as IBT and linX for long periods of time physically shorten the lifespan of your CPU (even prime is bad for the CPU) but it also is pointless as leeghoofd says there is no REAL PROGRAM that will stress your CPU as much as prime and therefore it is not worth clocking for stability in anything more stressful than prime. period. You all think your proving something by running these extreme torture tests but in reality your just proving you have no regard for reality or the health of your CPU.

From my experience Linpack is better for stability testing atleast - and I never ever run it longer then an hour, more than that is just dumb. An hour has always been enough for me, after that I go on with running WCG 24/7, playing games and other everyday uses.

As far as Prime goes, it requires a longer period of time to show stability then compared to Linpack and I have to run it for 24h+ too feel satisfied...

Only speaking from my own experience on this one, but I personally prefer Linpack for my stability testing. The only one I am proving anything is to myself, that my system is infact stable.

[Leeghoofd]

running a few hours of linX is not going to significantly reduce the life of your CPU
and yeah these are extreme torture tests, but this is xtreme systems

I'm at about 24 hours of total testing just for this H50 unit per CPU (due to numerous test variables)
I'm awaiting 4 more new I7 blocks, got 3 new Enzotech base samples to test. I've got more than 120 hours of priming ahead for the next I7 roundup, if you guys think I'm gonna do LinX or such on this nice i950 you must be joking...

Secondly you must consider this is a single rad unit. I'm testing at 3.8Ghz for the I7 as this CPU seems to be a hot head too. It was already throttling with the stock cooler at 3.9GHz, so I backed down to 3.8 to just keep it under 100°C . This cooler is not foreseen for high OC's, it's an easy step up from a stock cooler and performs alike and depending on the case used and airflow, better than most high end air coolers.

Thirdly FYI LinX nor IBT give me a 100% stability test tool, I can succesfully run IBT at max mem and yet cap out in eg Hyperpi 32mb ( 8 threads) in the first iteneration. There's no such thing as the ideal tool, a mix is required... and the best test is daily usage.
Linx stable means absolutely nothing. It's all in between ya ears ! If you think you are extreme by using LinX or IBT good for you then.

Fourthly Corsair gave each reviewer a well documented what-to-do-list, I already deviated from that by using my own prime test instead of the large test ( this to keep the results in conformity with the already available database at Shrimps )


Cyber , my E7400 makes my phase unit running so hot at 5ghz ( while that modded Unit can hold the E8600 at 5.75Ghz bench stable ) So don't tell me about heat. This is one hot motha.

But to each his own... ask the other reviewers to use LinX or IBT...

Sorry,i'am also not agree..i have here more then 100 chess engines and many off them stress the cpu more then Prime and these are my daily programs!!
My system has to be stable with my chess programs,no matter what other bench or stress program you use!
So,the only result that matters me is to see how good the Hydro H50 can handle my chess program!
And if it can give me much better temps (like -7°C from my max. now) then i will buy it.

JP.

Link me to them plz, so I can compare temp results and will pm ya the results...thing why we use prime is as we are expanding the current database. Switching to another program means redoing all them coolers and waterblocks...

[sosna]

Firstly Leeghoofd thanks for the info-pic about this cooler.Hopefully you inish soon and post more info.

I test my machine for an hour with occt.If i get no errors then thats fine for me.
Untill now in my 24/7 use ive never had a bsod or other abnormalities.And only with an hour of occt.

I also dont get why guys want to torture this much their machines?

Everyone knows the use and potential of their machines.So why do i need to strees test for lets ay 24 hours when i m not going to do something that will require the system to be @100% for 24 hours?

ps:Leeghoofd,any guess when the reiew will be ready.Sorry but i really cant wait to see the results -good or bad-

[Leeghoofd]

I hope end of this week mate, gonna see AC/DC tomorrow and need to recuperate from that :p. Most of the writing is done. Just checking with other reviewers and Corsair for pointers...

One thing I discovered it that the fan runs about 70rpm faster on the I7 mobo then on the S775. So I need to check that too and hook the fan and pump straight to the PSU maybe

Btw Sosna Metaxa, Retsina rock !!

[sosna]

I hope end of this week mate, gonna see AC/DC tomorrow and need to recuperate from that :p. Most of the writing is done. Just checking with other reviewers and Corsair for pointers...

One thing I discovered it that the fan runs about 70rpm faster on the I7 mobo then on the S775. So I need to check that too and hook the fan and pump straight to the PSU maybe

Btw Sosna Metaxa, Retsina rock !!

Nice!

If the results are good enough i could send you some bottles!

ps:you should also try tsipouro!

[Dante80]

Better yet, both at the same time...the feeling is priceless....

Leeghoofd gj, we will be eagerly waiting...

[Ender17]

appears to be instock here
http://www.ncixus.com/products/40137/CWCH50/Corsair/

[Smartidiot89]

appears to be instock here
http://www.ncixus.com/products/40137/CWCH50/Corsair/

Butt were ze reviews?

[Ender17]

Butt were ze reviews?

I'm tired of waiting
I think it will have good results in my P182 as I have much less airflow compared to Kyle's case on the hardocp review

[Leeghoofd]

i7 results tonite lads... cooler works better on a quadcore and with a bigger case (better results with added fans etc than with the s775..) gonna pop in the TRUE this afternoon and will keep ya updated chaps... need to get a second mount to do push pull on the TRUE

[Ender17]

i7 results tonite lads... cooler works better on a quadcore and with a bigger case (better results with added fans etc than with the s775..) gonna pop in the TRUE this afternoon and will keep ya updated chaps... need to get a second mount to do push pull on the TRUE

how many hours? I'm going to bed in like 2

I've been thinking about the instructions to have the H50 fan blow into the case

seems like this would be bad for a case with a lower mounted PSU (Antec P19x) as the PSU will be exhausting hot air and it rises up along the back and then gets sucked back into the case by the H50 fan
also having the rear fan as exhaust seems like it would disrupt the normal airflow in the case
I'm definitely going to experiment with different fan configurations when I get mine
hoping the tubing is long enough that I can try mounting the radiator on the front of my P182 where the 2x hard drive cage normally is

[Leeghoofd]

Mmm let's say when you wake up results will be online. This if you have 6-8 hours of sleep . I'll measure the tubing length for ya, so you can see if you can make it to install the rad in front of the case...

[Ender17]

Mmm let's say when you wake up results will be online. This if you have 6-8 hours of sleep . I'll measure the tubing length for ya, so you can see if you can make it to install the rad in front of the case...

awesome, thanks man

[WhiteFireDragon]

Early test show that the air flow in the case is crucial to get it working right, even an extra Nb fan reduced it's performance by 1-2°c.

Case setup is very important, hence why I think we will see different results depending on the case and setup used.

can you please explain why this NB fan cause it to have lower performance? the airflow near the area of the rad is disrupted or something?

[Leeghoofd]

Yep it hindered airflow in my little antec case, after I retested everything in the Antec P160 case with an Antec spotcool temps weren't influenced...(other NB cooler was a 90mm on a Zalman bracket.

Good air flow is a must if you want this unit to perform optimal

[djalex]

Mmm let's say when you wake up results will be online. This if you have 6-8 hours of sleep . I'll measure the tubing length for ya, so you can see if you can make it to install the rad in front of the case...

I would like to see if H50 fit in front of Antec 902 in the HDD cage. May it give a better results than back mounting? Tubing length in centimeters?

[PhenomII2]

i7 results tonite lads... cooler works better on a quadcore and with a bigger case (better results with added fans etc than with the s775..) gonna pop in the TRUE this afternoon and will keep ya updated chaps... need to get a second mount to do push pull on the TRUE

Waiting eagerly ,

[Smartidiot89]

i7 results tonite lads... cooler works better on a quadcore and with a bigger case (better results with added fans etc than with the s775..) gonna pop in the TRUE this afternoon and will keep ya updated chaps... need to get a second mount to do push pull on the TRUE

Tonite Belgium time is tonite sweden time grovy!

[Ender17]

damn, I just woke up and no results

[Leeghoofd]

I would like to see if H50 fit in front of Antec 902 in the HDD cage. May it give a better results than back mounting? Tubing length in centimeters?

My unit max length is 28cm, realistically I would count at 25Cm max to allow some flexing (measured from the start of the tubing on the RAD till the mid of the cooling unit.)




Lol ender 17

I just had Gareth of Corsair on the phone, ( I had send him some of my early results ) Apparently there are some units that have some manufacturing issues. Mainly some of the screws on the bottom are not countersunk enough and sadly my unit is involved. On top of that my unit seems to be convex on one side and flat on the other. I didn't check the flatness of the bottom plate till a few minutes ago.

Other thing I asked which TRUE fan Corsair used in their comparisive test. He couldn't confirm i f it was the boxed one or a higher rpm version. At the moment he was using a noctua cooler and still got that 10°C difference in advantage of the H50. I confirmed him on my unit the TRUE got the better of the H50. This with the TRUE bundled with the 1600rpm fan (not the stock 1000rpm fan which is in the box) The TRUE gains by a few degrees.

He then advised me to check the screws and the flatness of the bottom plate. I also got this mail from him :

I’ve just been informed about a possible manufacturing fault with some H50 units. It’s limited to a small percentage of the units, and we discovered it before many units went out to reviewers or customers, but it’s possible that it may affect the unit you have.

The problem is with the countersink bore on the cold plate. In some cases, the screws that attach the cold plate to the pump unit may not be fully seated, which may reduce the surface contact between the cold plate and the CPU, lowering the cooling performance.

You can visually inspect your review unit to see if the problem exists. If your unit is affected then the screws in the cold plate will not be fully countersunk.

If you have a H50 with the problem, we will swap it out immediately, and we’ll do the same for any customer who has the problem, though we don’t believe there will be many. In fact, we haven’t had any reports of the problem from customers so far, so this is mainly a precautionary measure.


If anyone of you already got this unit and performance is not as good as advertised plz check the screws on the bottom

[Ender17]

guess I will go ahead and order it

[Leeghoofd]

Hold off mate till some more data pops up... I'm gonna cranck up the speeds to 4Ghz and see what happens then, bad unit or not. Once I get them results I post them, but be warned that other units might perform better

[Ender17]

Hold off mate till some more data pops up...

I'm tired of waiting
plus I would kinda like to do my own testing as I've said before, my P182 is pretty starved for airflow I think - seems like my case would be the optimal environment for the H50