Corsair revisits CPU cooling: introducing the Hydro Series H50

[mav2000]

For those of you asking - the initial shipment will not have AMD brackets inside the box, they will be available from Corsair upon request for just shipping & handling costs.

That really makes no sense to me. Why pay extra shipping and handling for something that should come with the box. Its ok for guys in the US or EU and UK, but what abt in Asia and India??

[PaVi90]

Directly from Corsair to me:

Please make sure that the 3pin connector of the unit which powers the pump, will not be regulated by the mainboards bios – it needs full power.

[Jamesrt2004]

Directly from Corsair to me:

makes sense to me, dont want it being undervolted too much so then it doesnt run!!!

[Leeghoofd]

You need to hook up the fan to the CPU fan header, mine drops to 1000rpm at idle and normal use (clearly stated in the reviewers manual), once I load a game or prime it spins up to 1660rpm. Pump is hooked to a fan header (always 100% set in bios) , always at 1450rpm...

[Ruroni]

You need to hook up the fan to the CPU fan header, mine drops to 1000rpm at idle and normal use (clearly stated in the reviewers manual), once I load a game or prime it spins up to 1660rpm. Pump is hooked to a fan header (always 100% set in bios) , always at 1450rpm...

Numbers now! :p

[PaVi90]

makes sense to me, dont want it being undervolted too much so then it doesnt run!!!

Yes, in fact... However it runs but in an "under-performance" way

[Leeghoofd]

CPU : E7400@4ghz ( 10 x 400FSB ) 1.35Vcore
Mobo : Gigabyte EP45 UDP3P
RAM : 4Gb OCZ Reaper
Case : Antec P180 (medium tower case)

With both air coolers, case tested in normal fan config ( front and rear 120mm fan)

With H50 :

Tested with front fan
Tested with extra top fan
Tested with extra fan on the rad (extra fan at Medium fan speed setting 1600rpm)
Tested with 2 high performance fans on rad (max fan speed setting 2000rpm)
Tested with 2 high performance fans on rad plus replaced TIM with MX2


This is really a hot CPU, I can bench it under air at 4.5 but that's where it stops, it's benchable till 5Ghz with Phase, with Cascade till 5.3

Some test screenies :



TRUE test



H50 Test (no added case fans)


I7 test is for the review :p more piccies and numbers there... don't jump too conclusions too soon, await full review and numbers from other reviewers !!

[Redbeard]

That really makes no sense to me. Why pay extra shipping and handling for something that should come with the box. Its ok for guys in the US or EU and UK, but what abt in Asia and India??

It's a matter of the AMD brackets being available from our supplier. The first run of them had significant issues so we had to scramble for another solution.

Our goal is to have brackets that are compatible with 775, 1156, 1366, and AM2/3 in the box. Right now it's 775 and 1366, and all units come in a white box instead of a retail box as it is still being printed.

[EagleRock]

It is very clear now that I'm not going to replace my TRUE. Was hoping for +10c different.

[Manicdan]

random questions, how many watts is the pump using? it might have been stated already some 5 pages back, but i cant keep up

[HousERaT]

It would be much appreciated if someone could test this water setup on an open air system (another words outside of a case).

[Redbeard]

It would be much appreciated if someone could test this water setup on an open air system (another words outside of a case).

It will not outperform a high-end aircooler outside of a case. It's designed to outperform high-end aircoolers in typical case configurations where ambient temperature is higher than outside the case.

[Cyber-Mav]

he is using prime to generate the load on the cpu? what happened to using linpak to do the job properly?

[Manicdan]

It will not outperform a high-end aircooler outside of a case. It's designed to outperform high-end aircoolers in typical case configurations where ambient temperature is higher than outside the case.

so is it recommended to set the fan to pull in air from outside the case? i would much rather have the radiator away from the case using standoffs. kinda like how you see radiators on the back of cases. it would use the outside air, and it would dump the heat outside. something no air cooler can ever do (unless you make your case custom to isolate the heatsink from the rest of the case, but good luck)

[Leeghoofd]

he is using prime to generate the load on the cpu? what happened to using linpak to do the job properly?

I don't like IBT and Linx, stresses my cpu too much. I'm not wasting a good CPU on a silly testing tool that heats up ya cpu way further than any app in this world will ever do... plus it's not the ultimate stress tool, it's THE tool that generates the most heat :p

so is it recommended to set the fan to pull in air from outside the case? i would much rather have the radiator away from the case using standoffs. kinda like how you see radiators on the back of cases. it would use the outside air, and it would dump the heat outside. something no air cooler can ever do (unless you make your case custom to isolate the heatsink from the rest of the case, but good luck)

I'll try to get that done mate for the final review using the Swiftech standoffs...

[HousERaT]

It will not outperform a high-end aircooler outside of a case. It's designed to outperform high-end aircoolers in typical case configurations where ambient temperature is higher than outside the case.

Thanks. Then I'll keep in mind that this isn't really for benchers but for people with closed case towers. I still might have to get one of these for my tower once all the data has been displayed. This looks like a good setup so far. This could replace the Tuniq on my AMD setup.

[Cyber-Mav]

I don't like IBT and Linx, stresses my cpu too much. I'm not wasting a good CPU on a silly testing tool that heats up ya cpu way further than any app in this world will ever do... plus it's not the ultimate stress tool, it's THE tool that generates the most heat :p

strange, it just seems to me that linpak would be the ideal tool to run especially if your doing cpu cooler reviews. i just cant see the point in clocking a cpu to 4ghz and using prime on it when the same cpu at 3.6ghz running linpak would generate more heat.

however i do understand that if its the only cpu you have available to test with then you dont want to be breaking it. but then the question arises, why overclock the cpu if you have fear of it getting hot.

looks like linpak is all it takes to fry shrimps these days.

[Smartidiot89]

I don't like IBT and Linx, stresses my cpu too much. I'm not wasting a good CPU on a silly testing tool that heats up ya cpu way further than any app in this world will ever do... plus it's not the ultimate stress tool, it's THE tool that generates the most heat :p

Which is exactly what people want in a cooler review...?

[Baron_Davis]

Unless this thing pulls a miracle out of its ass, I have a feeling a lot of people are going to be dissapointed.

Megahelms + San Ace's ftw?

[Cyber-Mav]

Which is exactly what people want in a cooler review...?

atleast someone sees things the way i do. id hate to go by the madshrimps review, e.g i see X cooler doing 55c on a q6600 at 3.8ghz, so im like ok ill get that cooler for my q6600 to run at 3.6ghz, then im like wtf why my cpu running 10c hotter at 3.6ghz with less voltage than the cpu in the review running at 3.8ghz?? turns out i used linpak and review used prime.

[Smartidiot89]

atleast someone sees things the way i do. id hate to go by the madshrimps review, e.g i see X cooler doing 55c on a q6600 at 3.8ghz, so im like ok ill get that cooler for my q6600 to run at 3.6ghz, then im like wtf why my cpu running 10c hotter at 3.6ghz with less voltage than the cpu in the review running at 3.8ghz?? turns out i used linpak and review used prime.

LinX and IBT has become stability tests many people use so it should be used in reviews to produce maximum heat...

But if people like Priming 24hours+ be my guest, I am using LinPack and I detect instability and reach stable clocks much faster then I ever managed to do with Prime...

[Cyber-Mav]

i agree, these days i use linpak (linx) to do stability testing for the cpu since its the fastest way to find out stability on the cpu. i do still use prime in blend mode to hit my memory controller and goldmemory to hit the ram. hopefully one day there will be a dos based stress tester that you just boot from and it tests everything for you.

but yes these days more and more people are getting clued up and are using linpak for stressing the cpu. i cant understand why a review site would want to review a cooler and not present the cooler a decent challenge. using dual core for the testing is bad enough as it is. prime on top of that and its just being silly now.

i just do a 24hour linpak to slap the crap out of my cpu for overclock testing. prime for cpu stressing is no good these days.

[SNiiPE_DoGG]

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.

[Baron_Davis]

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.

I like your style buddy...lol you remind me of myself

[Ender17]

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.

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