DIY Corsair H50 killer...?

[nitromullet]

I've been considering picking up a Corsair H50, but then I started thinking... "why not do watercooling right?" I mean, all the people who are in the know about water say that a decent DIY setup will destroy an H50.

So after looking over the sticky and establishing the criteria that I didn't want to have to drill or Dremel anything, here is the setup I'm thinking about:

Watercool HEATKILLER® CPU Rev3.0 1366 LC CPU Water Block - $58.70
Swiftech MCRES-MICRO™ Rev. 2 - $24.95
Swiftech MCP655 - $79.95
XSPC RX120 Extreme Performance Radiator - $55.45

Total: (w/out hoses/fittings/accessories/shipping) $219.05

...now to the questions...

Of course it should be pretty clear to anyone that the pump is overkill (and a little overpriced) for the intended purpose, but the Swiftech MCP655 looks like a decent investment should I decide I like water cooling and say want to cut up my case and install a 360 rad. Is this a good call, or should I go for a cheaper pump?

What type of performance should I realistically expect from this setup?

...Last but not least, does anyone else have any other recommendations on a relatively cheap H50 killer? I'm trying to stay away from components that wouldn't really be re-usable in a larger system. I looked at the Swiftech rads with the built in pump, but they have limited mounting options. They also only come in 240 and 360 sizes, and not 120. I'm also staying away from the Swiftech cpu blocks with the built in pump, as they would also probably be useless in a bigger setup.

Thanks!

 

[Red Squirrel]

I not sure if this would apply to cooling, but I've been researching underfloor heating and they say a pump will actually not work at all, you need a circulator, which I guess is similar to a pump but much slower flow rate. It might be different for water cooling though, but I think the logic is that if the water is flowing too fast it does not have time to absorb or let go of heat.

 

[nitromullet]

I not sure if this would apply to cooling, but I've been researching underfloor heating and they say a pump will actually not work at all, you need a circulator, which I guess is similar to a pump but much slower flow rate. It might be different for water cooling though, but I think the logic is that if the water is flowing too fast it does not have time to absorb or let go of heat.

The Swiftech MCP655 is sold as a PC water cooling pump, so I'm going to assume that it works.

 

[cubeless]

considering that it's costing 4x what i got my h50 for, i sure hope it rocks...

 

[Schmide]

$43 Cooler Master Aquagate DuoViva

$45 Swiftech MCR220-QP Res get 3/8 barbs.

$10-20 more than a H50 and it will destroy it.

Drain the kit and save the fluid.
Dump everything but the powered block, mounting, and tubing you need. (you can cool your GPU/Chipset with the extra block) Basically dump the flow indicator and aluminum radiator.
Stretch the tubing over the 3/8" barbs (it will fit), fill saved fluid, top off with distilled water.

Oh and you need fans.

Edit: if you want to cool your 920 and a video card, you'll need a Swiftech MCR 320 QP-Res - Triple Rad with integrated reservoir.

 

[nitromullet]

considering that it's costing 4x what i got my h50 for, i sure hope it rocks...

Well, yeah... This isn't really a stab at the H50. It is probably the best you can do for the money with no prior water cooling experience, and people seem to be pretty happy with. This was part of the reason I asked what sort of performance I could expect - because I honestly don't know.

I also realize I could do this cheaper...

XSPC Delta V3 CPU - Acetal - i7 and 775 Model $41.95
XSPC Dual Bay Reservoir Pump $44.95
Swiftech MCR120 Quiet Power Series 120mm Heat Exchanger - $36.95

total - $123.85.

Still more than the H50, but the H50 at $70-80 isn't exactly budget cooling either.

...with the dual fans and fan holders, I probably paid $100 for my TRUE. Although, all I had to do when I went from C2Q to i7 was pick up the bolt through upgrade kit. This is sort of what I'm thinking I'd like to do with a wc system - over build a bit so it has room to grow.

 

[nitromullet]

$43 Cooler Master Aquagate DuoViva

$45 Swiftech MCR220-QP Res get 3/8 barbs.

$10-20 more than a H50 and it will destroy it.

Drain the kit and save the fluid.
Dump everything but the powered block, mounting, and tubing you need. (you can cool your GPU/Chipset with the extra block) Basically dump the flow indicator and aluminum radiator.
Stretch the tubing over the 3/8" barbs (it will fit), fill saved fluid, top off with distilled water.

Oh and you need fans.

Edit: if you want to cool your 920 and a video card, you'll need a Swiftech MCR 320 QP-Res - Triple Rad with integrated reservoir.

This is interesting, but I think I'd want slightly more "plug-n-play" components for my first time water cooling excursion.

 

[aigomorla]

It might be different for water cooling though, but I think the logic is that if the water is flowing too fast it does not have time to absorb or let go of heat.

are you being serious...

*Shakes*

The reason why we say those kind of pumps would be better is because they ignore restrictions.

Meaning regardless of whatever restrictions, you would have a X flow rate.

However more flow = ALWAYS better.

More flow = More passes at the exchanger in the same given amount of time.
More passes = higher Exchange rate.

Higher exchange rate = less delta
Less delta = happier temps.

So no.. higher flow = always better

Water does not need a lot of time to pick up heat.
The more heat you pick up the hotter the water gets, and that would destory the point in water cooling.

Your using waters natural Specific heat, to pick up that heat, and have very little change on the water itself.

 

[WoodButcher]

I've been considering picking up a Corsair H50, but then I started thinking... "why not do watercooling right?" I mean, all the people who are in the know about water say that a decent DIY setup will destroy an H50.

So after looking over the sticky and establishing the criteria that I didn't want to have to drill or Dremel anything, here is the setup I'm thinking about:

Watercool HEATKILLER® CPU Rev3.0 1366 LC CPU Water Block - $58.70
Swiftech MCRES-MICRO™ Rev. 2 - $24.95
Swiftech MCP655 - $79.95
XSPC RX120 Extreme Performance Radiator - $55.45

Total: (w/out hoses/fittings/accessories/shipping) $219.05

...now to the questions...

Of course it should be pretty clear to anyone that the pump is overkill (and a little overpriced) for the intended purpose, but the Swiftech MCP655 looks like a decent investment should I decide I like water cooling and say want to cut up my case and install a 360 rad. Is this a good call, or should I go for a cheaper pump?

What type of performance should I realistically expect from this setup?

...Last but not least, does anyone else have any other recommendations on a relatively cheap H50 killer? I'm trying to stay away from components that wouldn't really be re-usable in a larger system. I looked at the Swiftech rads with the built in pump, but they have limited mounting options. They also only come in 240 and 360 sizes, and not 120. I'm also staying away from the Swiftech cpu blocks with the built in pump, as they would also probably be useless in a bigger setup.

Thanks!

This is a great start, all these parts will integrate into a new loop when you upgrade. You have done your homework.
Will this loop "kick-ass"???? .
not mine,,,,,,
but it will out perform the H50 though and will upgrade well. Those kits that have been posted have no potential unless your ready to make serious mods to each and every part. Stick to your first post choices. If at all posible get the bigger rads but keep in mind that the XSPC 120 you chose will always be a good add on to any loop if you need more cooling so this is not a bad start.
good luck!

 

[Kenmitch]

I've been considering picking up a Corsair H50, but then I started thinking... "why not do watercooling right?" I mean, all the people who are in the know about water say that a decent DIY setup will destroy an H50.

That is a true statement....But if you catch it on sale it's a good bang for the buck item. Best Buy seems to be putting them on sale quite a bit lately for $59.99 and at that price it's kinda a no brainer....Unless you like to dabble in the real water cooling arena

I tested one of these on my e5200 @4.16ghz and I was impressed by the results. Using the stock compound and included fan the H50 maintaned temps in the mid to upper 50's during Intel Burn Test and OCCT with the fan blowing in as directed. I switched to fan blowing out and it only made a couple degree diff when testing. I'd imagine that with a good quiet high flow fan or 2 and arctic silver the thing would be way better.

Anybody who has a e5200 overclocked to 4ghz or more knows it's hard to tame the little beast when it comes to heat.

Anyways going the real water cooling route would be more fun. You just gotta plan very well as sometimes things don't work out as planned the first time or even second time. That and order some xtra hose just in case your plan doesn't go as planned.

 

[GAM3RIG]

the Modded H50 killing the real wc loop! You can customize or mod it for 100 buck or less than you spend 219 on the WC Loop. :biggrin:

this is not my modded H50.. this is for example

Isnt this the MODDED H50 the WC LOOP KILLER!?

 

[aigomorla]

Isnt this the MODDED H50 the WC LOOP KILLER!?

no.. :\

Would kill that setup even tho its outdated.

 

[GAM3RIG]

no.. :\

Would kill that setup even tho its outdated.

Ha this is MINEEEE! ()

anyway I dont think this box is outdated. I did look at waterblock, pump made by 5/09. :\ *Mmm* but those looks classic loop than saying *outdated*. :biggrin:

 

[aigomorla]

good luck finding one available.

LOL... everyone i can think of ran out to grab one.

The D5 alone is priced higher then that kit on clearance.

 

[nitromullet]

good luck finding one available.

LOL... everyone i can think of ran out to grab one.

The D5 alone is priced higher then that kit on clearance.

My local Fry's seems to have that in stock - listed as "includes display item". I might have to see what that's all about.

Having never done any water cooling before I don't have anything, so any fittings, accessories, etc would be helpful as well.

 

[aigomorla]

My local Fry's seems to have that in stock - listed as "includes display item". I might have to see what that's all about.

Having never done any water cooling before I don't have anything, so any fittings, accessories, etc would be helpful as well.

that will have everything you need.

You can add extra's like a res, and probably get better barbs.

But basically you should be able to setup a fully working water setup with that 1 box alone.

 

[nitromullet]

that will have everything you need.

You can add extra's like a res, and probably get better barbs.

But basically you should be able to setup a fully working water setup with that 1 box alone.

Sadly, the frys.com was mistaken. My local Fry's didn't have one. I think that would have been a really decent kit to start with. Oh well...

 

[biostud]

Personally I'm switching from water to air in my next build. It has been fun to have a watercooling setup, but to me it's to much of a hazzle to build new one when I upgrade. To me to modern aircoolers are sufficient, and the extra power/heat produced from high overclocks are not worth it (not to mention the price/performance ratio), unless you just want to push your system to the edge, which I totally understand.

 

[Billb2]

More flow = More passes at the exchanger in the same given amount of time.
More passes = higher Exchange rate.

Well yes, more passes. But no matter what the flow rate is, the water spends the same amount of time in the radiator giving up heat.

Race car analogy:
One mile track, two lines painted on the track 88 feet apart (60 mph = 88 ft/sec.)
One car goes around the track at 60 mph, another at 120 mph. In an hour, both car spend one munute between the lines.

The reason higher flow rates work better is because theer is more turbulance and less laminar flow inside the radiator (or waterblock, for that matter).

To the OP:
That Corsair thingey is about as good as a good air cooler.
These watercooling websites might interest you.
http://www.skinneelabs.com/
http://martin.skinneelabs.com/

 

[nitromullet]

Well yes, more passes. But no matter what the flow rate is, the water spends the same amount of time in the radiator giving up heat.

Race car analogy:
One mile track, two lines painted on the track 88 feet apart (60 mph = 88 ft/sec.)
One car goes around the track at 60 mph, another at 120 mph. In an hour, both car spend one munute between the lines.

The reason higher flow rates work better is because theer is more turbulance and less laminar flow inside the radiator (or waterblock, for that matter).

To the OP:
That Corsair thingey is about as good as a good air cooler.
These watercooling websites might interest you.
http://www.skinneelabs.com/
http://martin.skinneelabs.com/

but doesn't the car also cover more distance in between the lines?

10560ft @120mph vs 5280ft @60mph...

it seems to me that doubling the flow rate has loosely the same effect as using a rad with double the surface area. ...or am I missing something?

edit: oh yeah, thx for the links. I have seen a lot of references to martinsliquidlab.com, which is parked now... I did not realize he is now at martin.skinneelabs.com.

 

[aigomorla]

Well yes, more passes. But no matter what the flow rate is, the water spends the same amount of time in the radiator giving up heat.

Race car analogy:
One mile track, two lines painted on the track 88 feet apart (60 mph = 88 ft/sec.)
One car goes around the track at 60 mph, another at 120 mph. In an hour, both car spend one munute between the lines.

The reason higher flow rates work better is because theer is more turbulance and less laminar flow inside the radiator (or waterblock, for that matter).

bill its physics..

the 120mph car will go arround track twice while the 60mph car will go around once.

Well pretend each time they go around they drop off a passenger.

That passenger is HEAT.

So the 120mph car dropped off more heat in the same amount of time.

Now the limitations are usually on the waterblock side, in how much heat it can take away from the cpu and dump into the water.

Thats where your turbulence is more played.

But in a straight exchanger aspect, more passes at the rad means lowers temps period.

 

[skinnee]

Its not a linear relationship though. C/W does change with flow rate, but the scale is not linear.

edit: I test radiators @ 1.5GPM since that is the point of diminishing returns. However, I think its time I hike up my boot straps and start doing variable flow rate testing on radiators... man that is going to take even longer.

 

[aigomorla]

Its not a linear relationship though. C/W does change with flow rate, but the scale is not linear.

Yes... because of this comment...

Now the limitations are usually on the waterblock side, in how much heat it can take away from the cpu and dump into the water.
.

And vapor showed us very nicely on why...

Roughly physics states it takes 350W of heat @ water moving 1.5gpm for it to go up 1C.
Having lower flow, would make that heat number go down, having more flow would make that number go up.

 

[skinnee]

to stay with the Car analogies...

You can have the best engine in the world, but if you have a crap transmission all that power is wasted.

 

[Rubycon]

Once you reach the optimal dwell you must decrease ingress water temperature in order to have augmented thermal transfer from your target.

This can be accomplished with a larger radiator, faster fans, or (more properly) a chiller.

Very low temperatures - more than 150 degrees below zero centigrade - are possible with a properly engineered solution of TE cooling and supply of chilled water.