226w Peltier chilled with a Black ICE Extreme is possible???

[atomicminds]

A system with an MCW462-UHT Waterblock or Maze 4 by Danger Den powered with an 226W TEC can be chilled using only a Black ICE Extreme Radiator?

Is safe to put on the same loop an MCW50-T on VGA too, using this choice of radiator?
Any recomendations on a specific WaterPump or another radiator for this setup?

In this setup will be used an Athlon XP processor and a ATI 9800 Pro as VGA...

Thanks in advance for the help !!!

[Belgian_Dude]

I dont think that radiator will be powerfull enough to chill your water with the Peltier :s Certainly not with that VPU block on top of it. If I were you, I'd go for a bigmomma Heatercore ... ;) Or two :evil:

[atomicminds]

Thanks for the tip Belgian_dude !!!

I´ve asked because i got some problems with space surronding my CPU tower right now...

I think that i will mod a little my Xaser tower and put a Black Ice Xtreme II on the top of it and stick 4 120mm fans on it, like 2 inside of the tower pulling air to the rad and 2 on the top of the rad outside of the tower...

I can´t go more Xtreme than that... ;)

[Hardass]

The Swiftech Quiet Powers I run use the 676 rad found here to cool a 226W and 80W in same loop. http://www.xtwerkz.com/liquid_cooling_parts.htm

[atomicminds]

Originally posted by Hardass
The Swiftech Quiet Powers I run use the 676 rad found here to cool a 226W and 80W in same loop. http://www.xtwerkz.com/liquid_cooling_parts.htm

This a great solution and it isn´t too expensive like Black Ice Extreme II, you have any pictures of your setup ?

Thanks !!! :D

[mdzcpa]

I also have a set up similar to Hardass. It's the Quiet Power plus TECs.

A pair of high cfm 120mm fans keep the 676 rad cool enough to handle the 226w CPU and 80w GPU TECs just fine.

Here's a few older pics:

http://mywebpages.comcast.net/mdzcpa/images/QP1.jpg

http://mywebpages.comcast.net/mdzcpa/images/IC7incase.JPG

http://www.liquidninjas.com/mdzcpa/reviewphotos/TECupgrade/Picture_052.jpg

http://www.liquidninjas.com/mdzcpa/reviewphotos/TECupgrade/Picture_092.jpg

[Tedinde]

Get a Car heater core, The one in this case i have down in the basement, has a Heatercore out of a 79 ford pickup with AC, 6x6 inches 4 core and 120mm fan in the front. I cut the front out so is through the sheet metal of the case and is out of the way, Nice shroud on it and a 5/7/12v switch.

Sold it last night. 2.8b @ 3.3ghz, Used to handle a 226 watt tec on it @ 3740mhz, but since i was selling it to someone, He wasnt gettting the PS, and didnt have a clue what a tec was anyway,

He was more thrilled about the neon fan!!! hehe

Hoses are going up to a CDROM resovour.

http://mywebpages.comcast.net/tedinde/photogallery/XFpics/IM000759small.jpg

[zippyc]

If you are going to add peltiers to your water cooling system, I should tell you the first lesson I learned is that they generate way more heat then expected. So you are going to need at least one Very Good heater core and a lot of CFM's (150+) forced through it.

IMO the radiator solutions you see marketed out there for overclockers are attractive AND expensive (and usually painted heatercores you can source yourself, not so ideal for best very best heat transfer. It’s not bad to have paint, it’s just not best.)

I also recommend a heater core for your situation (actually I'd use two if going with more than one pelt). I used GM 9078, which were $23.75 each at the local Van's Autoparts. They measure 5.5" x10.5" x 2", so they have a good amount of radiating space. There are some out there that measure 8x10x2 that work well too. Any auto parts store should have a catalog of cores you can look through, in hundreds of shapes and sizes. Bigger is better, and the 2” thick heater cores work better than 1” radiators any day, given all other dimensions are the same.

I also got this size so I had the flexibility to put one in my standard tower case (I think the will fit in my my Antec 1040B, removing the drive bay or putting the core on top), although I eventually chose to build external enclosures to house them all. I have over thee pelts (400W+) I'm cooling with four cores now (I'm adding another 400W over the next week or so.)

Keep in mind that your pelt cooling is only as good as how cold you can keep your fluids on the hot side of the pelt. Drop your hot water 5C, you'll get 4-5C reduction in your cold side of pelt. temps.

And make sure you seal a shroud (no air leaks allowed) to whatever radiator and fans you choose...shinny silver HVAC tape found in the insulation section at Home Depot or Lowes (not "duct tape") is very useful in this effort, as it's surface radiates heat better than other tapes.

One last piece of advice...either add both pelts at once or consider going with two separate cooling loops. That 226W pelt in the same loop as a standard GPU block without a pelt will raise you GPU temps by 10 ~20 degrees or so, because your entire loop’s temps will go way up.

Here is a picture of the inside of my wooden boxes…. The two fans (110V so no power supply needed!112CFM x 2) suck through one radiator from the bottom and exhaust out the top through the second radiator. The water travels in series through the top then bottom…

[Illicit Tweakin]

Just found out about you guys. Finally a place that looks cool. And no stubberon mean people.

The Black Ice X2 works good for me with the swifty 226 and 80 watt. I use 3x80mm vantec tornado(90cfm ea). My buddy uses 2X120mm delta (190cfm ea), runs with 1 for less noise with no problem. Eheim 1250 (320 gph) for the both of us. I can run windows and most of 3dmark at 2687mhz 215x12.5 board limitations (I think) (old a7v333) and rock solid 2475mhz 198x12.5

[zippyc]

The boxes are really ugly, highly effective, temps no higher than 38C on the hot loop with over 430W (plus the CPU and GPU heat) to dissapate.

Dual loop setup with cold water loop running parallel to my Whitewater CPU block and Gemini GPU block.

My cold loop stays at 18C or less, even after hours of heaviest load (In my setup nonstop gaming is more of a test for heat than Prime or Madonions). At Idle it hangs around 12c, and with no load I get to 2C.

[jinu117]

BIX is cutting it. I've used it before with just BIX for mcw-5000pt and mcw-50t and that was a mistake. Heatercore will net you about 6-8c better temperature. Now... dual heatercore would be better though :P

[Weapon]

a heatcore out of a '71 caprice would handle it.

[Mr Footlong]

Hmmmm, perfect topic for me to ask my question :)

I am planning on using a Swiftech MCW50-T peltier (80 watts) GPU cooler with an Eheim 1048 pump, purple Ice wetter, Dangerden Reservoir and a Blackice extreme Radiator with only an Antec 120MM 70.6cfm fan due to space restrictions and noise restrictions. Only the gpu will be in the cooling loop and it is a 9800 Pro.

The rad is being mounted into the top of a Black PC65 case but inside the case.

Will I get away with it using that fan, or am I gonna have to go noisy?

Cheers guys,

Nick :)

[zippyc]

Sure you could get away with it, the warm loop would be pretty warm though, which means your GPU temps will be appx. that many degrees hotter.

A second fan of the same specs stuck side by side with the first (and shrouded to the radiator) will get you much better temps by doubling your airflow and would not increase the noise very much.

Good luck.

[Mr Footlong]

Cheers for that, I had thought about that too, but the space is gonna be too tight I think. Thanks for that :)

Cheers,

Nick :)

[2fast4u]

I have an Xaser III case with the Black Ice Ext with a 105cfm 120mm. I drilled out the hdd cages and put a high performance fan in the front fan bays to help pull the air out.Also turned back case fans to suck in as prommie takes care of cpu and psu is dual fan. Probably would be more efficient outside but I take this beast to lans and wanted to keep everything inside the case. 9800 and nb run around 15*C under load with a MCW50-T on each. Meanwell psu is also inside on mb tray. Just make sure you have lots of cool air comming in! I don't have a digital cam at the moment but am gatting one this weekend and could send you some pics if you like.

[Shroomer]

if I am reading this right a MCW50-T with a 176 W TEC (24v) at 12v should be fine on a procore by itself.

[Jabo]

On general note:

Radiators are as good as big is their heat dissipation area and how big iar they recieve.
Restrictivenss of internal design has bearing on pump choice (the more restrictive the stronger pump is going to be needed).
Therefore I think that any rad will do as long as you are able to provide sufficient air flow and coolant flow to achieve your target results:).
Don't be fooled into thinking that it's doable thou :)
The CFMs and GPH needed in some cases are so high that it just doesn't make any sense to proceed.

The bigger rad the less CFM you need the less noise you are going to suffer :)

[Epsilon]

As lol as u run not more than 1x226w peltier

a BI extreme is enough :)
expect a watertemp of Ambient + 10C

Fan doesn't make MUCH difference btw. A good radiator is more important.

[Jabo]

Originally posted by Epsilon
As lol as u run not more than 1x226w peltier

a BI extreme is enough :)
expect a watertemp of Ambient + 10C

Fan doesn't make MUCH difference btw. A good radiator is more important.
Again, it all depends on setup. If you have restrictive rad nad weak pump no matter how big rads you install othing will really improve since blocks will be starved for coolant. Having good rad in your loop allows you to increase performance of your system by increasing air flow through it (provided you got appropriate pump.
It's never like 'Good radiator is all' or 'Top notch blocks rulez', not true at all, system has to be balanced and h'ware sized for each other.

Fans in general make MASSIVE difference unless you have joke-of-a-pump in your system.
Go and have a ead trhough BillA's rad test results at overclockers.com and you'll see. They are flawed (he used pitiful pumps) but are of help since he's extremely comprehensive in his methods and resulta presentation.

[mdzcpa]

Originally posted by Jabo
It's never like 'Good radiator is all' or 'Top notch blocks rulez', not true at all, system has to be balanced and h'ware sized for each other.

So very very true. And many users miss this most important point.

Fans in general make MASSIVE difference

Again very true. Even more so from a practical point of view. Sure, users could go without fans if they wanted to put a massive radiator outside the case (killing asthetics and system portability). Those who want to go that route are few and far between. For most folks who want a cleanly modded case with a self contained H2O system, fans are a must.

[Epsilon]

Some1 who says a pump of more than 200l/hour (effective) is needed, hasn't learned phisics at school :).

200l/hour (0.5gph) = 55ml/s

water needs 4.18KJ to warm 1 liter of water up 1Kelvin.

To warm 0.055 l up --> you need 230J.

A Normal peltier load, is about double that value (460w). Every second the water is in the block, the water will warm up 2 degrees.

How long is the water in the block? well, that depends on the inside volume of the block. Lets say the block is 27.5cm^3 on the inside (guess why i am chosing that size ;)). And that is a LARGE block indeed !!

The water warms up 1 degree between entering the block, and leaving the block.

The one who tells me to buy a better pump, is a complete moron :)(well, only if has read this but then still tells me to ;)).

and offcourse, the system has to be ballanced. But a good radiator is more important then some Jet fans :).

[zippyc]

You are correct in the assumption that even a low flowrate in a ok waterblock is sufficient/ better than air cooling.

However, if you gave an me XYZ 200 l/hour pump and expect it to overcome the greater restrictions of several waterblocks in series and the multiple radiators of a complex water cooling rig assuming the system will cool as well as if you used a larger pump with 3x the head rating, you're in for a rude awakening.

The xtreme community is interested in all solutions, and I'm not saying a low l/h rated pump wont work well in a simple rig...it can.

However there are those of us looking for a few degrees cooler here, a coulple of mhz there, etc.

And is cases like this a larger pump (more l/h and higher head rating) can make a large difference in your systems overall performance.

Or (in other words) if you want to simply have a more efficient system, you may need a larger pump.

And because a pump is rated for 200l/h "UNRESTRICTED", once you add restrictions like waterblocks and radiators the system flow becomes way less than 200 l/h.

[Jabo]

Originally posted by Epsilon
Some1 who says a pump of more than 200l/hour (effective) is needed, hasn't learned phisics at school :).


Back to school then.... :)




200l/hour (0.5gph) = 55ml/s

water needs 4.18KJ to warm 1 liter of water up 1Kelvin.

To warm 0.055 l up --> you need 230J.



Blooming h311, I didn't know that gallon equals 400l!!! WOW, that's some major surpirise in my education ;)
What you are writing about is pure academic since it is assuming 'no time' environment and equal acces to energy source by each and every molecule of your one liter of water. What you wrote is right but has no direct application to PC water cooling. Laminar/turbulent flow and wetting of block/rad surfaces, thermal energy density, temp gradient, thermal energy transfer time window lenght (lenght of time each molecule stays in contact with heat radiating blocks surface) - which directly links us to pump's power btw - all of this is way more important than constant physical properties of water m8:)




A Normal peltier load, is about double that value (460w). Every second the water is in the block, the water will warm up 2 degrees.

Are you able to back it up with at least single bit of maths?
Where did you get that TEC load from? What TEC, running what volts, cooling what die @ what vcore?
And how on earth did you come to conclusion that it takes 4.18kj over one second to increase water temperature by 1K???
By simple relation that if your imaginary heat load is twice as big as it takes 1l of water to raise it's temp by 1K you cannot just conclude that temp will raise by 2 degs in 1 sec. Why not 1 hour?
Waky, waky :) Enery numbers you quoted tell you how much of Energy is needed by all the water molecules contained in 1l of water to raise temp of this 1l of water by 1K, nothing else. Period

Calcs you want to exercise here are done for sizeing heat exchanger's side of cooling loop. Water will take as much thermal energy as block would give. This in turn is dependant on the variables I mentioned above



How long is the water in the block? well, that depends on the inside volume of the block. Lets say the block is 27.5cm^3 on the inside (guess why i am chosing that size ;)). And that is a LARGE block indeed !!

The water warms up 1 degree between entering the block, and leaving the block.

The one who tells me to buy a better pump, is a complete moron :)(well, only if has read this but then still tells me to ;)).



I must be a complete moron then...
I think I'd better go upstairs to get my bath tub since the bigger block's internal volume the better...
Goodness me, finally I am enlightened! What a relieve!!!





and offcourse, the system has to be ballanced. But a good radiator is more important then some Jet fans :).

You don't get it do you?
You may have the best radiator on earth made of daimond with square miles of surface area but if your block is crap and your pump even worse you may shove it pal! ;) :) Will do your system as much good as having a radiator in a form of 20meters of garden hose. Actually with crappy pump overall system's flow would be so pathetic that water would have sufficient time to cool even in garden hose LOL:rolleyes:
If you have great rad you may use low speed low noise fans and have the same results as crappy rad with jet engine strapped to it :)

Do some reading first boi before you start 'moroning' people here at Xtreme.
You got different view just express it and discuss differences.
There's alwyas somebody more experienced and with better knowledge in a field you deem yourself an expert at.
I learn here every day and I seriously enjoy it!

[zippyc]

Yeah, I was going to go down the road of detailing the errors in Episilon's theory, but decided to keep it in simple terms everyone could understand. Hell, I make errors all the time, no big deal to me. Just don't want first timers misled.

Epsilon, IF you decide to implement water-cooling of your own, you will find your textbook knowledge will only answer some of your questions as you will most likely encounter results different from what you'd might expect, then it could be time to come here and seek advice. ( I suspect that is not in your nature from your posting though)....

I doubt that you intended to alienate anyone here, perhaps a less bold statement may have sparked a different reaction.

I did review you post again and you qualified the pump with an "effective" flow rate of 200, so which pump provides that level of flow?

THAT depends entirely on your hose lengths, diameters, radiator restrictions, blocks, fluid choice, almost everything.

Very few folks on this forum have a flowmeter that can accurately determine flowrates anyway, so your statement, although somewhat accurate, does not give those without flowmeters much direction except for a theoretical target. So lets all be happy and move on, no flames nescessary...and no calling people morons please. That makes people :argue:

[Jabo]

Amen Bro! :)

[Epsilon]

First off all, I never intended to insult any1, Sorry for that if it sounded that way (it's hard to interpretate the impact of some text as a foreigner :( )

Originally posted by Jabo
Blooming h311, I didn't know that gallon equals 400l!!! WOW, that's some major surpirise in my education ;)
What you are writing about is pure academic since it is assuming 'no time' environment and equal acces to energy source by each and every molecule of your one liter of water. What you wrote is right but has no direct application to PC water cooling. Laminar/turbulent flow and wetting of block/rad surfaces, thermal energy density, temp gradient, thermal energy transfer time window lenght (lenght of time each molecule stays in contact with heat radiating blocks surface) - which directly links us to pump's power btw - all of this is way more important than constant physical properties of water m8:)

That of 0.5Gph was a typo, sorry ;)
But you are correct here, and clearly have made a point.



[Are you able to back it up with at least single bit of maths?
Where did you get that TEC load from? What TEC, running what volts, cooling what die @ what vcore?

Well, i intended to keep it simple just by giving a fixed value. But at startup, a 226w peltier will give u that heatload (226w + power consumption)



And how on earth did you come to conclusion that it takes 4.18kj over one second to increase water temperature by 1K???

Not in 1 second. It just takes 4.18kJ to warm up 1L of water 1K
Watt is J/s



By simple relation that if your imaginary heat load is twice as big as it takes 1l of water to raise it's temp by 1K you cannot just conclude that temp will raise by 2 degs in 1 sec. Why not 1 hour?
Waky, waky :) Enery numbers you quoted tell you how much of Energy is needed by all the water molecules contained in 1l of water to raise temp of this 1l of water by 1K, nothing else. Period

In this part, i think u did not understand what I was trying to tell.
Btw, the time part simply disolves when u are talking W and l/s.
So that actually was correct.



Water will take as much thermal energy as block would give.
Well, If u put 400W in a block. The 400W will come out (even without ANY surface). That's a law. The big thing is, the temperature will rise and rise untill it reaches a high enough temperature to exchange the heat (the famous K/W value)


I must be a complete moron then...
I think I'd better go upstairs to get my bath tub since the bigger block's internal volume the better...
Goodness me, finally I am enlightened! What a relieve!!!

Sorry if i insulted u, it was DEFINATELY not on purpose. I was just trying to explain something.

The internal volume of the block OFFCOURSE doens't matter the performance. It was just an example. Now i think of it, that doesn't make sens. U are complete right. I was trying to calculate the temperature difference between the water that was going in and the water that was going out.



You don't get it do you?
You may have the best radiator on earth made of daimond with square miles of surface area but if your block is crap and your pump even worse you may shove it pal! ;) :)

True, did i mention that it wasn't? I said a system needs to be ballanced. A pump that does 200L/hour (effective !!!, so a 600L/hour eheim or something !) is enough. More = better, ofcourse, but the difference isn't a much as u thing. 2C isn't worth to buy a better pump, Sorry.



If you have great rad you may use low speed low noise fans and have the same results as crappy rad with jet engine strapped to it :)

You would be suprised. At SOME point the fans just can't handle it anymore.

190CFM 120mm delta fans (talking about jets ;)) and a Blackice extreme doesn't perform as good as a Blackice Extreme2 with a 120CFM papst fan or something. By the way, 2x 60CFM papst doesn't make as much noise as a 190CFM delta.



Do some reading first boi before you start 'moroning' people here at Xtreme. You got different view just express it and discuss differences.
There's alwyas somebody more experienced and with better knowledge in a field you deem yourself an expert at.
I learn here every day and I seriously enjoy it!
I do to and i was just trying to express my view on it.
And offcourse, there are allways experts ;).

[Epsilon]

Originally posted by zippyc
Epsilon, IF you decide to implement water-cooling of your own, you will find your textbook knowledge will only answer some of your questions as you will most likely encounter results different from what you'd might expect, then it could be time to come here and seek advice. ( I suspect that is not in your nature from your posting though)....
Well, it wasn't in my nature to do that, sorry again.

About implementing watercooling, i allready have a nice system. And even with some peltiers.

It's true that you can't callculate the exact temperature of your system. There are allways elements that create better or worse results (turbulant water flow, contact area, heatspreading, etc etc).

Flow rate is indeed just a wild guess. I think, 200L/hour is doable with a eheim 1250 or hydor L30. If u have a radiator with a low flow restriction, a eheim 1048 (150GPH) probably will be enough. Especially if u have a low altitude difference in your setup.

When u put the pump on the ground for example, and the blocks on the table, the flow will be much less than when you put all the elements on the table.

[Jabo]

Epsilon, sorry for getting so hard on you, got a bad day y'day and it looks like I just blew my steam on ya, apologies m8 *me feels like a :banana::banana::banana::banana:*
Will get back to you on meritorical (spelling?, I am not english neither ;) ) points later tonight, massively bussy in the office and my boss pisses mo off constantly!!!

[Epsilon]

massively bussy in the office and my boss pisses mo off constantly!!!

np m8 ;)

Southpark "cartman"s way to get rid of that -->
Just smack him in the face and say: "THATS ENOUGH" ;)

[Jabo]

He he he, don't tempt me here ;)
He'd end up at Intensive Care Ward (I am rather large bruiser)

I'll just go home at normal time and leave him to clean his own $hit.

Back to the chase.
As I see it what we want is to measure temperatures at inlet and then ot outlet of w/block. Do it to hundreth of a degree and that's all you need, really.
As far as theory goes I think time does not reduce itself there. I think we need to calculate Watts per lier of water pumped which will give you W/liter = J/liter*s. This tells us about physical properties of our block i.e. how much thermal energy a liter of water pumped through water block recieves. Now we have to introduce time since the amount of energy sucked by water is heavily dependant on flow rate produced by our pump. This gives us (W/liter)/s = J/liter*s^2.
Lets write it down like this ...
run out of time, been writting this bit for llast 2 hours, will finish later tonight :)
I will try to produce an easy to use equation for pump sizing:)

[kayl]

i had a 130w tec then upgraded to a 226w tec
i was using a black ice extreme rad, i had very little gain temps wise.
i then added a heatercore to the system, now there is a difference, im still not happy though now i think i need another pump

[Jabo]

Small update, I am working on this equations, they will be more like rad sizing help than pump since it's a different story although connected very closely.

Btw. I've installed domestic central heating circulating pump into my brand new coolin system (detailed account to follow after UK Patent is granted). It does not requie any fans to run (the only fan is Enermax'es PSU fan and as soon as I can I'll convert it into water cooled PSU to create first trully fanless multi TEC system

Full load temps (F@H) are around 5-7C higher than 300CFM heater core cooled system. At idle I can still stay below zero until ambient temp rises too much.

Back to pump business. It's Wilo Gold 50 producing 3m^3/h at 1m head, at 5m head it pumps 0.5m^3/h. Depending on pump brand and model it's anything from 20 to 4 times more water pumped for £36....
It has one problem admittedly... it produces quite a lot of heat at max flow setting (it has three settings)...bummer.... I solved it by installing quiet Zalman 92mm fan @5v under it blowing upwards.

Strangely enough the biggest temp increase I got from my North Bridge after I removed fan cooling heatsink. GPU temps rose massively and CPU was also higher than it should have been. Had to install another Ys-Tech to cool it. Have to pelt NB too to get aroud it :)

[kayl]

have you eva thought of making a bong
i nearly made one but i think the humidity here is too high

[Zejtan]

It's not only the size of the radiator that makes a count.
The amount of air that goes thrue the radiator is atleast
as important as the size. more air, more cooling.

[Epsilon]

True zejtan

But airflow stops at some point, but radiator sized doesn't :)

And besides, i'd rather have a BI2 extreme with 1x papst 95CFM than a BI extreme with 1x delta 120HP ;)

cooling is the same but the noise...

[kayl]

with a good bong you can actually go below ambient temperatures

[Epsilon]

True, but it isn't a closed system and with that kind of heatload (say 400W for a 172w pelt and 500w for a 226w pelt) u need to vaporize a lot of water to cool it down.

Then u have to keep filling the bong constantly