Cooling Est. size of a passive radi for approx 250 Watt 24/7

Hi all.
I'm new into this wartercooling, and I'm serching for knowledge.
I need alittle help from experienced people, therfore I choosed this forum. Thanku for a great forum btw.

Anyways. I'm planning to make myself a 25U Rack with my gamercomputer in a 4U, a fileserver in another 4U and 6*250gig HDD's in a 2U. All must be watercooled. Since I hate noice i wanna go water. I have trashed way to many HDD's due to overheat, so i want the watertemp. on a temp like max 30 degrees C* My room temp is at like 22 or soo, normal temperature.
I was planning to put a cu plate on the back of this huge case, the plate will be almost 1 m2 big. the plan was to solder cu. pipes on the back of the plate, and connekt it to the backside of the rack.
Below is a est. guess on what will be on water.
2*CPU's
1*GPU
2*Chipset
8*HDD's
2*PSU
The gaming computer will only be turned on when I'm not sleeping. aprox 10-15 hours per day.
the fileserver will be online 24/7.

I would love if it could be passive, but if needed i can put in a radi in the top, and put 4*120mm coolers on 7V.


I have found a pump for the projekt, It's a 230 Volt, 20-60 Watt.. it's a bit special.
It's a Grundfoss UPE 25-40 with a VLT. It has a buttompanel, where yo can change the frekensy up and down, It is very powerfull, maby even to powerfull, but there is a way to work around that.

What do you think guys? Is it completly insane??
And No, I dont do LAN-parties
--- Editet ---
I just found this.
http://www.vcore.dk/shop/default.asp?ProductID=2210
It's a radi that is so big that you can place 9*120mm on it.
If the passive radi on the back isent enough, I could put this one in, set some automatic cooling on it so if the temperature reaches 30 C*, it must run for 1 hour at 7 v.
Not that pricy to make, since I'm a elektrisian, and can get alot of the components cheap.

 

Got a room temp of 22? Expect it to be nearing 40 or more once you're pumping about 400W of heat into the air. You're best off to do things with big, quiet fans and noise damping, so that the rad doesn't overload and turn into a super heater...

 

thats nothing mine can fit 15 *120 mm .. aqua computer 1800 rad look into it

 

well first of all there already are passive or silent psu's and there aren't standard ways to cool those. It has been done, but I think the passive or silent solutions are much more eloquent and effective. Also don't expect a pump like that to be silent. Also I believe water cooling hdd's are just as bad if not worse than still air. (I'll explain if requested.)
Anyways I'm not a master water-cooler. but I believe here the issue would more be holding a large reservoir of maybe 10+ Liters. That coupled with a large car radiator as you showed would probably be able to run passively. Also 30C is very low and pretty unreasonable. My room is at 18C and with the water at air temperature I get about 28C after its been running full blast maybe 32. Is there any reason you want 30C? planning on over clocking? At any rate the more water and bigger the rad the smaller the delta between room temp and water temp.

oh also a pump that big is probably not going to accept anything smaller than 5/8" I.D. you're going to need to creat a high flow system or step it down as that sounds like a good way to rocket a connector off, running at high PSI.

 

I will start with installing 1,5 TB into the projekt.
I'm a old server geek, have been moderator on a clusternetwork that had alot of space, I wanna do the same, just with nada noice.

I know the flowrate must be high or I will pop connektions. I can take a Y splitter after the pump, an lead some of the water direktly back into the resovair. I dont know the english word for this little devise I have in mind, but you can move the resistance up and down with it, so I can deside how much pressure/flow i want throw my system.

The reason I would like to hold down the sice of the resovair is that if I get a leake, I will pump the water direktly out on my untreatet wooden floor.. my floor will suck it right down, and the floor would have a large spot with blue/black that couldent come off.
The other thing why not, is that the resovair will be placed inside the case, and will most likely be made of stainless steal, since I got that by hand "it's free".
Please do explain why you think that watercooling is bad??

What about this then..
Place 2 of those rads in the top, place them on top of eachother, and have 9*120 mm's coolers running at 7 volt. pulling in colder air from the bottom, and ejekt it in the top.
Or.
2/3*aqua computer 1800 "the one that can have 15*120mm fans on it. Strap them to the backside of the case, without any fans.

i hate this guessing I dont wanna do overkill*10 but I dont wanna do watercooling and end up with it overheating and sounding like a tonado.

throw me some info please
Much apreciatet for the feedback!!

 

If you're planning something that big, it might almost work using an old heating radiator (the ones to keep your house warm) on the back. At a guess that should work...maybe.

On a more sensible note: Standard watercooling radiators are built to be used with fans, so they are designed with the fins horizantal, but for effective passive cooling, the fins work better aligned vertically. You can always turn a normal rad round but it will take up lots of space.
Instead you could go for a rad thats designed to be used passively, like the Cape Cora or the Zalman Reserator. You'd probably need more than one reserator (but they do stack on top of one another), and I'm not sure about the cooling capacity of the cape cora, but they'd do the job pretty well. The zalman would have the effect of being your resevoir as well.
There's still be a place for some temperature controled fans as well.

 

Yeah. i have been thinking of those...
I think I'll come to the conclution that i will use 2 large rads.. the ones with either 9 or 15*120mm coolers. mount the coolers and mod them down to a lower voltage.
Bring up air from the bottom and spit it out of the top. If it needs more cooling than this, I must place some zalman reso's on the back... i will be ripped off (broke)

Oooh yeah, there was one that says that I shuldent expect that this pump was noiceless..
I'm a elektrisian, and I work with thise kind of pumps daily. They are ultra silent.
sometimes when they are running, I haft to tjek with putting a screwdriver to the rotor of it, and hear the vibes it gives to the scewdriver. I'll go as far to say that they are as "laud" as a PABST cooler.
It's not a mistake I choosed this pump, allthow I did know that it might was to powerfull. Only time will tell.
Thankyou for nice intel.

 

I'm requesting.
Please.

 

Me too, i can agree with you if it's moving air but still air ?

 

on most or all of the hdd watercoolers I've seen they are generally just pipes running along the outside edges. Here's the thing running some really effecient watercoolers would just create 2 "cold spots" along the perimeter, which isn't even where the heat is coming from. this uneven cooling in my opinion would be more likely to induce read/write errors due to the slight epansion that would take place in the center of the driver (where the heat is made IIRC) and the lack of expansion in the edges.
At any rate my hdd's are content running still air cooling (nothing) and are stacked on top of each other, with no space in between, been like that for 1.5 years now. I'm failed to convince that hdd coolers are needed and that they would just be dumping heat into the loop.

(note the expansion effect would be super small) but I think it would also be partly an issue of thermal shock, aka relatively cool and hot areas next to each other) as most any hdd I've seen has been steel ala not a good heat sink to begin with. So sinking one area of them wouldn't be good if there are tight tolerances

 

cooling on the sides and on the bottom vs. airflow is a widely disgussed subjekt, and i dont feel like turning this tread into a disgussion about that

but. I think that some cooling is better than nothing.
I dont know what the manufactors have written in their guidence when you'r pluggin a new one in.
I normaly just plug&play
------------- EDIT---------------
If you ask diffrent manufactors of HDD's, engineers of the designs, you will properly get multiply answers I guess...

 

yah, I'm assuming that most hdd's are well within tolerances in terms of heat. I just don't see a need to cool them at all. My drives run toastey but aren't anywhere near the 80C where you starting have issues, besides if they wanted to run them cooler, they would make sink them or at least not make them outta steel

 

Watercooling hard drives in areas with little airflow is a good idea. Side-mounting blocks are best, as hard drives are designed to give off most of their heat through the sides (typically so the case can act as somewhat of a heatsink).

I'd set up an Aqua Computer setup, with three independant loops that share a single Airplex EVO 1800 radiator (space for 15 120mm fans, but it can cool a fireplace passively).

I'd set up the gaming comp on one loop, the fileserver on another, and the hard drives on a third. Use three pumps and three reservoirs. For each loop, have a res->pump->waterblocks->Y-fitting->rad->Y-fitting->res, and have all three loops joined at the Y, and split back out into the reservoirs at the second one. Unlike Y-ing within a single loop, you won't have unbalanced flow issues as each pump is attached to it's own Y-section.

It'll cost a fair bit, but you can rest assured you won't have toasting hard drives. Performance-wise, you should be great considering it's passive - my water temp is a shade under 30c with my fans as low as I can set them (effectively stopped), and I've got a CPU, GPU, chipset and power reg chips (which tend to run REALLY hot) on a pair of 120mm Aqua Computer rads. Right now with my fans as high as I can set them while remaining inaudible, it's about 26c h2o temps, and about 25c at full tilt (fairly loud, and my case temp reads 24c). I'd imagine it wouldn't be too hard to mount the EVO1800 in a 1U space or so, but I'm not too familiar with the formfactor so it could be tricky.

Realistically, though, you'd be talking much more sense to just use a quiet 120mm fan (or two) for the hard drives. Or the biggest you can fit in the 2U space. As long as there's some airflow, they should be pretty safe. Pricewise, you're looking at about $300 in just hard drive waterblocks, versus probably $30 tops for a pair of quiet fans.

 

Discussion about watercooling hard-drives here...

 

wow my last message doesn't make much sense, but I feel that thread tends to vindicate most of what I've said. In all honesty the delta between the sides and center of the drive woudln't be that great thermal expansion is at a minimum this low on the thermometer anyways. still my point rests the perimeter of HDD's isn't even the hottest part or the part where the heat is generated or goes. I can run mine passively just fine and in my tighter case they are right behind a 120mm fan. WCing them seems execissive inefficient and just lowers the integrity of the WC loop.

I have enough faith that if engineers decided that hdd's needed more cooling they would at least mill the bodies with aluminum (Raptors are steel) or add some semblence of passive cooling

 

(I'm guessing that making HD's out of steel helps keep the magnetic fields controled, but that's just a guess)

 

The absolute maximum temp before Maxtor drives fail is 60°C, 55°C for WD and even lower for some others according to the datasheets. Not sure where you got 80°C from.

 

ahh yes magnetic shielding would make sense althought some drives are aluminum (probalby has more to do with cost) I believe some laptop drives are aluminum bodies with steel shielding on the rare earth magnet. I believe its more a casse of shielding the magnet/arm against damage, then shielding the platters.

80C is the the temp solid state components start having issues, I was just using that number but now that I've looked into it the 55C number looks correct, my apologies.

I couldn't find temperatures on storage review but if its my understanding most of the drivers around around the 32-40 deg C range in still air that of course doesn't take in account room temp and such but at those rates I believe just still air or moderate air cooling should be sufficient

 

Would have to disagree with SteveyG. I had a maxtor 80G IDE drive sat running happily away on a wooden test bench at 72°C for a month or two while i was "between cases." It was sat with no airflow and still runs fine. Foolish i know, however such is the learning curve.

 

Isn't that the whole point of a water block?

If you have access to your laser thermometer again try firing the laser down the screw holes for mounting the drives as i remember reading somewhere that are the main heat dissipating element.