I'm new to XS but am immensely impressed with the information to be had here.
I've built water cooled rigs in the past and recently got the itch to upgrade. The thing is, I like to take extreme departure's from the "every-day" so this project is a bit non-standard. I want to make an external cooling system that removes the rad, fans, and pump from the room altogether (i hate heat and noise!). I've identified some parts but after reading some of the stickied topics I really need to check some of my assumptions. This sudden uncertainty comes from reading the likes of HESmelaugh's CPU block tests, Cathar's effect of tubing sizes/fittings, and MaxxRacer's pump recommendations among others.
Originally, I approached this project with the following basic assumptions:
Assumption #1: Radiator: The more surface area the better
Assumption #2: Pump: The higher flow/head the better
Assumption #3: Fan: The more CFM the better
Assumption #4: Tubing: The bigger the better
Based on these at least partially invalid assumptions (based on Cathar's tubing study), I've made the following choices:
Rad: I found an auto-heater core made of copper and brass on eBay that's roughly 8"x8"x2". It's not "PC sized" but since it's going into a custom enclosure, it doesn't matter. Also, I felt more surface area was better since I'd be supplying a beefy air supply. It's also dirt cheap at around $8US from eBay.
Pump: Following the theme of "external" and "overpowered" I decided that I'd only pump enough DC current to the remote location to keep a relay circuit engaged and use that to power an AC pump. I chose the Eheim 1250.
Fan: If you think I'm crazy now... check this out http://www.grainger.com/Grainger/items/2C647. In case the link is broken, it's a 134CFM squirrel cage blower made for commerical machinery and has an exhaust port that's about 3x4 inches.
Tubing: At the moment, I've got it all spec'd out with 1/2 I.D. and compression fittings
My home office shares one wall with my garage. That's the wall that I'm proposing to pass the tubing/wires through. I'm not at all concerned about making the enclosure and have only minor concerns about insulating the hole in the wall and dressing it up to look "commercial." I just don't want to go down the path unless this combination can actually achieve my goal of removing additional heat and noise from my office while still leaving me with an over-performing PC.
Specifically, I'm curious about things like the added heat dump from the 1250. Will it be too much for 1/2" tubing or could I still switch to 3/8" as Cathar suggests in his review? What about the fan/rad combo? The squirrel cage blower is powerful enough that even if my enclosure is somewhat restrictive, I'm feeling confident I'll end up with 50-75CFM through the rad. In winter, it will make a nice "garage heater" lol... but will such an assembly be able to overcome the high ambient temperatures in a Chicago summer?
your 4 assumptions are correct, and I like the time you take to post. Welcome to xs. The only thing that I'd say is that bigger tubing results in diminishing returns, because as cathar said in the stickies, there's only little gain to be had from getting more flow. (1/4 to 1/2 results in less than a degree worse iirc) Also, don't go with the eheim 1250. If you want pressure, go for multiple ddc 3.25s in series or an rd-30 at 18 volts.
I would go for it, using 1/2" tubing and the RD-30 as suggested above. I'm not experienced enough to predict how effective the heater core would be, but I guess it should be fine.
I've been dreaming of doing something like that for a long time, but it's just not possible where I live at the moment. No noise and heat would be heaven.
How much do you plan on cooling?
The rad is about the same size as a 140.3 so keep that in mind. Also, what is the fpi count on it?
The pump suggestions above are good. The Eheim is actually quite weak, more so than a stock DDC 3.1.
Also, I am not entirely sure that fan will be great. In order to get that blower to fit on the rad you will have to make a custom shroud where the rad end would be drastically larger than the size of the blower. This will really hurt the air pressure and the results of the fan probably won't be as good as you would expect and certainly won't justify the cost. You would be better off with 3-4 strong 120mm fans.
Thank you both for your replys. I checked out the LD-30, that's a nice pump but it's also DC powered. I know you can get PSUs to support the load or maybe use a 2nd PSU but I'm leaning towards an AC pump. Besides, given my design concept of removing heat/noise from the room, I kind of worry about driving that extra load with the PSU thats inside the case. It's going to be bad enough with the pair of 295s I'm planning ;-)
I like the design look of the Iwakis but can I ask, what's the concern w/the Eheim? I thought that it was a common workhorse a few years back? Alternatively, what about the slightly smaller WMD-20? That's AC powered and about $100 cheaper than the LD-30.
a few years ago our water blocks weren't as restrictive as they are now, most people end up splitting up loops to keep good pressure/flow and temps. I don't think an 8"x8" will be enough, as said its only about a 140.3 rad size, that alone would handle the i7 but not the 295's, if you could pick up a second rad you could probably set them up with a shroud and blower may be enough for both of them.
Meaning just add a 2nd rad into the same loop? The shroud construction wouldn't be a problem as I have a full workshop in the basement. In fact, building the enclosure is part of the fun ;-)
Regarding faster3200's point about the necessary shroud... I do realize that the shroud would need to expand in size between blower and rad which would cause a pressure drop, ie. slow down the air velocity. But it's still 134 CFM isn't it?
Don't know how well it translates to your heater core, but Martin did a review of a Pontiac Bonnie heater core. Dimensions are about the same. With one 133CFM (Scythe Kaze Master 3000RPM) fan he got 800W heat dissipation out of it at a 10C air-water delta. That's pretty good, but even a $50 rad (MCR320) can do the same with some high speed fans.
The reason we recommend a different pump is that the Eheim 1250 is a little weaker than the usual pumps (MCP355, MCP655). Nothing wrong with it otherwise.
Regarding faster3200's point about the necessary shroud... I do realize that the shroud would need to expand in size between blower and rad which would cause a pressure drop, ie. slow down the air velocity. But it's still 134 CFM isn't it?
You would be much better off just buying standard fans. A 3000RPM UK would be roughly the same as that blower but 1/6th the cost and take up much less space. For the same price as the blower you could buy 3 crazy rpm Deltas that would outperform the blower by a mile. This is my point about that.
Quote:
Originally Posted by alacheesu
Don't know how well it translates to your heater core, but Martin did a review of a Pontiac Bonnie heater core. Dimensions are about the same. With one 133CFM (Scythe Kaze Master 3000RPM) fan he got 800W heat dissipation out of it at a 10C air-water delta. That's pretty good, but even a $50 rad (MCR320) can do the same with some high speed fans.
I don't think you can compare that review to what he plans to buy. For one, different sizes, as Boogerlad mentioned. Second, fpi don't look even close to the same.
For this effort and money to end up with a 10C delta would be awful. For all this work I can't imagine anything less than 5C delta to be acceptable.
what? no they aren't. This is almost twice the width, and a bit shorter.
I don't follow. OP says his is 8x8x2, or a surface area of 64. Martins is 6.5x10 7/8x2, or a surface area of 70.
Quote:
Originally Posted by faster3200
I don't think you can compare that review to what he plans to buy. For one, very different sizes, as Boogerlad mentioned. Second, fpi don't look even close to the same.
For this effort and money to end up with a 10C delta would be awful. For all this work I can't imagine anything less than 5C delta to be acceptable.
Either I'm blind or we're not looking at the same heater cores. Surface area is roughly the same, and counting FPI I also got roughly the same result, one around 10 and one around 11.
i don't see it mentioned in my quick read, but i would be worried if your garage in not insulated or heated/cooled if you have drastic temperature changes outside of your house.
check out my build, same idea, but mine is right above my case, wall mounted.
Either I'm blind or we're not looking at the same heater cores. Surface area is roughly the same, and counting FPI I also got roughly the same result, one around 10 and one around 11.
Yeah, idk whats going on but I can't maths worth a damn today. Must be finals week tiredness.
Still, 10C delta is crap for all this effort and the blower will be weak sauce.
Hmm. Well this is why I posted in XS hehe. So based on what I'm hearing:
a) I don't have enough rad surface area, especially considering the temp changes that are likely to occur
b) the blower is out since it wouldn't produce a high enough air velocity
c) the pump (Eheim) is out due to general weakness
I'll go back to the drawing board and re-post w/some new ideas.
Oh and Spawn... that thing looks like some kind of alien Rad-tree lol!
where do you live? depending on the temperature there you may be able to pick up a big enough rad and just run it with some slow speed fans or passive.
Actually, those small centifugal blowers work quite well... that style is often used in ventilation systems because of excellent static pressure, and low noise. I installed almost the exact same one on the exhaust hood of the vapor degreaser I built at work... I chose it because it was mounted inside the clean room and a noisy 100+ CFM fan wasn't practical.
That said, ebay might be a better place to look, for a slightly larger model.
Also, I would go with a copper car radiator.
For pumping, something like one of the little March mag-drives would be nice... they have similar head pressure to most DC pumps used in PC watercooling, but much higher flow... you could run 5/8-3/4" ID for the remote lines and run a couple parallel loops for your components, that would get the GPM up and take advantage of the low resisance and large surface area of the car rad, plus it would help negate any resistance from the longer run.
__________________
Phenom II 720BE
Gigabyte 790FX-UD5P
OCZ Platinum AMD edition DDR3-1600 2x2Gb
XFX 5850 1Gb
X-Fi Xtremegamer
Lian Li PC-A05B/Corsair 750HX
A separate computer PSU for powering pumps and fans, one or two DDC/D5 and Fast 120 mm thick fans seems like a easier and possibly cheaper solution. Then the blower and AC pump solution. Sure a RD30 is cooler but considering the cost just go with "standard" parts.
__________________
"I don't have a masters degree, but even I know that mixing water and electronics is a stupid thing to do." My girlfriend.
"It's easy to get sideways at first thinking there is ONE BEST product for each area of water cooling. But, that's not always the case - depends on your exact setup and needs." shazza
Ok, to satisfy my personal curiosity, I think I need to clean out the workshop this weekend and setup the blower/rad in a test rig. I want to see how much air it can actualy force through the heater core with a sealed shroud. Meanwhile, I'll hunt for a larger rad and blower. I'll try to come back with some pics in a couple days.
Nothing beats cooling your water with cold air ... If the temperature in your garange is in the mid teens (c), I wouldnt worry too much about what you listed in your post, just make sure you have sufficient pump power and your good to go...
Linear Mode
