Custom Watercooling Part II [Pictures to follow]

[Overconfidence]

Awright, I posted this a few weeks back, but now I'm at home and ready to tackle my watercooling loop. If you don't know, I'm making a watercooling setup with the least amount of money possible (Especially since I'm young, and don't have the money to spare in the first place). I picked up quite a few parts while on summer vacation, and will have pictures of them up as soon as I can. They are:

Heatercore - approx 18cm square, copper, 5/8" connections, 2' thick

Fan - Commercial fan for air conditioning. About an 18cm fan, 4 bladed. AC motor and fairly loud, but will attach it to a DC stepper motor and see what I can do.

Pump - Not sure, looks about 3/8 or 1/2 inch connections (Haven't mesured yet), but it's a fuel pump. I may have to buy another pump, but this one'll do for now. Runs on 12v, a fair size (pics will show)

Reservoir - Washer fluid container from a car for now.=W will make an acrylic 2 bay res myself, but the washer fluid container will do for trial runs. The 2 bay res will go behind the front window, and the thermometer on the front of my case will be in the water. This pic explains better:


http://img.photobucket.com/albums/v644/Overconfidence/comp3.jpg

Tubing - I plan on buying some milk hose from a local place, since my dad says that that type of tubing will be better than your standard clear tubing.. but prove me/him wrong.

Connections - I'll either buy standard barbs/hose clamps (Though I think I have some hose clamps around home), get some threaded connections (origionally restrictive but if I bore out the inside it'll be non-restrictive and more secure than barbs), or I could turn some custom connectors out on a lathe. To put whatever I choose onto the heatercore, I assume I'll just braze or silver solder.

That covers pretty much all of it (unless I'm forgetting something) except for the waterblock.

For that I've been thinking of using a whitewater-esque design. I just got a hold of a CAD program today, and I came up with something quick (It's not that great). Of course, it's disproportionate except for the size of the barb connections which are fairly close, and the size of the channels/O-rings. I can't figure out how to make the one O-ring (looks like 2 circles with lines connecting, is supposed to be like a racetrack but ended up like a cassette tape, best I can explain), but everything else is in the right place/shape, just not scaled correctly. The bottom chunk is .4 inches, and goes down .3 inches (are those proportions decent?), the middle chunk is only .1 inches, and the top is .3 or so, but will be as large as needed for the barbs/connections/whatever. Anyway, here's a wireframe and also a shaded model of my basics.

http://img.photobucket.com/albums/v644/Overconfidence/customwater5.jpg
http://img.photobucket.com/albums/v644/Overconfidence/customwater4.jpg

Any input on this is appreciated, pictures of the remaining 'junk' will follow!

[cartmanea]

"Milk hose" is probably just medical/food grade tubing, which is what Tygon is.

[Overconfidence]

Ah, alright, it probably is.. It's not completely see-through, though.

Anyway, I'll do the pics tomorrow.

[Overconfidence]

Alright, here are the pics as promised (if anyone cares):

Heatercore: About 7" across, doesn't look copper but it is, I'll take some more pics without the rubber guard. 5/8" connections as mentioned before.
http://img.photobucket.com/albums/v644/Overconfidence/wc-core1.jpg

Core again, with comparison:
http://img.photobucket.com/albums/v644/Overconfidence/wc-core2.jpg

Next 3 pics didn't turn out very good, I'll fiddle with the focus/zoom and show a few more.

Fan:

http://img.photobucket.com/albums/v644/Overconfidence/wc-fan.jpg

Reservoir: (Yeah, yeah, it looks really bad. I'll be giving it a thorough wash. It's only being used for the ease, because all of the holes are already there.)

http://img.photobucket.com/albums/v644/Overconfidence/wc-res.jpg

Pump: I'll get the mesurements when I go to take the second set of pics.

http://img.photobucket.com/albums/v644/Overconfidence/wc-pump.jpg

It may not seem like much right now... But I'll be cleaning everything off, repainting whatever I need to (contemplating sandblasting the heatercore and giving it a black finish). Second set of pics to follow.

[Overconfidence]

Okay, here's the second batch of pictures.

One from the side of the pump - Apparently it's a Mitsubishi pump.

http://img.photobucket.com/albums/v644/Overconfidence/wc-all001.jpg

This one shows the front - The inside diameter of that threading is 1/2 inch.

http://img.photobucket.com/albums/v644/Overconfidence/wc-all002.jpg

One of the heatercore from the side without the rubber.

http://img.photobucket.com/albums/v644/Overconfidence/wc-all003.jpg

Heatercore - All copper.

http://img.photobucket.com/albums/v644/Overconfidence/wc-all004.jpg

The fan from another angle - Will replace the AC motor with a DC one and slow it down to speed with a (I believe) 3077 transistor.

http://img.photobucket.com/albums/v644/Overconfidence/wc-all005.jpg

Any problems you guys see, point them out to me right away!

Another thing I just found out... I may have access to a submersible pump with larger diameter connections.. are submersible pumps decent to use? I'd assume they'd be quieter...

[cartmanea]

Good dimensions on the heatercore, what car is that from?

[3Tripnip]

That heatercore looks awesome, I'm also curious to know what car it's from.

[Activate: AMD]

lol, thats gonna be the ugliest WC setup ever :p:

[Bloody_Sorcerer]

unless he stealth blacks everything like weapon... then things look fantastic ;)
the only big problem with submersible pumps is that they have to be, erm, submerged and that they dump 100% of their heat into the water (inline pumps also dump some into the air)

[Overconfidence]

That heatercore is from a 1/2 ton Ford truck that I found lying around the dump. I'm not sure what year it was, but it was fairly old.. say earlier than 80's. It was really easy to get at, because it was accessible from under the hood.

Alright, I couldn't find the submersible pump anyway. I'll stick to the other one (I think it's a diaphragm pump, inline). And yes, I know it looks ugly right now. I'm going to sandblast the heatercore (I'll take a pic when it's all shiny), and then paint it black with an airbrush (I forget what the paint's called, it's a black finish my dad uses for gun parts after he turns them out on his lathe).

I'll do the same with the fan, no sense having it all green and ugly =P. The reservoir is pretty much the ugliest thing in existance, like I said, I'm going to make a dual bay acrylic (Does lexan work? I have some of that stuff right next to me..) reservoir in the front of my case, which'll look awesome. The one in the pics is just for prelim tests, before I make it all look good.

The pump will still look the same, but I'll make some sort of aluminum case (I know a place I can get it bent into shape for fairly cheap) for the heatercore and pump. Now that I think about it, I could have that hook up directly to my actual case, and have the hoses going through the motherboard side of the case. Or I could have the hoses from my reservoir go through the front of my case to the other mini-case.

And I'm also making ground on the waterblock, I took apart my computer yesterday and got all the mesurements for the mobo mounting holes and clearance room. If anyone's interested, you can make a waterblock 3.5 by 1.9 inches, with the mounting holes 2.6 by 1.4 inches apart. All on socket A, mind you. Thankfully, the holes are all at easy to mesure distances. I used calipers accurate to 1/1000 of an inch, and one was exactly 1.4 inches and the other was 8 thou off of 2.6. I'll start designing the real block today.

Just one question... how much room between the water and the CPU core do you want? Would it be profitable to have 1/8 of an inch of copper there, or do you want more? Less? How much room do you want the water to go through? In that prelim design, I have .1 inches of copper at the bottom, with the channels .3 inches tall. If anyone has knowledge in this department, any help would be great!

[Bloody_Sorcerer]

hunt around for cathar's development of the WW threads over at procooling or OCAU (i'm not sure where exactly they were...) he found some "magical ratio" at which things worked best or something. all i know is that all his blocks have reeeaaaally thin bottoms.

[cartmanea]

Well, thinner bottom and heat is transferred to water faster, but cpu temperature range is greater. Thicker bottom and cpu temperature is more constant but transfer to water is slower.

For some reason I would think thicker bottoms would be better. You might not achieve as low temperatures on average, but there would be less variation.

[Overconfidence]

Yes, but your reasoning works this way, too (Numbers are all made up, btw):

Thin bottom: 30 degrees idle, because the cooling loop can keep up with it, and 37 degrees under load.

Thick bottom: Heat transfer from core not that great, so runs at 35 degrees idle. Nearly the same amount of heat transfer from core to waterblock, so 37 degrees load.

Anyway, I'll look for that "magic ratio", thanks!

[cartmanea]

That didn't really make sense. The thicker bottom wouldn't raise core temperature at idle, I think it would just slow the transfer of heat to water since the copper could absorb more heat. So smaller range of temperatures from the cpu and slower transfer to water. Not less transfer, just slower since the extra copper could hold more.

[Bloody_Sorcerer]

it could hold more, yes, but it would be overwhelmed by the CPU alot easier, as since the heat slows down the farther it gets from the source, it doesn't transfer into the water as well, so suddenly the waterblock becomes a blast furnace.
If thin bottom was worse than thick bottom, we'd all be using thick bottoms.

[cartmanea]

That makes sense, thinner is better since thermal resistance in the copper increases as it gets thicker.

[Overconfidence]

That didn't really make sense. The thicker bottom wouldn't raise core temperature at idle, I think it would just slow the transfer of heat to water since the copper could absorb more heat. So smaller range of temperatures from the cpu and slower transfer to water. Not less transfer, just slower since the extra copper could hold more.
Yes, because you have more heat in the copper, more heat on the core... *shrug* The explanation the next post gave was what I meant.

[Activate: AMD]

if you're on the cheap a T-line is really the way to go IMO.. the res is ugly and the t-line is like a 2 dollar part at most, sure bleeding is slower, but for a cheap and easy setup theres no really compelling reason to not use it

[Overconfidence]

Err, So I shouldn't make a res out of acrylic that I have lying around?

[Bloody_Sorcerer]

T-Lines are actually around 5.60 for 10 at mcmaster... :)
acrylic cracks over time. people still use it; just words of caution.

[Overconfidence]

Hmmmm, okay. I might make a copper res, instead, then. Or I'll just use a t-line. That's the least of my worries ATM.

I just tested the big fan at 7volts, and it seems to run fairly well. The only noise is really the motor noise, and even if I put it right next to the heatercore, I can barely hear any turbulence. It seems to push more air than the 70 CFM fan I have in my case. Then again, I can't REALLY tell.

http://img.photobucket.com/albums/v644/Overconfidence/bottomlayer.jpg

Here's a pic from the final version of the waterblock design! I'll finish the two next layers tomorrow, or at least by the weekend. I'm still awaiting a few mesurements before I move on.

[Overconfidence]

Okay, here're a few general questions that I'm also posting at another forum:


1. Right now I'm looking at some threaded fittings. They fit into my design quite well, but they're only 3/8" ID because they're intended for high pressure, but I can bore out the inside of them to be almost exactly 1/2", since there'll be almost no pressure in my system. The only problem is that if I bore them out, it exposes the plain aluminum (since they're anodized or whatever). Would that be alright (since it's only on the inside, or should I keep it at 3/8" or not use 'em all? Also, if I do bore them out, should I do that only to the middle one, since you can't put double the volume out than in anyway?

2. Is glycol alright to use as corrosive inhibitor?

3. Can I make the top 2 pieces of my waterblock (that are not used for transferring heat) out of aluminum, as long as I protect the outside from corrosion? Would it be a problem on the inside? I don't see why, considering my dad's truck has an all aluminum heatercore, and it's perfectly clean (Glycol in the system).

4. Also on aluminum, does tubing made out of it work? I could get some thin-walled stuff that would work pretty good... again on the corrosion.

5. Do I have to make a backplate for my waterblock (To mount on motherboard)?


I'll edit this post if I have any more questions!

[cartmanea]

DO NOT USE ALUMINUM AND COPPER IN THE SAME LOOP!

Who cares about protecting the outside, the inside is where corrosion occures. And about mixed metals in cars, I'm not sure why there is a difference, but it could be due to the greater distance between copper and aluminum components. Also, they use a 50% glycol mixture in cars and mixes tend to be lower in watercooling setups.

[BGP Spook]

DO NOT USE ALUMINUM AND COPPER IN THE SAME LOOP!

Who cares about protecting the outside, the inside is where corrosion occures. And about mixed metals in cars, I'm not sure why there is a difference, but it could be due to the greater distance between copper and aluminum components. Also, they use a 50% glycol mixture in cars and mixes tend to be lower in watercooling setups.


I can tell you form experience that when they used to use mixed metal in car cooling loops there deffinitally was corrosion happening. If you have ever looked inside a radaitor from a iron block motor with about 100k miles on it you would easily identify corrosion build-up. (Every heard of "flushing" your radiator? Ever wonder why they do that?)

In modern cars most of the metal is aluminum. High wear parts that don't have direct contact with the cooling loop are steel, and a few other parts may be made of copper by necessity.

[Fairydust]

I can tell you form experience that when they used to use mixed metal in car cooling loops there deffinitally was corrosion happening. If you have ever looked inside a radaitor from a iron block motor with about 100k miles on it you would easily identify corrosion build-up. (Every heard of "flushing" your radiator? Ever wonder why they do that?)

In modern cars most of the metal is aluminum. High wear parts that don't have direct contact with the cooling loop are steel, and a few other parts may be made of copper by necessity.

And why did this corrosion take place so "quickly"? Maybe because people hardly ever change the anti corrosive? While most Antifreeze/Anticorrosive mixtures keep their thermal properties over long periods of time, the corrosion protective properties degrade over time.

Not mixing metals is the easiest was to approach the problem, but corrosion in mixed metal loops can be reduced by a rather large degree with the right inhibitors. Mixed metal blocks are harder to protect as the electro chemical reactions will happen a lot quicker than with components divided by several inches or even feet of water (water is not a good conductor). But in my experience the corrosion in mixed metal loops with anticorrosives is less of a problem than you guys make it sound. I do however agree that prevention is easier than protection and removing any mixed metals from the loop is more effective (Oxygen corrossion can still happenn though and necessitate a corrosion inhibitor).

[Overconfidence]

Alright, thanks.

Is it bad to use 50% glycol in a cooling loop? Again, I'd like to know the deal on this.

I'll probably make my entire block from copper then.

For the tubing I can do clear stuff, but for metal, copper tubing would work (if the block is all copper)? Just a thought, would stainless steel be okay, too? I know it's not too bendable, but I guess that gets back to your mixed metal stuff...

And then, would I not be able to bore out the inside of the aluminum fittings, because then I would expose the aluminum?