i'd like to build gerwin's radiator
 

i have about 25 feet of copper pipe laying around, and from my other thread, you can see i want to cool passively.

would anyone happen to have instructions on how to make this? it looks fairly simple, but i want to get it right the first time around

You might check the price of solder-on elbows and "T"s vs. Gerwin's all-compression joint setup. Soldering pipe is not difficult. You can use 50/50 solder for this non-potable application too - it's easier to work with than lead-free solder.

Why not play around with the design, make it both functional and attractive? I believe the form of such radiators non-critical.

Place it under a window if possible. The (usually) colder air sliding down off the window offers some natural draft and is of course the coolest air in the room. We place electric heaters under windows for these reasons.

Looking forward to your creation.

You could make a large coil and bury it underground. You could do something like this. You should get tools to sweat all the joints; propane torch, solder, flux. You can use copper FPT adapters for ataching hose barbs.
http://www.brewtree.com/i/plumbing/Dsc06534.jpg
You'll need a pump that can supply a large head, someone else should make a suggestion on the pump but a D4 should be more than enough I would imagine.

Edit: For reference you can use this page for determining the pressure drop of you run of tubing in psi. You can use this page for adding in however many fittings you end up using. If the back pressure gets to high for a D4 you'll need another, or a different pump.

thank you very much. that was quite helpful. i shot a PM over to gerwin and he's going to give me a few tip's before i start my project. i may try make some original modifications of my own.

and yes, i will be soldering.

the problem with the garage radiator you showed me, razor, is the size. i'd make something using the same amount of copper, but it'd have to be wider so i can mount it on my wall. hopefully performance won't be affected too much.

if i used 90 degree elbows on gerwin's design to have more verticle pipes would performance be decreased or increased? i would imagine flow rate would be hurt.

oh and for pumps i was planning on using the swiftec mcp600 or the aquaxtreme equivilant. is that a bad idea? maybe i need something with higher gph like the L30?

There is probably a tangible benefit from mounting the radiator on the floor since it is much cooler and because the concrete slab can absorb a lot of heat, that goes for burying it too. If it's wall mounted then you just have air for cooling but maybe that's not so bad. :shrug:

As for what pump to get I'm not all too sure. Obviously the MCP600 and the D4 are highly recommended for their pressure but the 15' of copper pipe by itself doesn't give much head loss so the L30 wins by a wide margin. When you add in all the fittings and whatever length of tubing to get to your pc, plus the water block(s) it will be a different story. The pressure drop will probably fall around where the two pumps are equal. I made a graph of the L30, MCP600, and the copper tubing.

http://forums.procooling.com/vbb/att...tid=3464&stc=1

I'm wondering how you could estimate the pressure drop of the Gerwin style with all those T's in there. Copper.org's table shows an elbow as 1 extra foot and a side branch as 2 extra feet, I don't understand that part though. Why does a T have twice the pressure drop of an elbow?

I wish I had one of those 180 GPM MCP600s. :p

T's have twice the pressure drop as elbows? that doesnt make sense to me either... i suppose the radiator design you introduced to me would be the better choice then?

I don't get it, Groth? I swear the two flows were 165 (US) and 180 (Imperial)

Looking at the pressure drop of the tubing alone is irrelevant as the block makes up most of the resistance.

The Imperial gallon is larger than the US gallon, at a ratio of 1.20095 US gallons to 1 Imperial.

So, if its 180 US gallons, that would be 149.88 Imperial.

Ok it's fixed now. :D

basic fluid mechanics

for water to turn 90 degrees in an elbow, it slides up the copper
for water to turn 90 degrees in an T, it smashes into the wather flowing straight through

as for the water flowing straight trough, it catches the edge of the inside of the T, if that makes sense

I wonder how many ft. of line a res equals

If the reservoir is properly built, zero.

The pressure drop of both a tee and an elbow depend upon the tee and elbow in question. Some tees are sharp (all I've seen are) and some corners are equally sharp. For these sharp ones, the tees should have less resistance. A lot of elbows, however, are curver softly and not sharp, so greenman's comment is correct in that aspect. I think the issue with the straight through of the tee is that the sharp corner causes a lot of turbulence, maybe?

res "properly built" still has internal friction, unavoidable, so what relationship would that have to ft of line (not head)?

obviously dpenedent on the res in question, but I was making a point that resivors are generally more restrictive, IMO....no test to back me up, though

internals of a T do cause turbulence

A res is like a large diameter pipe so the resistance is extremely small for the ammount of water we flow. The way in which the water enters and exits the res can obviously affect the head loss though, I think that's what AngryAlpaca meant. If you have a res where the inlet comes straight in the bottom and exits straight out the top then what you've added is a big pipe. You can't increase back pressure by adding a larger diameter section of pipe can you?

I also added the D4 to my above graph.
http://forums.procooling.com/vbb/att...tid=3467&stc=1

I wonder no one's mentioned "Y"s then. "Y"s branch cleaner and merge cleaner, and require not much more effort to set in. They're typically shaped like "K"s without the front leg; branch is 45 degrees. Very common fitting. Then use an equally common 45 degree bend to get 90 or parallel.

Good plumbers consistently use "Y"s over "T"s except where flow will be greatly restricted anyway (by a waterblock or vanity sink) further down both lines. Ahem.

Turbulence shouldn't be an issue in a big passive pipe rad. There I think the priority should be getting the largest surface area for the time and money invested, and if the rad's sprawled across a wall it ought to please the eye too.

hmm - surely what you've added is a big pipe with some air trapped in it

Wouldn't a better alternative to 15ft of continuous pipe be 5 parrallel pipes of 3 ft, or any of parrallel setup. It would mean less resistance so the flow would be higher.

It was a hypothetical example. Still, the air would be pushed out the top to be removed elsewhere in the loop or you could see the system as being already bled.

but surely the whole point of having a res is to trap the air...

well, i've begun cutting pipe. my paralells are 2.5 feet 1/2" piping, and i'm using 3/4" piping horizontally (t's and pipes inbetween them). and i'm soldering not using compression fittings

i'm only going to have 12 paralells when im done, as i only had three 10 foot pipes. hopefully the radiator will be able to cool the water down to acceptable temperatures. afterall, the only thing thats being cooled is my processor...

Making the branches narrower was smart. So you used reducing "T"s?

yup 3/4" to 1/2"

I tried building a Gerwin style radiator with soldered joints...it did not work very well. The copper pipes expanded with the heat and broke the adjacent solder joints. :cry: If I were to do it again, I would put towels soaked in cool water around the area I am soldering to minimize the heat expansion. Good luck :)

i dont understand, are you talking about the heat from the propate torch, or the heat from the water flowing through your radiator once it was complete

From the propane torch while you are building it. I doubt heat from the water would be an issue since the temps are much lower, and all the rungs would expand equally.