My personal Mods. (New D.I.Y. water cooling project)

[Outlander_04]

Not sure if you have used epoxy as a glue , but if you have it will probably fail .

Most epoxies have a low heat distortion temperature , Its not resistant enough to heat.
Additionally if you have metal sealed with epoxy its likely to fail after some heating and cooling cycles because the expansion and contraction are different .

I think your mods are awesome BTW . Im almost inspired to make a waterblock for my computer .......

 

[Shadow703793]

^+1. You need some thing like this:
http://webtronics.stores.yahoo.net/hiteepenandp.html

(There are different epoxy ones used in furnaces, cars,etc so check them out too, usually under $15 on ebay)

 

[ir_efrem]

+1 "gray text on black background annoyed me."

Made my eyes bug out on me a little. Good information for the most part. Please keep in mind that an informative article needs to be professional. A professional write up needs proper grammar. I am not the spelling/grammar police but if you want people to take you seriously................

 

[ir_efrem]

My epoxy is not used as a glue, except in non critical areas. Like the plastic pen tube being epoxied to the heat retention clip, it doesn't matter if the epoxy fails there because the plastic has no where to go.

The epoxy is used as a filler though, on the reservoir and the water block itself. I had some reservations on using it actually, but heat resistance has never been a worry. JB Weld has very high heat resistance. http://jbweld.net/products/jbweld.php
Shadow > the epoxy you linked is good up to 527F and JB Weld is good up to 500F

Almost all types of seals can fail, even well soldered joints suffer from what looks like "cold weld" when in all actuality it was the different expansion of different metal types. I have used quite a few types of epoxy and I will probably stick with JB Weld and I am not worried about it leaking. I wasn't planning on using it on my next water block anyhow so it's a moot point.

 

[Shadow703793]

^I have heard of JB Weld how good is it (bond wise)? The one I personally use is Durapot 865. It's hard to get hold of but very strong bond for metals (or so I am told). I'll see if I can get my hands on some JB Weld to play with . I got the 865 off a friends dad who works for Northrop (or Lockheed; not sure which lol) .

 

[ir_efrem]

I have never had it fail on me if that tells you anything.

The harshest application I have used it on is an external crack on a Chevrolet 350 engine. The crack was formed by some dork leaving water in the engine during winter, luckily the only crack was on the exterior of the block. The epoxy held for 4 years until he sold the truck... When a surface is prepared properly it holds its bond very well. All that really needs to be done is sanding and cleaning with a non residual cleaner. Residue is the bane of all bonding agents.

The outside of an engine block gets up to 200F easily... In the winter here it gets below zero. So if it can stop antifreeze from leaking in that application, I don't think I will worry to much about my water block....

 

[Shadow703793]

^True, if it can survive in the engine it should survive in a PC. Now I have to see if the local hardware places are selling this (else it's ebay lol).

 

[Outlander_04]

You will get the best bond between epoxy and metal if you apply the epoxy to the metal and then wet sand with coarse sand paper . That stops a layer of oxidized metal forming in the time the fresh surface is exposed to the air .

I still have my doubts that you can make a 100% water tight system with epoxy that will last but I wish you luck

 

[Proximon]

Epoxy works good for boats I patched dry rot in the hull of an old 36' wooden cruiser with a product called Marinetech, which is basically the same as JBWeld. Patch is a mild word actually. It was more like replacement. Worked fine.

 

[Outlander_04]

Epoxy works good for boats I patched dry rot in the hull of an old 36' wooden cruiser with a product called Marinetech, which is basically the same as JBWeld. Patch is a mild word actually. It was more like replacement. Worked fine.

When Im not ranting about computers I have a day job as a composite engineer and designer/builder . My company designs and manufactures boats , kayaks and automotive components .
The kayak on my profile is one of my products . Its 6 meters long , around 500 mm wide and the carbon and kevlar versions weigh in around 15 kg [ 20 ft , 20 inches, 33 lbs ]

I wish the OP and others all the best with these metal/epoxy joins but I personally would not rely on joins like that to keep all water off sensitive electronic components long term .


Proximon I hope you removed 3-4 ft of planking past where you could see your dry rot . The spores will have reached well into the timber and your problem will probably come back if you havent . Copper oxychloride saturation could help and so could boric salts on the timber which should both kill the spores

 

[Shadow703793]

Hey man, thought you'd be interested.

Here is my work so far:

^Testing the pump/res (The pump is inside the res; the red sandwich box). The block still needs work, and held by clamps. The rad has copper lining inside so galvanic corrosion is not a problem. Cost so far: $25 (excluding tools). Not bad eh? Pump (an AC pump) is rated at 130 GPM (~10 Gallons per second). Everything is epoxied using JB Weld (Yes! I bought it to check it out).


Any comments OP?
mid plate is done, doing testing

 

[Proximon]

Proximon I hope you removed 3-4 ft of planking past where you could see your dry rot . The spores will have reached well into the timber and your problem will probably come back if you havent . Copper oxychloride saturation could help and so could boric salts on the timber which should both kill the spores

We were actually going to replace the transom properly later, but as it turned out the whole project got scrapped before we completely lost our shirts
We got discouraged after about 30 hours spent just trying to get one bolt loose on the outdrive. It looked good, but the body was completely fused with electrolysis.

Just an old Chris Craft type of cruiser from the early 60s that had spent the last 20 years in an enclosed shop. We picked it up for a song, got a little use out of it, spent too much on it and dumped it cheap We made sure the actual hull was sound.

This was all many, many years ago, but I do recall using something on the wood... there was some sort of pre-treatment followed by some sort of thin epoxy that soaked into the wood to preserve it.

 

[Outlander_04]

Id like to hear back about how this cooling system performs in operation , and again in 6 months to check that all is going well .

 

[ir_efrem]

Looks good so far Shadow. I have to wonder about the radiator though, Might be a good idea to get plenty of air flow through it. By the looks of it, air should flow with very little back pressure though. BTW what kind of retention bracket do you have planed?

I have a new design that I am working on. It has been halted until I get access to a drill press. It is somewhat similar to the copper block image that I posted at the bottom of page 1. I got a 2"x2"x.5" block of copper now and I need to drill holes in it. I will be drilling 7 Holes in the block at .25" each and copper tubing will be soldered into the block about .25" into it. I will be building 2 manifolds made of 1/2" hard copper pipe with .25" holes drilled into it for each of the 7 smaller copper tubes.

The tubes will all be slightly crimped right at the block to add turbulence to the water before it enters the block.

It wont be long before I get the holes drilled and I will post some images then.

Here is what I have drawn up for the block of copper and hole placement.

 

[ir_efrem]

I wouldn't mind hooking up my original water block to the E5200 to see how well it does on a dual core. It's quite a bit of work to drain the system though; in fact I'll probably add a drain system once I add the new water block.

I wasn't unhappy with my temps on the Q6600 but I wanted to tinker some more to see if I could build a better mouse trap. So I'll wait to test the E5200 for when I get the new block designed. Then when I have a drain built in, I'll test the old water block. I'll add some complete temperature numbers for both blocks with both CPUs at normal and overclocked speeds.

 

[Shadow703793]

BTW what kind of retention bracket do you have planed?

Working on it for LGA775.

I do have a ad-hoc one for S478 tough. Will post pics.

 

[ir_efrem]

Well I took the block to my brother-n-law's workplace where he has a drill press. Well it was a hunk of crap....

It drilled a hole at a god awful angle so I brought it home and I've been hand drilling the holes myself and with much better results. And yes I squared the table with the bit and squared the block in the clamp, the problem was the horrible wobble in the chuck. I went with 6 holes total. I will post some pics when I get the tubing soldered to it, probably tomorrow or the next day.

 

[ir_efrem]

Here is the final water block. I went with a slightly different method for bringing all the copper tubes together, this time they are soldered into 1/2" hard copper pipe. That's 6 holes drilled through the 2"x2"x1/2" copper block, and because of the crappy drill press there is a bit of a stagger in a few of the holes; surprisingly the hand drilled ones are almost dead straight through. Both ends of the 1/2" hard copper have a cap and a threaded fitting.

I vacuum tested the block and it's as water tight as a frogs ass. Now I need to get some more water - someone drank the jug I had....... I tossed a couple quarters on there for size reference. It's quite a bit bigger than I need, but I wanted to try a block with alot of copper to see how well it cools. I have some aluminum flashing now and it's time to make the fan shroud.

 

[Shadow703793]

What solder are you using? You may be loosing a heat transfer that way depending on how well the solder conducts.

 

[ir_efrem]

Solder type is unimportant, the tubing doesn't go all the way through the block. they are only about 1/4" into the block. Water flows inside of the block itself. If ya still want to know it's 40/60

 

[ir_efrem]

Thought I would throw up an image of the block lapped. It's all pretty and shinny.

 

[Tech-Boy]

That is really cool! Keep up the good work!

 

[Shadow703793]

Solder type is unimportant, the tubing doesn't go all the way through the block. they are only about 1/4" into the block. Water flows inside of the block itself. If ya still want to know it's 40/60

Ahh.. In that case, solder type is not much difference.

 

[ir_efrem]

"Pump (an AC pump) is rated at 130 GPM (~10 Gallons per second)."

Is that a typo bud? the pressure would be phenomenal if this were the case..

 

[AKM880]

Nice work on lapping that block, looks like you have alot of time on your hands and have mad skillz lol.