Custom made water cooling

T

TheArchitect

Weaksauce
Joined
Feb 18, 2004
Messages
89
I was able to start and finish this project in the time the fourms were down. But here are the results.

Please forgive for how slow these pictures come up, my school site seems to be bogged down.
004.jpg

021.jpg

027.jpg


the water block is 3x2" and has 12 1/16" wide and 1/4" deep and each channel is approximately 2" long. When you add up the surface area along the channel it come out to just over 15 in*in of surface area!
The radiator is a little overkill comming in at 7.5"x17"x0.75" but ill add in the GPU, NB, and possibly another CPU in the future.

I bought quite a bit extra metal stock to make redesigns and also make some for my friends and hopefully it will pay for itself. :D

The temps are about 8C lower in idle than my 900A w/SF2 on full blast and about 15C lower under load. Temps are higher than i was expecting, but it is starting to get hot here in Sacramento. I suspect that my water block is too open to get the turbulance I want, so im going to redesign the top piece to have some inserts into the channels to increase restriction and water velocity.
 
Nice waterblock. When you say slow you really mean it! :D Try photobucket or something.

I am always amazed when people build their own waterblock esp with all those fins and stuff :eek:
 
Nice waterblock. I was also thinking of building my own, but I didn't have everything together.

Also, what type of radiator is that?
 
Nice custom waterblock :cool: I see that jet impingement attempt :D Is it just me or does that rad seem a bit restrictive? :confused:
 
Its just you. ;) the radiator is just a moster, it is only 0.75" think so its less restrictive in terms of airflow. It uses 5/8" tubing for the core and inside of that tubing it has aluminum insets that divide it up into 3 channles of a little less than quarter inch tubing to increase rad to water contact. But it only has 6 passes vs. smaller radiators that have like 3 or 4 rows or tubing that make a lot more passes.
 
That block looks great, what kind of tooling was used to cut the slits? To really take advantage of the slit/turbulating features, you'd need to have the inlet directly above the center of the block where the turbulence-inducing features are. Otherwise it's not going to matter what kind of turbulence you induce at the inlet, the flow most likely will be re-laminarized whle flowing through the slits. It's also not necessary to have slits across the whole block, as the vast majority of the heat flow is directly above the core of the CPU.

Also, that Al (?) lid is going to corrode very quickly in contact with the copper base, you should consider making a copper or plexi/lexan/acrylic/whatever lid.

I'm not trying to crap on your design here, because it looks awesome and seems to perform as well... but just a little bit of constructive criticism :D
 
zer0signal667 said:
That block looks great, what kind of tooling was used to cut the slits? To really take advantage of the slit/turbulating features, you'd need to have the inlet directly above the center of the block where the turbulence-inducing features are. Otherwise it's not going to matter what kind of turbulence you induce at the inlet, the flow most likely will be re-laminarized whle flowing through the slits. It's also not necessary to have slits across the whole block, as the vast majority of the heat flow is directly above the core of the CPU.

Also, that Al (?) lid is going to corrode very quickly in contact with the copper base, you should consider making a copper or plexi/lexan/acrylic/whatever lid.

I'm not trying to crap on your design here, because it looks awesome and seems to perform as well... but just a little bit of constructive criticism :D
Click to expand...

Thanks for the suggestions, i put those turbulance inducing feature there to make sure that over the cpu the flow was not laminar. I was going for maximal surface area and unless you use a bunch of jets over the cpu i would figure putting a inlet right over would just make dead-spots in the center of the jet where the water on the edge of the jet would flow outwards keeping teh water on the inside confined. Putting an inlet/outlet on either side i would think would make the most of the surface contact as long as it stays turbulent.
There is a nice 1/8" of copper between the channels and the cpu, im not expecting a huge temp gradient between the cpu and the channels because of this base. Could be wrong though, i'd figure a good water block would be an encapsulated 900a or something aside from being incrediably restricting.
I have access to a machine shop so i am using nice CNC mills :D
how bad is this Al-Cu corrosion? is it only direct contact? or just close proximity, cuz i got a silicon seal layer between the two.
 
galvanic corrosion will get bad fairly quickly if you have aluminum on copper and you run straight water through it. adding water wetter + antifreeze (the kind with anticorrosions additives) can help but the best bet is to avoid having copper and aluminum in the same loop (i.e. 100% copper block, radiator with solid copper tubing, etc. with no aluminum in there) If the aluminum is anodized, that will stop galvanic corrosion unless the anodizing wears off (which can happen over time).

search the forum for galvanic corrosion or battery effect -- that should bring up quite a bit of info on it.
 
weapon-- said:
galvanic corrosion will get bad fairly quickly if you have aluminum on copper and you run straight water through it. adding water wetter + antifreeze (the kind with anticorrosions additives) can help but the best bet is to avoid having copper and aluminum in the same loop (i.e. 100% copper block, radiator with solid copper tubing, etc. with no aluminum in there) If the aluminum is anodized, that will stop galvanic corrosion unless the anodizing wears off (which can happen over time).

search the forum for galvanic corrosion or battery effect -- that should bring up quite a bit of info on it.
Click to expand...

Thanks for the heads up, i am using anti-freeze with anti-corrosion stuff in it so hopefully it wont be too much of a problem but im going to look into getting some acrilic for my next top.
 
TheArchitect said:
how bad is this Al-Cu corrosion? is it only direct contact? or just close proximity, cuz i got a silicon seal layer between the two.
Click to expand...


Direct contact would be horrible, close proximity will still probably be bad considering the lid and base will still essentially be in direct contact (connected by bolts).
 
zer0signal667 said:
Direct contact would be horrible, close proximity will still probably be bad considering the lid and base will still essentially be in direct contact (connected by bolts).
Click to expand...

being connected by water is bad enough for corrosion to happen

better have waterwetter in there to stop it, or at least make it slow down

how much would you perhaps make one of those blocks for me, and also it looks like the tubes to the block are kinked.
 
kronchev said:
being connected by water is bad enough for corrosion to happen

better have waterwetter in there to stop it, or at least make it slow down

how much would you perhaps make one of those blocks for me, and also it looks like the tubes to the block are kinked.
Click to expand...

I was going to say the same thing about kinkage
 
kronchev said:
being connected by water is bad enough for corrosion to happen

better have waterwetter in there to stop it, or at least make it slow down

how much would you perhaps make one of those blocks for me, and also it looks like the tubes to the block are kinked.
Click to expand...

The tubes were just twisted when i took that shot, that was when i just installed it. They aren't there anymore.

Its more of for friends kinda thing cuz the materials run about $10 to make that block but there about 3 to 4 hours of CNC machining, and since that was my first time machining copper i broke 2 bits in the proccess finding what rate and speed i could cut at. I wouldn;t feel right charging for them cuz i would make them at my leisure, and charging even $10 and hour would make it expensive as is.
 
TheArchitect said:
The tubes were just twisted when i took that shot, that was when i just installed it. They aren't there anymore.

Its more of for friends kinda thing cuz the materials run about $10 to make that block but there about 3 to 4 hours of CNC machining, and since that was my first time machining copper i broke 2 bits in the proccess finding what rate and speed i could cut at. I wouldn;t feel right charging for them cuz i would make them at my leisure, and charging even $10 and hour would make it expensive as is.
Click to expand...


how about 20


:D come on i'm trying to buy custom blocks from now on ;)


who am i kidding i cant even afford one
 
kronchev said:
being connected by water is bad enough for corrosion to happen

better have waterwetter in there to stop it, or at least make it slow down

how much would you perhaps make one of those blocks for me, and also it looks like the tubes to the block are kinked.
Click to expand...


Look up galvanic corrosion, you'll see that 3 conditions are needed:
1. Electrical contact
2. Ionic contact, as in a solvent with available electrolytes or ions
3. Metals having dissimilar corrosion potentials

The two paths are needed to form a complete circuit through which electrons can flow. It's basically the same idea as a battery- electrons are not flowing in or out of it until you connect the two terminals.
 
Galvanic corrosion = bad

Distilled water has little or no contaminants in it but bacteria growth can still happen. At the very least add some water wetter which increases the pH so it delays bacteria growth for a very long time.

Having only distilled water and mixing metals in a loop will cause corrosion. Galvanic corrosion happens when electron transfer can be created between two metals. A circuit will always have the most positive current. If you look at the reduction potentials of Cu and Al, you will see that Al will oxidize and lose electrons to Cu.

Reduction potentials:

Al3+ + 3e- --> Al -1.66 V
Cu2+ + 2e- --> Cu +0.34 V

The circuit will always gravitate toward the most positive charge so you flip the Al reduction equation. That makes it +1.66V, which makes the circuit have a positive charge. So the oxidation of Al will happen and Cu will gain those electrons. It is not a matter of if, it is a matter of when.

If you mix metals (even if you have an anodized aluminum top since it only takes one break in the aluminum oxide layer to cause corrosion), I think it would be best to use corrosion inhibitors like methanol or anti-freeze. But adding more than 20% IMHO will degrade performance too much by increasing the viscosity and decreasing flow.

This was from another post of mine at another forum. BTW Cu-Cu corrosion does happen also.
 
DaveX said:
BTW Cu-Cu corrosion does happen also.
Click to expand...


If you're still talking about galvanic corrosion, then yes, it can happen within the same metal. It is generally in alloys that don't have homogeneous composition throughout the alloy though, not in high-purity metals. For example, many steels and iron alloys are processed so that they have a microstructure that has separate phases- one rich in carbon and the other mostly iron with much less carbon. The carbon-rich phase here could potentially cause the other phase to corrode galvanically even though they're both part of the same alloy.
But then again, most of the copper we use for waterblocks is pretty pure and not likely to have crap in it that would cause problems like that.
 
You must log in or register to reply here.
Back
Top