Discussion in 'Computer Cases, Power Supplies and Cooling' started by Toast, Mar 2, 2013.
How are you attaching the block to the motherboard?
Or did I miss something?
Based on how the rest of the mess looks, I'm guessing he will let the thermal paste hold the block on, as well as cool the CPU....
Looks have little to do with ghetto. Go used a block of copper and a drill/drillpress to make a waterblock. Then there is the fountain pump and glass jar. You are in downtown ghetto.
Based on how it is looking right now, I would guess something like a x pattern with a strait bracket (http://www.homedepot.com/p/t/202077...roductId=202077353&storeId=10051#.UTwtEBzrz-Q) would work and fit the theme.
Also why one should go with market instead of DIY.
Seems like I'm in the forum where everyone only thinks that decent cooling is possible by buying a H100i or a kraken x60, or by dropping hundreds of dollars on a custom loop... Sad world we live in.
The fountain pump has variable flow control, and is much more heavy duty than any specifically designed computer water cooling pump. The block is solid copper, something you won't find from other blocks, and thus highly thermally conductive.
Everyone somehow seems to be missing the fact that I have a radiator? It's a 2x120mm radiator that wasn't in that set of pictures. Yesterday I finally got it cleaned out and ready to go.
The block is going to be attached with a mounting bracket that I am yet to make, and the reservoir is an acrylic, watertight jar with a removable lid. The waterblock will get its leaks fixed today, and then it will be polished, and perhaps painted.
I paid less than 50 dollars for all of this, and I'm willing to bet that when it's 100% done, it will be on par with any $100+ water cooling kit, so I'd keep your pessimistic comments to yourselves until it's shown to not work well.
No one here is saying that decent cooling takes those components. But you could have very simply just grabbed a off teh market waterblock and had a much better component with very little out of pocket. You are taking a hell of a risk making your own block.
Power scales. If you pumped up a Swiftech MCP955 to 120 volts, it would be the same quality and around the same flow.
There is a good reason why you don't see that. It is a waste of money to make solid copper blocks. The only part that needs to be copper is the part directly contacting the CPU. Above that, it is purely there to hold the cold plate (the copper piece) and water flow. That aside, a highly conductive material will not overcome a poor flow design, nor make the difference between an amateur (you) and a team of thermal/Hydro engineers that put in 100000+ hours design work into a component (Swiftech, Asetek, corsair, heatkiller, etc.).
It is just an aluminum radiator. It will not perform that well, especially with a dated design. Though with your extremely high rate of water flow, I would look into getting a Transmission cooler for your rad. They are designed for that High flow rate. would also match the rest of your loop. http://www.autozone.com/autozone/ac...on-cooler/_/N-25io?itemIdentifier=151847_0_0_ for example. (you can get one at the pick and pull for round abouts $15-20, just have it flushed before using it).
Take a look at the strait brackets. That will be far easier. Just weld/solder a stud in the center of the block to attach them somewhat securely (5/16 should be small enough, without loosing strength. Just weld it to the top, slide teh center hole of the brackets over the stud, and put a lock washer and nut on the end to hold it together.) should come to about ~$5 or so depending on tax. Use them and a decently long bolt to match the mount points on the board.
There is nothing wrong with ghetto. It just means that you tried. The resv. will look better if you can paint it a decently dark colour to mask teh water inside.
Don't attempt to paint the block. As far as polish, you need to go through the grits down to 2500-5000 or so to make it worth your while as a waterblock. Anything less and your performance will not make the work worth it. (this is talking about the mating surface of coarse. The other sides don't matter what you do to them).
You will likely match a $100 kit, if you are talking about a Coolermaster or Corsair kit. Otherwise, you are not going to outperform a true loop with this. It could be made to do so if you put more work and $$ into it. But expecting a bucket of parts that someone with limited engineering knowledge to outperform parts specifically built to do that job is unrealistic. Expect it to outperform a similarly priced air cooler, that is where you will shine.
It is a neat project, but there are flaws.
Firstly, your copper block + aluminium rad will be causing corrosion. The channels in the block are already very limited, having gunk build up there and in those tiny channels that the radiator has, as well as the cheap pump, it isn't going to be great in a few months - a year or two when it has had time to really act.
Secondly, you have very little surface area for your block. That which there is, the channels aren't going to be all that great for allowng the water to flow uninhibited beacuse you didn't use the "proper" tools, so you will be having a hell of a lot of resistance up in there. Combined with the 90 degrees fittings, that pump that better be able to provide a lot of pressure to overcome this
Thirdly, The radiator leaves a lot to be desired, but I will leave that as the previous posts make this a well trodden path already.
I will still be following as it seems very interesting to say the least, but I don't hold a huge amount of hope. I would absolutely love to wrong though
Might seem a bit of an odd question, but if Aluminum and Copper cause galvanic corrosion in loops, why don't they corrode when together on an air cooler? Is it just a chemical reaction between them and the water?
The block is going to be polished for a few hours, first with sandpaper, then steel wool, then polishing cream, etc.
I'm using antifreeze as opposed to water. Less corrosion, and it's specifically made for cooling metal.
The channels in the block are quarter inch. There are three long channels which run 3/4 of the length of the block, and 2 channels which run half the length. Since it is a solid block, the heat will disperse throughout the entire block, thus making the channels more effective.
The pump is a 120volt and can pump up to 156 (I think) gallons per minute, but I think that may actually be too much. It has plenty of power to pump the coolant, trust me.
I appreciate the input, though. I hope to finish it up today.
I bought one for 16 euro shipped from eBay, it works great lol.
No. This will reduce performance. On the mating surface, the side of teh block that touches the CPU, do not use anything other than Alcohol and Sandpaper. For the sandpaper, start with 320 grit, and move through them to at least 2k grit. When you are done, wipe it down with alcohol (cleans the material removed off), and look. You should have a mirror surface. If you do not, then do it again, the right way.
Basically, from a scientific point of view, the smoother (more mirrored) the surface is, the better it will transfer heat. It is the same reason why some cheap coolers need to be lapped before use.
Negative. You are using 2 separate ideas to make it work, and it will not. antifreeze will reduce Oxidation (aka rust corrosion) in a system. It will not stop or reduce galvanic corrosion. It is specifically used in automobiles to stop the oxidation that occurs from having an open loop cooling system, with non galvanically reactive materials. Generally Aluminum Rad with Aluminum Block/Heads or Copper Rad with Steel Heads and Block. Some systems have carbon fiber or other exotic materials in the system, but as they are nonmetalic, it does not affect it.
Your system will have galvanic corrosion because of the Aluminum and Copper in the same system. If you want to stop that, you either need to get a copper rad, or get an aluminum block. One or the other will stop the corrosion, or at least reduce it (that is assuming that the prop inside the pump is not aluminum or copper. But they generally arent unless you get one with massive (like over 1000 GPH) flow rates).
The easiest way to fix this is to get a copper radiator. Like said above, get a transmission cooler. Affirm it is copper before you buy it. Should not cost much at all.
This is not exactly right. Yes Copper is conductive. the whole block being copper will to a point aid cooling. But it will not in the least make up for sloppy design, and poor workmanship.
The way you have it set up, you are trying to use more coolant to make up for poor surface area, and it is not going to work. Nothing will make up for a lack of surface area, and you have to give the water time to adsorb the heat.
Don't get us wrong. This is a neat project. But it is not going to work out how you want it to because you are not doing the proper research, and are half arseing the work. You could end up with a massively superior solution, or a low end solution. It all just depends on how much work you put into it.
What I'm wondering is where the other $48.50 was spent...cause quite frankly looks like you went to a junkyard.
I'm all for DIY projects, but it looks like crap, won't work worth a damn, and is essentially a waste of time---
I mean..cmon you used a spice jar as a res...and .10c fittings from a hardware store--you're not even trying.
And FYI, I've seen some impressive DIY jobs and was blown away at the skill, creativity, and performance..
Yours looks like a bunch of crap you found in a dumpster and superglued together while drunk...
and I mean that in the semi-nicest way possible.
Why are you breaking down his project? It's pretty interesting starting from 0 with no money. Just wait until he's finished and shows his results.
You can give tips, comments,... But don't be rude, if he likes his project and he thinks it's good, then it is ok.
It is also part of the learning curve.
Reminds me of the comments about Smile's cheap block not looking pretty.
All I want is the report on how good it works.
Certainly an interesting project. I'll be interested to see how it all looks once it's done and how well it performs.
I will, when I oc my athlon to 5ghz ;P.
Just waiting for the board.
Update on how it looks- It isn't done yet, though.
Everything is set in place, basically. I fixed all the leaks that were in the block and then spent about an hour polishing it... I think I spent enough time on it. Cleaned up the sides as much as I preferred.
Not the cleanest looking block, but I'm proud of it. First attempt ever, and basically my first time working with metal on such a large scale. All I've done before is soldered copper plumbing.
Bet you're all gonna get real mad at me. I used the normal AMD mounting hardware, and it works great. Sue me!
The coolant is just plain antifreeze/distilled water (50/50) with lots of dye added. Not the EXACT color that I wanted... but it's pretty darn close. I like it.
The radiator and reservoir are sitting beside it on my computer desk/nightstand. Right now, this is being passively cooled because I do not have my fans for this up and running.
It's very decent at transferring heat, as far as I can tell, as it only took a matter of minutes for the coolant to warm up once I ran COD4. Obviously, the coolant heated up more than I'd like, because there are no fans on the radiator, so it simply doesn't have any way to stay cool. My hope is to eventually do a push pull on the radiator- maybe with some better, non-computer oriented fans.. we'll see. There's always plenty of options when you know how to do anything you want (;
Passively cooled, the temperatures reached 50c while gaming, and that's about as high as they wanted to go. This is a big improvement from my previous 65-68c while using the stock air cooler, which would ramp up to 60-70 percent usage during gaming. This is silent, obviously, and I'm yet to put the fans in to better my temperatures.
The flow is just fine. Right now I have the pump running at about 1/3 speed, but I may lower it when I get my fans to see if having it move slower is going to help the liquid pick up more of the heat and also disperse it throughout the radiator.
Do I know what I'm doing yet or do I need to try harder?
You still are not doing it right. What exact method did you use to polish?
The mount method looks decent enough.
Rad is still going to end up galvanicly corroded. But it seems like you don't give a damn about that, so have fun killing your system a few months down the road.
"Galvanic corrosion (also called ' dissimilar metal corrosion' or wrongly 'electrolysis') refers to corrosion damage induced when two dissimilar materials are coupled in a corrosive electrolyte."
The metals are not touching each other, nor are they electrically charged. I don't see a problem with corrosion here.
Your coolant is the electrolyte.
The dissimilar metals make a battery.
Why do you think that water heaters have a zinc sacrifice rod?
Run water over anything and the water will pickup pieces of said material. You are running an Aluminum Rad with a copper block. You are creating a galvanic cell between them.
But as said, keep going. Don't do you research. It is only your system. Not ours (thank god).
Separate names with a comma.