My Aluminum block Rev. 2.0.

[jaydee]

Here are a few pics of the revised version with the gasket:






So far no leaks overnight. I will hook it up to the XP1600+ tofay sometime. I have a few minor things to change like the screw holes are to close the the mounting holes butit still works with the springs. I may move them slightly still though. My test so far have not included ANY other form of sealant other than the gasket itself. If I do ever sell any of these I think I will add some Silicone to it anyway.

[[dannzig]]

your blocks seem to evolve into something really nice ! personally i think that Al blocks are not much worse than the Cu ones... i also like that thay aren´t as heavy. you can count on doing with me in a couple of months when i get my economy together again.

[thebigNil]

Could you easly counter sink the screw holes so that the screws are flush with the top of the water block top? It looks good so far but I think that counter sinking the screws would make it look a bit better.

[[dannzig]]

and personally i would use hex-bolts or whatever they are called in english...

[Fixittt]

very nice.

What I would do for the screws, is use a flush mounted hex screw. The ones that are flat on top and have a chanfer on the underside. That way you dont weaken it if U totally countersink the screws. I wish I had some pics of what I am talking about, if I remember, I will take some tonight, I did it in my very first open chamber block I made.


So, both top and block are totally flat? with only the gasket in the middle? that has some potential.

man I wish that My T mill software would run my desktop mill. I really like it. is that what U use to make your G code? or do U use some other CAM software?

[jaydee]

Originally posted by Fixittt



So, both top and block are totally flat? with only the gasket in the middle? that has some potential.

man I wish that My T mill software would run my desktop mill. I really like it. is that what U use to make your G code? or do U use some other CAM software?

Yes pretty flat anyway, yes about the gasket but I am still not 100% sure it will work. I am thinking about using rubber sheet type gasket(if i can find any). I think this stuff might start leaking after time as it is just fiber type stuff.

Yes, I draw the object in ACAD2000 then make my tool path and save as a DXF file. I then load it into the "My T Mill" software using the import DXF file deal and it writes the code. Right after that you have to select "optimize tool path" and then that code is nearly ready to use. I then just copy and paste each path I want to make and change what ever depth of each path. it takes some minor code changes to get it the way you want butit works pretty good.

[jaydee]

Yes I was thining about counter sinking but i will need to move the screw holes to do that as they are to close the the mounting holes. This is just the first prototype.

Thanks for the input any more is welcome.

[Brad]

is your water channel the same, or have you modified it in anyway with this new revision of block?

[jaydee]

Originally posted by Brad
is your water channel the same, or have you modified it in anyway with this new revision of block?

Nope it is exactly the same. I am trying everything possible before I have to change the channel. I like the serface areas of it and don't want to have to change it. Anyone know where I can find sheets of rubber about 1/16" thick? I think rubber will be the best.

[Brad]

maybe an auto shop? for sealing car windows, or something like that?

[jaydee]

Well my thougts about it seeping through after time have proven true. It started leaking through the sides. I have since took the old gasket off(completely soaked!!!) and put a new on on but this time smeared it with silicon. I think this will work. I am going to let the silicon setup completely then I will retest it. I still think a rubber gasket is the way to go. Trying to track down rubber sheets.

[jaydee]

Found a place locally that sells sheets of rubber gasket material!!! Have no idea about cost or anything yet. I will stroll down there tomorrow and pick some up if the price is reasonable.

[[dannzig]]

i dunno what its called in english but we have some kind of gasket-goo that we used as gaskets on your mopeds when i where young.... maybe you could trie that.

[jaydee]

Originally posted by [dannzig]
i dunno what its called in english but we have some kind of gasket-goo that we used as gaskets on your mopeds when i where young.... maybe you could trie that.

I used to use permatex, kinds black tar looking stuff. The problem with that is it will blow out the sides with pressure. The gaskets i was using and will use are the same deminsions as the block. One solid peice with just the holes trimed out. That way they cannot blow out on the thin sides I have on the block.

[webmedic]

Jaydee on the block you gave me for testing it was giving temps about 10c higher than my dd maze2. This was with a duron 1gig overclocked to 1240 with 1.85volts. Sorry I can't confirm this as my duron fried so I will have to rma it.

On another note Alpha tech sent me a pal8045 for modifications so that it may become a watercooling monstrosity. This is going to get interesting. The only problem will be that it would probably take at least 55$ - 60$ to build a water block out of one of these.

[jaydee]

Originally posted by webmedic
Jaydee on the block you gave me for testing it was giving temps about 10c higher than my dd maze2. This was with a duron 1gig overclocked to 1240 with 1.85volts. Sorry I can't confirm this as my duron fried so I will have to rma it.

I would have to say it was mounted wrong or not tight enough. I am running an XP1600@1523mhz 2.05Vcore with the original Aluminum block I made and It is running a toasty 30C under SETI load at this moment. I will drop my newest one on shortly and report back.

Not having the top on correctly will easily make it mount wrong. If the top is not 100% level with the rest of the block then it will mount crooked or not fully. Also if you got anything in the channels it will cause those eddies to form and just make a hot whirlpool of water.

I also ran it in my Abit KT7A system for a day and got 34C under load which is the same board you had it mounted to? That had the XP1600@1523mhz at 1.85Vcore. No problem with heat. I am running a lesser system aswell with 170GPH pump and 1/4" ID barbs and a tranny cooler. Can't see it being the blocks problem but I will double check on the new one.

10C is way over what any design should be. A chunk of Aluminum with a hole drilled through it will get better than 10C from a DDMaze 2. The worst design there is shouldn't have but a 10C difference in the best there is.

On another note I got the drawing back and ran it through the software but it came up wrong on the code with my mill software. The drawing in ACAD looks good though. I am not sure what to make of it yet.

[Brad]

yeah, you must have had it mounted on an angle, or with not enough pressure. Or maybe stuffed up on the AS2 use.

just try remounting it again.

[jaydee]

webmedic:

I am not sure why that block was running that much higher. I installed an identical one exept with the mods above and I am getting almost identical temps with the original. Actually a little better. Right now water temp is 28.5C and load under SETI is 34C. I know it has to be pretty close to accurate or I wouldn't be able to get 1523mhz out of the XP1600+. Anything over 40C and she locks up.

I am wondering if the faster water speed of your system is causing it. I am running real slow water speed through my system. Maybe the extra time the water is in the block it picks up more heat??? I am only geting about 60GPH out of my system. Damn tranny cooler is restricting most of the water speed which maybe a good thing actually. The water is comming out of that rad only a few C above room temp. My bets are the slower speed and cooler water coming out of the rad. AND/OR it wasn't mounted quite right. Not sure what you used to get that top on there after I pulled the epoxy loose.

I am going to work on a better testing setup tomorrow. I am going to either mill a channel in my block and put a probe nect to the core or do it like the proccoling article and grind a channel in the socket. I feel the channel in the block would be more accurate though as it should touch the core. If you touch the bottom of the CPU then the heat has to go through the hole CPU before it gets the the probe.

I am going to pick up that rubber gasket stuff tomorrow and stop by radioshack and seee what they have for probe parts.

[OcPunK]

Hey could you make custom Blocks for Video cards?

[jaydee]

Originally posted by OcPunK
Hey could you make custom Blocks for Video cards?

Well I havn't decided yet. I do not have the specs of of newer cards nor a way to test them because I don't have a newer card.

Once I get my regular CPU water block down I may consider it. I would need to buy a version of each card I want to make a block for so I can test it properly and I don't have the reasources for $100's of dollars in vid cards just yet.

[webmedic]

1. I tried it a couple times before I crushed my cpu.
2. Yes I'm trying less flow but I don't have another cpu here to test with yet. I found a good german pump that carries a 2 year warranty and is dc. They are sending me a demo unit so I can test it out. On my system I also tried a 80gph fountain pump that gave the exact same temps as the ehiem I was using before. I have been watching the watts that these small pumps produce and I have been trying to stick to the one with the smallest wattage output. In this way less heat is introduced into the system to begin with. The rio pumps by comparison put out almost twice the watts. This is probably why the rios heat the water a couple degrees warmer than the ehiem.
3. The cpu I was using was an overclocked duron. They are known to be harder to cool because of thier smaller surface area. Maybe if increased the thikness on the bottom of the block it may help to spread the heat better at least for the duron. Shortly after this I crushed it so I was unable to run a second round of tests to confirm this. Also the arctic stuff does seem to give a few degree better temps after a few days.

[jaydee]

Originally posted by webmedic
1. I tried it a couple times before I crushed my cpu.
2. Yes I'm trying less flow but I don't have another cpu here to test with yet. I found a good german pump that carries a 2 year warranty and is dc. They are sending me a demo unit so I can test it out. On my system I also tried a 80gph fountain pump that gave the exact same temps as the ehiem I was using before. I have been watching the watts that these small pumps produce and I have been trying to stick to the one with the smallest wattage output. In this way less heat is introduced into the system to begin with. The rio pumps by comparison put out almost twice the watts. This is probably why the rios heat the water a couple degrees warmer than the ehiem.
3. The cpu I was using was an overclocked duron. They are known to be harder to cool because of thier smaller surface area. Maybe if increased the thikness on the bottom of the block it may help to spread the heat better at least for the duron. Shortly after this I crushed it so I was unable to run a second round of tests to confirm this. Also the arctic stuff does seem to give a few degree better temps after a few days.

I just got a new Duron 1gig today (to replace a Duron I crushed) and will see what it does. I will mill that little channel in the bottom of the block and put an external probe touching the core to get a more accurate result.

That 12V pump sounds interesting. It should be able to run off the regular comp power supply. I have a 12V 500GPH beilge pump for a boat that only takes 1.5amps. Works great on a comp power supply but is to big for what you are planning.

[webmedic]

The only thing to watch out for on the bildge pumps is that some companies don't make them for continuious use. They will burn out after about two months use. I have already confirmed this with becooling he makers of the aquastealth kits.

[jaydee]

Originally posted by webmedic
The only thing to watch out for on the bildge pumps is that some companies don't make them for continuious use. They will burn out after about two months use. I have already confirmed this with becooling he makers of the aquastealth kits.

Oh yeah, I only uise it for an hour at a time. It has a 3 year warranty and was only $10 so I am not to worried about it. I will be around if it fails as it will mainly be used to cool my Mills motor.

[jaydee]

Pic of thermoprobe in block. What do you think??? It touches the side of the core.


i will hook it up in a few and see what it says.