My PinGrid1 block.

[jaydee]

Built and tested this block sometime ago and have been waiting for pH and Joe to get somethings with the site setup but that hasn't happened yet so I am going to post this now in here.

This blocks development is solely for DIY'ers for an easy and effective block to make. Here is my first set of results. Interesting that the block with the outer pins removed faired better.

PinGrid1 (green) has all the pins in place. PinGrid1B (red) has the outer pins removed. PinGrid1C (blue) has ALL the pins removed.


More info and pictures of the block can be found here: http://customcooledpc.com/forums/index.php?topic=15.0

I am working on PinGrid2 right now that has deeper holes drilled in to the base so the pins are longer...

[jaydee]

Also here is a graph with other blocks for comparison:

[jaydee]

Errr, some pics?
The holes are only 1/8: deep on this one. PinGrid2 the holes are 3/16" deep.

On the last pic it shows my second test done with out the outer row of pins. I did a 3rd test with no pins at all. I didn't get a picture posted with the barbs. All they are are copper connectors soldered to the top. They have a true 1/2" ID.

[ljohn787]

Looks pretty good. I wonder why it performs better with the outer pins removed.

[brucoman]

Originally Posted by ljohn787

Looks pretty good. I wonder why it performs better with the outer pins removed.

I would think less restriction around area beyond IHS contact

[killernoodle]

Because it allows the water to first flow out to the outer edge where it will then return to the outlet rather than flowing straight to the outlet. Having pins on the outer edge keeps the water from utilizing all of the pins.

[jaydee]

Originally Posted by ljohn787

Looks pretty good. I wonder why it performs better with the outer pins removed.

Pretty much summed up by the last two posts. Still a lot more experimentation to do. I also am revising my die simulator for more accurate results. As it is now the probe is to close to the heater adding to much heat to the results.

[mikhail]

Don't flame me Im a newbie in WC, so If you don't mine a question or two?
How do you bind the plexiglass to the copper water block? Is it standard epoxy type or anything else. Can you recommend a brand? or a link step by step. please...
Thanks!

[Incoherent]

Like it Jaydee. Been tinkering with something similar.

Have you tried with sharpened pins? Would maybe reduce restriction a bit. The flat tops probably help turbulence a little but are probably more restictive.

Also, how about soldered pins drilled all the way through the base. Better heat path direct to the die then.

[FL3JM]

Originally Posted by Incoherent

Also, how about soldered pins drilled all the way through the base. Better heat path direct to the die then.

I agree with this one, i believe it should perform a bit better with the pins going right through the block as the current design prohibits a good heat transfer to the pins (metal against metal, even tho tightly squeezed against each other, ain't that good. ).
The best thing would be to drill right through the block, put in the pins, then heat the block with a torch or something untill the pins melt to the block. Though i dunno if thats possible, but it would be cool.

BTW, nice testing!

[Moto7451]

I think trying to fuse the pins to the base with a torch or something like that would warp the base. Soldering the pins in with some Silver solder wouldn't be bad though. Also putting some AS5 or AS5 Epoxy on the pins before putting them in would probably help with transferring heat if melting/soldering the pins in isn't possible. I imagine with the tight fit that the pins (probably) have the AS5 wouldn't be washed away.

[Etacovda]

Originally Posted by Moto7451

I think trying to fuse the pins to the base with a torch or something like that would warp the base. Soldering the pins in with some Silver solder wouldn't be bad though. Also putting some AS5 or AS5 Epoxy on the pins before putting them in would probably help with transferring heat if melting/soldering the pins in isn't possible. I imagine with the tight fit that the pins (probably) have the AS5 wouldn't be washed away.

See, theres this funny thing we do called lapping, which will remove the distortion from the base.

[jaydee]

Originally Posted by mikhail

Don't flame me Im a newbie in WC, so If you don't mine a question or two?
How do you bind the plexiglass to the copper water block? Is it standard epoxy type or anything else. Can you recommend a brand? or a link step by step. please...
Thanks!

I used clear RTV silicone on that block. See those 4 bolts? Those are holding all three peices together. Put a layer of silicone between each layer.

Originally Posted by Incoherent

Like it Jaydee. Been tinkering with something similar.

Have you tried with sharpened pins? Would maybe reduce restriction a bit. The flat tops probably help turbulence a little but are probably more restictive.

Also, how about soldered pins drilled all the way through the base. Better heat path direct to the die then.

I didn't sharpen the pins on the top but I did on the bottom to try and fit the V shape the drill bit leaves for more contact. I think sharpening the pins on the tops on the one's around the inlet might help some aswell flow wise.

The copper wire I uses is slightly wider than the 1/8" holes. I had to tap them in so they are not just floating in there. I imagine one could take a towel with ice in it and lay the base of the block on it and then take a torch to the pins and maybe they will expand into place better.

Drilling the holes all the way through and soldering then sounds like a good experiment. I could face it with my mill. For the average DIY'er though is might be hard to lap the base.

I havn't had any time to mess with blocks lately. I got everything packed up and ready to move. After I get moved I intend to do a lot more stuff.

[Incoherent]

Originally Posted by jaydee

The copper wire I uses is slightly wider than the 1/8" holes. I had to tap them in so they are not just floating in there. I imagine one could take a towel with ice in it and lay the base of the block on it and then take a torch to the pins and maybe they will expand into place better.

You'd be better off doing the opposite I think. Heat the base and chill the pins, that way they expand into the base and the base would shrink around them. Might be a little tricky doing so many but...

[Jag]

Originally Posted by Incoherent

You'd be better off doing the opposite I think. Heat the base and chill the pins, that way they expand into the base and the base would shrink around them. Might be a little tricky doing so many but...

It's been done before, by this user here

[jaydee]

Originally Posted by Incoherent

You'd be better off doing the opposite I think. Heat the base and chill the pins, that way they expand into the base and the base would shrink around them. Might be a little tricky doing so many but...

Probably right. Might be able to freeze the pins into a block of ice and then hit the base with the torch. Ice might melt to fast though.

I like the soldering idea to. What I am going to attempt is drill the hole all the way through the block and then chamfer the holes on the cpu side. That way the solder can fill into the chamfer. I will stick the pins a little farther past the base so when it is lapped the copper will be exposed and the solder will be inside the chamfer area. I think that would be the easiest way to solder the pins in as opposed to trying it on the other side of the block with all the pins in the way.

P.S. sorry for the pics being down, my host has it's head up their ass....

[Incoherent]

Originally Posted by jaydee

...the copper will be exposed and the solder will be inside the chamfer area. I think that would be the easiest way to solder the pins in as opposed to trying it on the other side of the block with all the pins in the way.

Thats the only way you want to do it really, in order to avoid getting solder around the base of the pins. That would impede heat transfer to the water a fair bit I think.