I'm finally making my own block

[zoson]

So here's my design.

You can see the fins below the inlet barb which sit over the core of the cpu (Athlons only, including T-bird through Barton will work). The block is EXTREMELY low restriction on flow regardless of the fin design because the total surface for the water to travel through the fins is equal to the surface in the 1/2" channel.
You can also see the added surface area created by the curvey maze design. This will also cause turbulence (along with the size change in the channel immidiately following the fins) and increase cooling potential. The block will be machined 100% if possible, but if need be I can do the fins with a dremel.
I was thinking about putting dimples in the bottom of the channel much like swiftech does, but I'm not sure if that would have any negative effects - so information on this topic is appreciated especially!
Please comment and pass along suggestions. Thanks
-Zoson

[phreenet]

[21:37] <Zoson> feedback appreciated
[21:39] <phreenet> looks nice zoson

feedback

[Spamz0r]

i think you like that design to much.... im not sure what that white stuff is

[pippin88]

Originally posted by Spamz0r
i think you like that design to much.... im not sure what that white stuff is

It's a watermark of his username.

[Spamz0r]

Originally posted by pippin88
It's a watermark of his username.

damn it, you ruined my fun !

[bigben2k]

That's a very good design Zoson. Glad to see you're taking the plunge!

Personally, I'd drop the whole wiggly channel around the central area, but if it's wide enough, it won't hurt anything. It certainly doesn't improve cooling though.

You design demonstrates the convection effect, optimized with coolant flowing through the fins.

It might be a little more restrictive than you think though: even if the cross section between the fins is equal in area to that of a 1/2 ID tube, you have to convert each channel to its hydraulic equivalent. You can find the calcs here:
http://www.lmnoeng.com/pipeduct.htm

There's also an "orifice plate" type of restriction, as the flow is split, entering the finned area, and leaving it.

The rest of the flow restriction will come from the 90 deg bend at the inlet, and through your wiggly channels.

As for cooling, it'll work very well, based strictly on the convection effect, and the flow rate. Using a higher flow rate shouldn't improve things a whole lot, so this'll work with a relatively small pump.


If you've been following Cathar's design, then you should know what he did: he throttled the inlet so that it would shoot water down between the fins: this adds turbulence, and improves cooling even more. There are other ways to do this.

[jaydee]

Another thing is you do not have enough free area on top to properly seal it.

[zoson]

@bb2k: Thanks for your input, I wanted a block that doesn't restrict flow because the rest of my system is optimized for high flow. Also thanks for that calculator, I had no idea about the hydraulic restriction problems. If I made the ends of the fins pointed like the one fin in the DDM3 would that help to reduce the problem? Lastly, on the issue of the wavy channels, I did that to induce turbulence, and to lengthen the channel itself - thus also increasing surface area.

@jaydee116: how much area on top do I need? I did that drawing freehand, so it is not quite to scale. The channels should have 2mm of free space at the closest point to the edge of the block. I'm planning on using a soldered on copper top, but if this also becomes impossible I can always make the block larger and add an o-ring, and also use the mounting holes around the socket. Currently I'm planning to use a clip.
-Zoson

[jaydee]

Originally posted by zoson


@jaydee116: how much area on top do I need? I did that drawing freehand, so it is not quite to scale. The channels should have 2mm of free space at the closest point to the edge of the block. I'm planning on using a soldered on copper top, but if this also becomes impossible I can always make the block larger and add an o-ring, and also use the mounting holes around the socket. Currently I'm planning to use a clip.
-Zoson

You can attemp tp solder it, but chances are there will be ALOT of solder dumped into the channel with that thin of sealing area. Give yourself at least 3.175mm or 1//8". Last thing to skimp on is the sealing!

[Spamz0r]

Originally posted by jaydee116
You can attemp tp solder it, but chances are there will be ALOT of solder dumped into the channel with that thin of sealing area. Give yourself at least 3.175mm or 1//8". Last thing to skimp on is the sealing!

a friend had some solder/whatever it is get into the channel, but wouldnt that create more turbulance ?

[LiquidRulez]

Originally posted by Spamz0r
a friend had some solder/whatever it is get into the channel, but wouldnt that create more turbulance ?

Yeah.......but turbulance = more flow restriction
So I would skip soldering it and change the dimensions of the block to accomodate an O-ring or something to that effect.

[zoson]

I didn't really see a problem with having solder in my block. I'm going to be silver solder, so it's extremely conductive, and will help cool the block! I'm also not worried all that much about flow restriction because the amount of solder in the channel isn't going to be TOO much, maybe some drips but other than that it will like to stay where there is flux.
-Zoson

[Spamz0r]

Originally posted by zoson
I didn't really see a problem with having solder in my block. I'm going to be silver solder, so it's extremely conductive, and will help cool the block! I'm also not worried all that much about flow restriction because the amount of solder in the channel isn't going to be TOO much, maybe some drips but other than that it will like to stay where there is flux.
-Zoson

wait a sec are you going to SOLDER the top on not weld ?

[8-Ball]

That is the most commonmethod of fixing to pieces of copper together other than bolts. I have never seen anything welded from copper, though I'm sure it can be done.

Welding is more approproate for joining two edges together, whereas soldering works well enough for joining two areas together as with most waterblocks.

8-ball

[Spamz0r]

so....

is silver solder expensive ?

[bigben2k]

Originally posted by zoson
@bb2k: Thanks for your input, I wanted a block that doesn't restrict flow because the rest of my system is optimized for high flow. Also thanks for that calculator, I had no idea about the hydraulic restriction problems. If I made the ends of the fins pointed like the one fin in the DDM3 would that help to reduce the problem? Lastly, on the issue of the wavy channels, I did that to induce turbulence, and to lengthen the channel itself - thus also increasing surface area.

Making the ends pointy would improve thins, yes.

As for the wavy channel, unless you're thinking about a pelt, that whole area doesn't even need a channel, especially with a 2mm thick baseplate: the heat won't spread out that far.

[Volenti]

Originally posted by Spamz0r
so....

is silver solder expensive ?

for the amount needed to seal a block? no

under $10, most probably under $5, you'll want oxy equipment to do it though, you have to get the copper almost cherry red to use it (your school will have oxy equipment though)

[zoson]

@bb2k: could you explain why having a 2mm baseplate would cause the heat to not spread to the edges? I was under the impression that 2mm thick was actually a pretty thin base. How thick should it be? If what you say is true I may as well just make a block with fins, similar to that new block from france. I thought giving the water more time to absorb more heat would improve cooling. I also thought that because my block is more channel than copper that this would cause more spreading of the heat.
Finally, about the peltier issue, I have no intentions of using a pelt with this block. Wouldn't putting a pelt under this block make the area with fins not too useful? The finned area currently sits directly over the cores of all the athlonxp cpu's.
Serious reconsideration going on now.
-Zoson

[Spamz0r]

Originally posted by zoson
@bb2k: could you explain why having a 2mm baseplate would cause the heat to not spread to the edges? I was under the impression that 2mm thick was actually a pretty thin base. How thick should it be? If what you say is true I may as well just make a block with fins, similar to that new block from france. I thought giving the water more time to absorb more heat would improve cooling. I also thought that because my block is more channel than copper that this would cause more spreading of the heat.
Finally, about the peltier issue, I have no intentions of using a pelt with this block. Wouldn't putting a pelt under this block make the area with fins not too useful? The finned area currently sits directly over the cores of all the athlonxp cpu's.
Serious reconsideration going on now.
-Zoson

well, the pic below is ilike to think of the heat spreading, as the plate is thicker more chancec it has for the bubble like area to expand and widen out, thats why Cathar had all the cooling direct and a very small area as his plate4 is like 1.someting thick

[bigben2k]

That's pretty much it.

Cathar found out that a 1.5 mm baseplate doesn't need fins beyond 2mm past the core.