MY Al Water block

[ibmkg]

Sorry for low quality pics. I was in a hurry to install this WB and thus took pics via cell phone.

Running windows vista and idle CPU temp is 38C with system temp 30C and one 120CFM fan running at 5V (120CFM@12V). With full load (using CPU Stability Test 5.0 by Jouni Vuorio) the max temp I got is 55C (FAN at 5V, 24hr test) with system temp reaching 36C max.

Soon as load is taken off, temp drops to 42C and eventually 38C.

Using a Heater Core, MCP655. The rig consists of about 1liter of tap water (no coolant). 1/2'' barbs with steel reinforced plastic (clear) pipes.

Currently I am replicating the same design using Cu. Awaiting Comments.

[bigben2k]

55C seems high, but we don't know what kind of CPU that's on...

There's a part covered in tape here: could it be a copper disc by any chance?

[jaydee]

Pretty sure I seen a design pretty much the same here before. Make sure you don't run those brass barbs very long. last time I used brass barbs on a AL block the top nearly got eaten through in less then a month.

[ibmkg]

bigben2k:

Its on a P4 3.2Ghz HT. The tape part is lapped contact area for CPU. Its raised to about 3mm thus to avoid block touching the caps. Its pure Al.

Room temp is 28C. I think the Al is not conducting heat fast enough. The thickness of the Al base is 2-3mm.

The temp of CPU is not constant (changes with load immediatly to 55C and reduces at once when load is removed to 38-40C) and I am currently blaming Al for that. While the water remains cold (almost near room temp i.e. 28-30C)



Jaydee:

I have been running this over for a month now. I have not seen sign of corrosion as of yet. The barbs are screwed in with nyrol spacer to seal it off.

This is a large block. The four 'big' holes are mounting holes for LGA775 (see pic 1).

Should I go ahead and replicate the same deisgn in Cu? Should I remove the outer most track and thus reduce the size of my next block?

[jaydee]

Originally Posted by ibmkg

Jaydee:

I have been running this over for a month now. I have not seen sign of corrosion as of yet. The barbs are screwed in with nyrol spacer to seal it off.

This is a large block. The four 'big' holes are mounting holes for LGA775 (see pic 1).

Should I go ahead and replicate the same deisgn in Cu? Should I remove the outer most track and thus reduce the size of my next block?

Well that design is not very optimal. What kind of equipment to you have access to to machine with? If you can go down a couple size end mills I have a design in mind that would require 2-3 more circle passes and several strait cross passes. It would take more work but should work a lot better.

But if you are set on that then yes, do it in copper. Not just for performance but reliability. When aluminum corrodes is gets eaten. When copper corrodes it makes a nice protective layer that significantly reduces failure time.

Good job on the block. Nice machine work.

[ibmkg]

I am doing this on a lathe machine. What is wrong with this design, less surface area (contact area)? Or..? How can I improve this design?

Actually I am not designing this myself. Someone is doing me a favor and he suggested this design (he said this is easy for him to do).

[jaydee]

Originally Posted by ibmkg

I am doing this on a lathe machine. What is wrong with this design, less surface area (contact area)? Or..? How can I improve this design?

Actually I am not designing this myself. Someone is doing me a favor and he suggested this design (he said this is easy for him to do).

Yeah more surface area. Add more fins.I was thinking add pins in the channels but if he just has a lathe that might take a lot of time.