Lapping Question for Modern Blocks

D

DarkenReaper57

2[H]4U
Joined
Oct 2, 2003
Messages
2,173
Hello,

I have a new D-tek Fuzion that I want to install. I've always lapped my blocks and placed them on a bare core (removed IHS), but now I have a Q6600. If I've read up correctly, the IHS can't practically be removed on these chips. Bare cores are almost always flat, but IHS's are often concave. As a result, the new Fuzion blocks are factory bowed, correct?

I want to get the best heat transfer possible, so I thought the most logical thing to do would be to lap both my Fuzion (while assembled) and the Q6600. Am I wrong here? I'm not as up to date with modern equipment, so I want to ask before I brick something.

Also, I have a MCW60 for my GPU. Are these blocked bowed at the factory now as well (I don't see why they would do this for bare cores, but wonders never cease). I was thinking of lapping this as well.

Thanks in advance!
 
Erasmus354 said:
Don't lap the block, all you can do is mess it up.
Click to expand...

Can't you get a much better contact between the IHS and the block if they are both assured to be flat by lapping? The curvature of a bowed block and IHS are two unknowns, and might not bring excellent contact. Lapping, if done correctly, assures that both the mating surfaces are flat.

Is there something wrong with my logic here? I think I remember reading a study here where someone compared the following situations: unlapped CPU/unlapped block, unlapped CPU/lapped block, lapped CPU/lapped block. Lapping both produced the lowest temperatures.

Then again, I read this about a year ago. Do people just not lap their blocks anymore? By the way, I have the original Fuzion block, not the second version.
 
People rarely if ever lap waterblocks. Some HSFs, sure, but waterblocks pretty much never.

Also, no do not attempt to de-lid the Q6600 -- you'll kill it. Pure and simple. You cannot de-lid any CD2/C2Q chip. The IHS is bonded to the core, and will destroy it if pryed off.

I suppose you could lap the CPU and block, but with water, why bother? 3C either way isn't going to matter, and that's the most difference you're likely to see. The IHS should already be fairly flat, as should the block's mating surface.
 
I have a new D-tek Fuzion that I want to install. I've always lapped my blocks and placed them on a bare core (removed IHS), but now I have a Q6600. If I've read up correctly, the IHS can't practically be removed on these chips. Bare cores are almost always flat, but IHS's are often concave. As a result, the new Fuzion blocks are factory bowed, correct?
Click to expand...

It is my understanding from a visit to their web site some months ago they sell both the "bowed" and "flat" versions. I recommend you go check. A check of your block with a single edge razor blade would help you determine if you block is flat or not.

Also, I have a MCW60 for my GPU. Are these blocked bowed at the factory now as well
Click to expand...

No, they are decently flat but I lapped mine because, just because.

I want to get the best heat transfer possible, so I thought the most logical thing to do would be to lap both my Fuzion (while assembled) and the Q6600. Am I wrong here? I'm not as up to date with modern equipment, so I want to ask before I brick something.
Click to expand...

Its not going to matter a lot but if it were me, and keeping in mind I am crazy with more time than I know what to do with, I would access the situation. Check the flatness of the IHS and the Block, and make a rational decision on what to do.

IHS flat, block bowed, it would depend on how much but I seriously doubt you could remove enough material from the block to damage it. It would be a lot of sanding.

IHS flat, block flat, lightly lap em both ( and see below)

IHS not flat, block bowed. Well is the IHS concave (bowed out in middle) or convex (bowed inward in middle) or something else. This would require you to make a decision as to what you wanted to do.

IHS not flat, block flat, lap the IHS.

If you really want to be insane, Lapping actually means to grind two surfaces to match one another, once you get them close with normal lapping techniques, rough jewelers rouge (despite the "rough" designation it is a very fine polishing compoundP applied to the IHS and about 2 full lengths movies worth of rubbing the IHS against the block where it will be mounted (centered in the water-block) using small movements would help match the two components exactly.

In short I recommend you access what you have, develop a plan, and go for it, unless you use an 80 grit in a belt sander you are not going to damage anything if you use some care. The CPUs now without the pins are much less likely to be damaged but don't scuff around on carpet or wear a fuzzy sweater etc. when handling it. Clean it well with alcohol when done, try not to drop it on a hard floor etc. I did mine and got it plenty wet, its sealed and a good drying in a sunny window after cleaning will do the trick.

Unless they are way out of wack you are not going to see much difference improvment but, as I mentioned, I am insane, I lap ramsinks and northbridge/MCH heatsinks etc. !!
 
I wonder if lapping to a mirror finish will actually harm thermal performance ...

IF the surfaces were really flat and shiney, I noticed that thermal paste doesn't really "stick" on, so it's almost like two bare surfaces touching, which I wonder will really give better performance, than say two surfaces that are flat but "slightly" rough to allow thermal paste to sit between the srufaces.
 
chanchan said:
I wonder if lapping to a mirror finish will actually harm thermal performance ...

IF the surfaces were really flat and shiney, I noticed that thermal paste doesn't really "stick" on, so it's almost like two bare surfaces touching, which I wonder will really give better performance, than say two surfaces that are flat but "slightly" rough to allow thermal paste to sit between the srufaces.
Click to expand...

Was any sort of rubbing compound used to make the surface shiny? Many rubbing compounds repel water, and it may do the same to thermal compound. This is why I've never used any sort of rubbing compound after lapping - it can actually make thermal conductivity worse. Yeah, you can clean up with alcohol afterwards, but that won't necessarily pick up 100% of the polish, especially the microscopic bits that have worked their way in (where the thermal paste should be).

BillParrish, thanks for the insight. I've lapped blocks and heatsinks many times before, but I the last block I used was the d-tek whitewater. This is a rather outdated block, and I'm trying to catch up on the current technology. Back in those days, everyone lapped their waterblocks, and I see that fewer people do now. I was just trying to understand why.

Anyway, that aside, I'll check the block and IHS for flatness using a razer. Unfortunately I don't own calipers, so a flat surface will have to do.

On another note, I remember reading that the Fuzion v2 offers up to 5C improvement. The problem is the study only tested with 1 mount, so you can't really say the results are accurate (you need at least 5 mounts). Does anyone know if the v2 offers significant improvement over the v1 with a quad-core system and a D5 pump?

Oh, one more question: right now, I don't have the pro-mount. Is it worth it? I read that the mounting is much more secure. The v2 includes the pro-mount, which is why I am thinking of selling my unused v1 and buying a v2.
 
Yes, the pro mount is without question much more secure and distributes pressure much more equally.
 
silent-circuit said:
People rarely if ever lap waterblocks. Some HSFs, sure, but waterblocks pretty much never.

Also, no do not attempt to de-lid the Q6600 -- you'll kill it. Pure and simple. You cannot de-lid any CD2/C2Q chip. The IHS is bonded to the core, and will destroy it if pryed off.

I suppose you could lap the CPU and block, but with water, why bother? 3C either way isn't going to matter, and that's the most difference you're likely to see. The IHS should already be fairly flat, as should the block's mating surface.
Click to expand...

Wow, I didn't know that. Thanks for the info. :)
 
You must log in or register to reply here.
Back
Top