PDA

View Full Version : Who's Lapping Copper...what did you use?



WillysNut
05-15-01, 08:45 AM
Does lapping copper require any different procedures/grit size as compared to lapping alumnium? I've been running my CAK38 for about a week and I took a look at the top of my chip last night.

I can see small lines of perfect cpu/sink contact (almost no AC2) and small lines of AC2. Basically, the high and low grooves of the machined surface contact on the processor die.

I believe lapping is def. in order for the CAK38. I checked some guides on lapping but they all apply to Alum. as copper is not as established. What did you all use...or should I just use the same grit sizes as Alumnium?

Mac42
05-15-01, 08:52 AM
Same stuff, though it might take slightly longer.

-=mac=-

LimeyGreg
05-15-01, 09:07 AM
I agree, same proceedure is fine. I did read that the prefered compound for lapping soft metals is garnet.

Read the article titled "Heatsink Too Flat" on the front of this site, I thought it was quite interesting. Basically it says that you don't want a mirror finish.

Hoot
05-15-01, 12:02 PM
I read that article yesterday and while I respect results over theory, it doesn't seem right in my minds eye. I sure wish I had two "virgin", identical HSFs to test and do an article on. One with a flat, matte finish and one with a flat, mirror finish. It would be an excellent test of the diminishing returns theory.

Hoot

ken257
05-15-01, 12:12 PM
Here is my theory on the to flat article. If you had a large contact area say 2 square feet the extra grease would have a hard time making it's way to the edges and oozing out so the course surface would give the grease a place to go. Cpu's have a small contact area and extra grease has no trouble escaping so flat is better for us.

WillysNut
05-15-01, 12:39 PM
I'll give it a shot tonight or tom. I'm going to review the guides again and test temps. in 3 steps.

1) Stock conditions (no lapping)
2) General lap (not mirror finish)
3) ...wait for some more input and decide if
I should go the next level (mirror finish)

Thanks for the replys....any more ideas on the "not to totally flat" lapping?

Phil
05-15-01, 03:33 PM
copper is soften than aluminium but does take longer to lap, you may need to change the paper more often as when I have lapped copper objects I've found the paper has become clogged up with copper quicker than with aluminium. As for how flat you want and how course I would say that beyond about 1500 grit paper you won't see much of a differance, and a mirror finish may look nice but won;t help that much. I have a very smooth slightly matt surface on the bottom of my FOP and using Artic silver this works great, I don't think being mirror smooth would do anything for it.

outhouse
05-15-01, 09:25 PM
I would not recommend using any kind of polishing compound to get a mirror finish [if you go that way] since most polishes have wax and other compounds in it, using them will leave there material in the grooves and pores which could keep your thermal compound from doing its job to its fullest potential resulting in a loss of thermal efficiency

It_The_Cow
05-15-01, 10:05 PM
Hoot (May 15, 2001 12:02 p.m.):
I read that article yesterday and while I respect results over theory, it doesn't seem right in my minds eye. I sure wish I had two "virgin", identical HSFs to test and do an article on. One with a flat, matte finish and one with a flat, mirror finish. It would be an excellent test of the diminishing returns theory.

HootThe author really should have tried not using any thermal compound at all for a good test. According to him, the TC got in the way, so why didn't he just take that equation out?

WillysNut
05-16-01, 06:35 AM
Would not using any compound at all really be safe? Note: I have not lapped my PIII as I'm not ready to go down that alley yet. So...There are still imperfections on half of the cooling equation.

Even with lapping, nothing is ever totally flat so I think a marginal amount of Arctic Silver is still gonna be applied to all my heat sinks. Does anyone agree?

Phil
05-17-01, 02:34 PM
you should always use a thermal compound as very little in the way of surface area of the cpu and heatsink come in to contact (if viewed through a really good microscope) even lapped there are gaps so always use a thermal compund.

It_The_Cow
05-17-01, 04:00 PM
Well let's think about it. The author said that a too flat of a surface would squeeze out all the TC and disrupt heat dissapation. With his resources, he probably made it the heatsinks very, very flat. Since there were few imperfections, the heat transfer would have been fine, if not better, without thermal compound.

KaHNZa
05-18-01, 12:22 PM
My Alpha PAL6035 with the copper inlay had a near mirror-like shine when I got it. It only had small imperfections. I ran my fingernail across it and couldn't feel any grooves or bumps. But I could see some slight swirling marks from machining. So I used just a little bit of AS2.

Phil
05-18-01, 04:55 PM
It_The_Cow (May 17, 2001 04:00 p.m.):
Well let's think about it. The author said that a too flat of a surface would squeeze out all the TC and disrupt heat dissapation. With his resources, he probably made it the heatsinks very, very flat. Since there were few imperfections, the heat transfer would have been fine, if not better, without thermal compound.

I don't care if he did it with a lazer, there are still gaps, at the molecular level, which the paste will fill and it wouldn't squeeze it all out if it is put it on evenly