Right & Wrong Lapping

[psychofunk]

Hey what is the deal with lapping. I have read do it with 600, 1000, 1500, 2000 so this is how I did it for the last year or so. It usually results in something extremely flat and very very shiney, I can watch TV off of the reflection.

I recently had someone tell me that this kind of lapping would make my temps worse. That I should never go beyond 600 as the surface becomes so smooth that there is no where for the AS3 to bond to. He is a pretty extreme oc'er too so I was kind of taken aback, but I still want a few more opinions. No need to flame I just want some facts and opinions based on real experiences not opinions. Thanks again for putin up with my newb questions.

[Alchemy]

So long as you're doing it right, the smoother your surface is, the better.

TIMs are there to fill in the tiny gaps in the surfaces so that there is extremely high thermal contact. The fewer and smaller these gaps are, the less TIM you need.

The only way you could make things worse is to lap the block to a very high grit and then use too much TIM. Even this shouldn't give you performance worse than a poorly lapped block.

Lapping over 600 grit or so will probably do next to nothing for your temps (unless you're such an expert at mounting that such a tiny difference matters), but it won't hurt them.

Alchemy

[psychofunk]

I have a large thick...um...glass table and I use that as my lapping surface. So the fact that it is so shiney is not a problem it is just not necessary for optimal performance? Thanks Alc.

[cybrsamurai]

when lapping the flater the better. If you had two perfectly flat surfaces you could put them together with no thermal paste and have the best possible thermal transfer but as we dont really have the technology to lap metal to fit together perfectly nor do we have perfect metal TIM's are good and shiney is good too! Although you probably wont notice a temp difference between a smooth piece of metal and a smooth shiney piece of metal.


just my $.02

[Alchemy]

Originally posted by psychofunk
I have a large thick...um...glass table and I use that as my lapping surface. So the fact that it is so shiney is not a problem it is just not necessary for optimal performance? Thanks Alc.

The shine helps, but only if you're sure it's a flat shine. I use a sturdy wooden desk, but any flat surface will work.

If you're using a glass table (or anything), I'd lay something across it, like a level, to be sure it's flat and won't bend when you're working on it. But yeah, a firm, flat piece of glass is optimal for a working surface.

Anyway, I've lapped up to 600 grit and gotten a mirror shine. I can't imagine it looking any better, so a 2000 shine must be either the same thing or shinier than God's ass. Or something.

Anyway, I noticed a significant temp difference when I took out a WB that looked sort of rough and brought it to 600 grit, so I'm fairly certain at least that much helps. I'd expect a law of diminishing returns on this line, so I personally wouldn't lap any further. If you want to, go for it, and let us know how it worked out.

Alchemy

[Phuzun]

so a 2000 shine must be either the same thing or shinier than God's ass.

Thats a pretty good comment.

[BillA]

try the search function on this forum
lots of good info has been posted
none of which was repeated in this thread

is that perfectly clear ?

[CheeseBall]

I read bigben's thread on lapping, and someone brought up rotating your block to get the thermal compound worked in. I thought you were suppose to move the block around as little as possible? Dosn't wetting the block w/ thermal compound accomplish filling most of the micro canyons? And how can you rotate your block w/ the threaded rods sticking out of the mobo? And if you put the block on, rotate it a time or 2, then try and square/center it up and run your rods through, aren't you just asking for trouble?

I am putting together my first WC rig, block is just about lapped (I ran outta 600 grit sandpaper).

So basically what do you think of rotating your block when mounting it to "work in" thermal paste?

[JSimmons]

You should actually apply the compound to the block and rub it into the surface of the block. Then, apply a thin film of the compound to the cpu core.

[CheeseBall]

Yes I know this, I refered to that as "wetting the block". Is that not what peeps call it? This is what I have always done, but someone mentioned rotating it while on the die just before clamping it down. I am wondering what everyone thinks of this.

[bigben2k]

Rub it, spin it, get down and Boogie with it... whatever works!

A lot of these compounds come in a somewhat liquid state in order to facilitate the penetration into the crevices, but that's just temporary: the compounds quickly loose their fluid part, and settle in nicely. Many report this as "curing time" and lasting up to 3 days.

What's more important is making sure that the surface is lapped and cleaned. Nail polish remover is *BAD*, because it contains oils, so straight acetone is better. A cleaning alcohol should work well too, but still watch out for that oil content.

[Ice Czar]

Damm it took awhile to find this link

This is the article over at overclockers that started this debate
its rather convincing

"Why Heatsink Polishing Might Be A Bad Idea"
by Simon Watkins - 5/14/01


Flat is important, 2000 grit polish isnt