The Truth Behind Lapping?! Which method is right..

afropuff

Limp Gawd
Joined
Oct 1, 2002
Messages
448
Hello all,

I was just wondering what is the standard for lapping a HS? I have just got done reading a few guides on lapping and they all say different things. One says you should get it to a mirror finish using high grit, others say that mirror finishes can actually make things worse. Some say to use figure 8 strokes, others say to only go over it in one direction while others say go both directions. Some say to apply pressure while others say to apply no pressure at all and let the weight of the HS do the work.

All of these "guides" took place between the last three years. What I want to know if there has become a standard for lapping. Is there a right way and wrong way to do it? If so, what is it? Has there been some conclusions as to which method(s) are best for lapping?

What is the deal with lapping? -Jerry Sienfield

-fro
 
Figure 8, and as flat as possible. If flat as possible means mirror finish - then so be it.
 
Well, doing it in figure 8's is what causes the base to be thicker in the middle and thinner on the edges because you are applying the pressure at different angles, but it is the faster way to get a lapped base. As far as mirror finish is concerned, I really doubt that it makes that much of a difference. I think where lapping might help is when you can see the machined lines or pits on a heatsink. Lapping it to 800 grit may help in that case. All of this talk about lapping, and to gain only a degree or two....too much work IMO :p. Better cooling devices yield more meaningful results :p.
 
Yeah, just start out with a great HSF and don't worry about lapping it. Most heatsinks nowadays come perfectly flat and with a mirror finish anyhow. No need for lapping heatsinks anymore. Now lapping northbridges and GPU's, that's an entirely different story! :)
 
Rix2357 said:
Well, doing it in figure 8's is what causes the base to be thicker in the middle and thinner on the edges because you are applying the pressure at different angles, but it is the faster way to get a lapped base. As far as mirror finish is concerned, I really doubt that it makes that much of a difference. I think where lapping might help is when you can see the machined lines or pits on a heatsink. Lapping it to 800 grit may help in that case. All of this talk about lapping, and to gain only a degree or two....too much work IMO :p. Better cooling devices yield more meaningful results :p.

I would agree, but for a stock intel or stock amd HS, lapping may be worth it in the end??

Reason I ask is because I have never lapped a HS before and was wanting to know if it is worth it and how it is "properly" done.
 
my 2.4c's stock heat sink was horrible.. lapped it before first boot up... now i know why intel puts thermal pads that are soo thick on them
 
Rix2357 said:
Well, doing it in figure 8's is what causes the base to be thicker in the middle and thinner on the edges because you are applying the pressure at different angles, but it is the faster way to get a lapped base.

Then we agree to disagree. Done properly, this is the best way IMO to lap a sink, barring having it re-machined.
Anyone who grabs a sink in their fist and starts grinding it into some sandpaper is obviously going to run into troubles.
A nice square weight sitting on top of the sink durring lapping will distribute the weight evenly and give you a nicely flat surface.

Do it right...do it once.
 
xplicit said:
Then we agree to disagree. Done properly, this is the best way IMO to lap a sink, barring having it re-machined.
Anyone who grabs a sink in their fist and starts grinding it into some sandpaper is obviously going to run into troubles.
A nice square weight sitting on top of the sink durring lapping will distribute the weight evenly and give you a nicely flat surface.

Do it right...do it once.

Do you only sand in one direction or do you go back and forth?

Also, does anyone have a link to a good guide on lapping that is accurate and good - I know there are so many guides out there.

thanks
 
afropuff said:
I would agree, but for a stock intel or stock amd HS, lapping may be worth it in the end??

Reason I ask is because I have never lapped a HS before and was wanting to know if it is worth it and how it is "properly" done.

I lapped the HSF that came with my Athlon XP 1700+. I didn't try it before lapping but I think it was worth it.. running at 2200+ I hit 40 C under full load and that's with case fans off..
 
I really don't understand this whole lapping thing...
Is it done for just a flat surface?
I mean, if you have a flat surface but it has that brushed metal texture to it, no mirror finish, there's no need for lapping, right? There's still the thermal paste... which conducts almost as well as the metal itself, in most cases...
 
krizzle said:
I really don't understand this whole lapping thing...
Is it done for just a flat surface?
I mean, if you have a flat surface but it has that brushed metal texture to it, no mirror finish, there's no need for lapping, right? There's still the thermal paste... which conducts almost as well as the metal itself, in most cases...

Flatness is relative, it can be measured across any length scale. A surface can be "flat" with ridges, or it can be "flat" with some curvature that is minimal within your measurement. What we want is a good mate between the heatsink base and CPU core, so we want it flat on a scale similar to that of the CPU core. Getting rid of ridges is a plus, that makes less space for thermal paste to occupy (which by the way, is nowhere near as good a conductor as metal). Probably more important is to eliminate curvature to a extent that allows the core to be in (near) complete contact with the heatsink. Without that, a mirror finish is useless.
 
A mirror finish is almost only useful to detect curvature. Finely machined metal usually gives a mirror finish, so usually a mirror finish tends to signify flatness. Just don't use any of that polishing crap on your heatsink like some people do.
 
There is flatness and then there is "flatness". A mirror finish doesn't really have anything to do with flatness other than sometimes when getting a surface flat the resultant surface also is quite reflective. Flatness is relative and I think for our purpose visual inspecting for flatness by placing a razer blade across the suspect surface under light is probably a good way to determine if you might need to lap.
 
If I were to lap a heatsink by hand and wanted to acheive as much perfection as possible, I would put a weight on the heatsink (if not heavy enough) then lap only in one direction making sure to set the heatsink firmly down before moving it across the sand paper. From my own personal experience, I never got more than a 2C drop in temperatures lapping to 800 grit doing 8's or the way I detailed above. What I do know is that it takes a good day to do a good job of lapping the heatsink for an Alpha. Alpha heatsinks aren't exactly smooth, but they aren't terrible either. I would put it on par with a stock intel heatsink in regards to the level of flatness if you only look at the machine lines.

Consider this, you work someplace and get minimum wage worst case. You get about $5 an hour maybe after taxes. Working 2 half days (8 horus) yields you $40 to buy a nice heatsink. Hence, why I feel that lapping doesn't net you much gains, and that investing in a better cooler is the way to go.
 
Are you guys talking about lapping or sanding? They are completely different processes.

lapping: Smoothing, finishing, or achieving an extremely close tolerance (e.g., 0.000002 inch in gauge block manufacture) on metal surface by means of a LAP I that is charged with LAPPING COMPOUND and rubbed against workpiece in rotary and reciprocating motion, either by hand or machine.

sanding: Rubbing sandpaper or similar abrasive over a surface before applying a finish
 
Back
Top