What methods do you use to lap your stuff?

[hardnrg]

Hmm! Very interesting!

 

The edges of the heatsinks I've lapped are rounded to be honest, but it's only *right* at the edge... I usually lap heatsinks on a mirror in a shower cubicle or on a draining board so I can rinse the paper frequently.

 

I've heard about the figure-8 method before, but had not heard anything one way or the other about which method was better or worse, so it's good to hear this from a machinist!

[kendellrt]

I would like to add something here.

 

Many years ago I worked as a machinist in a shop that made parts for communication sattelites. A lot of the parts required lapping to achieve a maximum of 0.0005" (yes, ten thousandths of an inch on parts up to 2 feet square) flatness per inch. A to and fro (or back and forth if you preffer) motion would yield horrendous results. The leading and trailing edges of the lapped surface would get rounded because of the build up of loose particles of abrasives and material that were being lapped and accumulating at the edges.

 

The best results were always achieved using a figure "8" motion while applying only the necessary force to keep the part in complete contact with the lapping surface or medium (paper or compound on a granite surface). The preffered lubricant that we used was mineral spirits (oil based paint thinner). If you're using sand paper, a mix of dishwashing soap (yeah palmolive or the likes) should tield satisfactory results.

 

Hope this can help.

I have heard others recommend a figure eight motion also. How long would you do it in a figure eight? I assume that you would go clockwise and counter-clockwise with each grit.

[hardnrg]

a figure-8 *is* clockwise and counter-clockwise though

[kendellrt]

a figure-8 *is* clockwise and counter-clockwise though

Ok, one way and then the other. Whatever!

[mazakman]

I have heard others recommend a figure eight motion also. How long would you do it in a figure eight? I assume that you would go clockwise and counter-clockwise with each grit.

 

Yes a figure 8 motion for all grit sizes and remember to apply only the force needed to keep the part flat on the lapping surface.

 

I'd suggest a trial run on something of comparable size and material just to get the hang of it.

A precision ground bevelled straight edge can be handy to verify the flatness of the surface to see if there is any light coming through between the lapped surface and the straight edge, but if done properly the straight edge isn't needed.

[Verran]

That's some good advice, mazak. Like nrg, I have also noticed that when I lap a sink the very edge of the sink usually gets a bit rounded. That never really concerned me since that part of the sink never actually contacts the chip, but it'd still be nice to get it completely flat. I'll have to give the "figure 8" approach a go the next time I'm lapping

[kendellrt]

Yes a figure 8 motion for all grit sizes and remember to apply only the force needed to keep the part flat on the lapping surface.

 

I'd suggest a trial run on something of comparable size and material just to get the hang of it.

A precision ground bevelled straight edge can be handy to verify the flatness of the surface to see if there is any light coming through between the lapped surface and the straight edge, but if done properly the straight edge isn't needed.

How much time do you dedicate to each grit?

[mazakman]

That's some good advice, mazak. Like nrg, I have also noticed that when I lap a sink the very edge of the sink usually gets a bit rounded. That never really concerned me since that part of the sink never actually contacts the chip, but it'd still be nice to get it completely flat. I'll have to give the "figure 8" approach a go the next time I'm lapping

 

Let's not get carried away here, completely flat is impossible, we're talking micro inches. and as you mentionned the center of the heat sink is where it's most important in this case but you can always brag that it's flat to the edge.

 

chavalcito, it basicaly depends on the material and what state it's in so all I can say is as long as needed.

 

I'd say check the surface after every few minutes of lapping at first and then at shorter intervals.

Once there is no noticeable difference then you can move onto the next grit. Remember to clean the part thoroughly before inspecting. Proper lighting is a must and a magnifying glass can help inspection to see fine scratches (a stereoscope with 7 times magnification would be great but not everyone has access to such equipment).

 

One thing I forgot to mention is make sure there is no dust. It will make a huge difference once you get into the finer grits, trust me on this one.

 

I hope you're patient.

 

Have fun!

[road-runner]

I have not lapped a cpu yet myself but I did take a razor blade and put the edge to my new cpu to see how flat it was across and it looked flat to me. I held it up to the light and could not see any daylight in the center or on the edges. I used like it was to not void the warranty...

[graysky]

@road-runner: the only way to know for sure if your CPU is flat is unfortunately to lap it. You'll know after about 20 seconds because you'll start to see the material come off the higher parts first like this:

 

 

I started out w/ my Q6600 using 800 grit thinking, "Hell, I'll just polish the surface a bit just to see how uneven it really is." After I saw how bad it was, I had to finish the job

 

 

Best $20 I spend in terms of cooling band-for-the-buck.

 

Link to the rest of the thread with more pics/temp results.

Edited June 29, 2007 by graysky

[kendellrt]

Why did you stop at 1000 grit?

[kendellrt]

I found this article today. I thought it was a good basic guide, and I like the idea of using a marker to see how even the surface is. The definitely are against using circular motions, though.