how would you know if the heatsink that you were lapping became uneven? Could you see it just by looking at it? Is there a certain trick I can use? or do I have to buy those leveling thingies with the water and the air bubble?
questions on lapping a heatsink
- Thread starter ku
- Start date
Jojo posted a very nice trick to tell if you are lapping it evenly, although it will not tell you if it was uneven to begin with. Basically, if you take a sharpie marker and draw a grid on teh bottom of your heatsink, when all the marker was gone, you would know it was flat.
Chiz
Chiz
If you lap with a fairly fine grit (600+) and use a circular or figure 8 motion, you should not have to worry about unevenness. If you're lapping back and forth... may god have mercy on your soul 
Do not lap an object in a circular motion - that tends to bevel the outer edges.
Use an "elongated" figure-8 motion - the emphasis is to remove material on the long, straight strokes - not as you make the "turn" at the end of the stroke. Use "wet or dry" 600 grit paper, and use water or honing oil - it helps the abrasive cut smoother, last longer, and keeps it from loading up, thus helping you to maintain light and even pressure. Patience, rather than pressure is required.
All lapping should be done on a piece of plate glass to ensure a dimensionally flat surface - a bevel-block mirror insert is cheap, flat, and works just fine. Wet the back of the paper slightly before you apply it to the glass to keep it in place.
A marker is a good way to check your initial technique and progress. If you're tilting the object, you'll see evidence of bevelling rather quickly.
Use an "elongated" figure-8 motion - the emphasis is to remove material on the long, straight strokes - not as you make the "turn" at the end of the stroke. Use "wet or dry" 600 grit paper, and use water or honing oil - it helps the abrasive cut smoother, last longer, and keeps it from loading up, thus helping you to maintain light and even pressure. Patience, rather than pressure is required.
All lapping should be done on a piece of plate glass to ensure a dimensionally flat surface - a bevel-block mirror insert is cheap, flat, and works just fine. Wet the back of the paper slightly before you apply it to the glass to keep it in place.
A marker is a good way to check your initial technique and progress. If you're tilting the object, you'll see evidence of bevelling rather quickly.
timber: could you explain your "elongated figure 8" technique more? I think I understand but I just want to make sure. You're saying make 8's using long STRAIGHT strokes?... to create sort of an hour-glass?
<< timber: could you explain your "elongated figure 8" technique more? >>
like a long skinny 8 versus a "8" that is nothing more than to circles put on top of each other
also use a flat surface and make sure your hand keeps the heatsink flat against the surface.
you will see how "uneven" your HS was to start in just a few minutes i use that as my guide
ku:
Sorry for the delay in responding - I dozed off to the Land of Nod.
But MSO4 described it nicely in my "absence". Your "hourglass" pattern would also be a good description. Again, the use of excessive circular motions (along with the impatience that naturally accompanies a tedious and boring task) tends to result in a nicely polished "dome effect", rather than a flat surface.
I've seen the VH article, along with a few similar "How-To's" on the web. IMO - The authors have confused the results of "creating a mirror finish" with the goal of "establishing a flat surface to facilitate an efficient thermal interface". BTW - Experience has taught me that any article that starts with "beer" on the materials list is questionable at best.
Nowhere in the article is a mention of the most fundamental requirement for the task: a FLAT SURFACE to work on. The laminated counter-top that he appears to be working on may not be entirely flat - but a piece of plate glass is. You can't tell just by looking at a surface, just as you can't accurately judge his end results by observing the reflection of an AMD case badge. Depending on the mounting method of your HSF assembly, it's entirely possible to invest multiple hours creating a beautifully polished mirror that still doesn't achieve maximum contact with the CPU slug.
A reasonably good quality 9" x 9" mirror block costs around $5 at many home stores. Generally speaking, any glass that is around 1/4" or more in thickness will be plate glass, which is manufactured with a rolling process that controls it's thickness and flatness; as opposed to the more common "float" process that can produce waves or other dimensional irregularities that are not detectable visually, or with common measuring tools.
Since there is no way to create two perfectly flat mating surfaces (with commonly available tools and materials) that would make the use of thermal compound superfluous, it's important to remember that the final finish can and should have a texture that retains the compound. Which is to say that you can stop when the surface is flat and smooth. I use 400 and 600 grit wet /dry paper and honing oil. A coating of machinist's blue dye, or indelible marker, will help you gauge your progress.
As to the "difference of opinion": I've been a machinist and gunsmith for about 20 years. I drive tractor-trailers now, mainly for the steady income and benefits; which lets me take on projects that fit my schedule, and suit my interests. I build computers for others, and geek 'n' tweak for my own knowledge and amusement. Most of the "differences" seem to be with "geeks" that want to learn machinist's skills.
Sorry for the delay in responding - I dozed off to the Land of Nod.
But MSO4 described it nicely in my "absence". Your "hourglass" pattern would also be a good description. Again, the use of excessive circular motions (along with the impatience that naturally accompanies a tedious and boring task) tends to result in a nicely polished "dome effect", rather than a flat surface.
I've seen the VH article, along with a few similar "How-To's" on the web. IMO - The authors have confused the results of "creating a mirror finish" with the goal of "establishing a flat surface to facilitate an efficient thermal interface". BTW - Experience has taught me that any article that starts with "beer" on the materials list is questionable at best.
Nowhere in the article is a mention of the most fundamental requirement for the task: a FLAT SURFACE to work on. The laminated counter-top that he appears to be working on may not be entirely flat - but a piece of plate glass is. You can't tell just by looking at a surface, just as you can't accurately judge his end results by observing the reflection of an AMD case badge. Depending on the mounting method of your HSF assembly, it's entirely possible to invest multiple hours creating a beautifully polished mirror that still doesn't achieve maximum contact with the CPU slug.
A reasonably good quality 9" x 9" mirror block costs around $5 at many home stores. Generally speaking, any glass that is around 1/4" or more in thickness will be plate glass, which is manufactured with a rolling process that controls it's thickness and flatness; as opposed to the more common "float" process that can produce waves or other dimensional irregularities that are not detectable visually, or with common measuring tools.
Since there is no way to create two perfectly flat mating surfaces (with commonly available tools and materials) that would make the use of thermal compound superfluous, it's important to remember that the final finish can and should have a texture that retains the compound. Which is to say that you can stop when the surface is flat and smooth. I use 400 and 600 grit wet /dry paper and honing oil. A coating of machinist's blue dye, or indelible marker, will help you gauge your progress.
As to the "difference of opinion": I've been a machinist and gunsmith for about 20 years. I drive tractor-trailers now, mainly for the steady income and benefits; which lets me take on projects that fit my schedule, and suit my interests. I build computers for others, and geek 'n' tweak for my own knowledge and amusement. Most of the "differences" seem to be with "geeks" that want to learn machinist's skills.
thanks so much timberwolf. I feel confident now in lapping my heatsink for the most part. I plan to this tommorow and hopefully, it'll be successful =P..
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter