How I Lapped my Heat Sink

Add Your Comments

DIY – Jeff G

I lapped my 9800 (stock) HS the other day just because I can – that and I’m picking up either an Artic Silencer or a Zalman VF700-CU in a couple of days (heck, maybe I’ll get both and write a comparison review). Anyhow, I thought I would share with you (pictures also) how I lapped my 9800 HS for those that don’t know how and even maybe give some new techniques to some that already do.

I would like to start of by saying

“I am not responsible nor is Overclockers.com/Forums for any damage that you may inflict upon your heat sink (HS). These are just guidelines on how (I) lapped my heat sink (HS) and should be taken as such (guidelines). What you do to your heat sink (HS) is your own responsibility.”

First off, get yourself the necessary stuff, most of which can be purchased at an auto parts store:

  • A flat surface – glass is best
  • 400, 600, 1000, 2000 Wet/Dry sandpaper
  • Masking tape
  • Jug of water
  • A sponge
  • 100% Cotton Terry cloths (important for secret ingredient)
  • Secret Ingredient (talked about later on)

I started off by taping a piece of 400 grit sandpaper to the flat surface (sticks well to glass), of a size at least 3″ bigger than the heat sink. I then proceeded to soak the paper with the water using a sponge – works good to soak up any spills also.

When the paper was good and wet, I started sanding the heat sink, at first using circular motions (clockwise and counter clockwise), but found it did not leave a desired finish. I then started sanding in one direction back and forth a dozen times, then turned the heat sink 90 degrees and proceeded to sand the same amount of times again.

Here is what it looked like after the first few sanding with the 400 grit.

Lap

As you can see, the heat sink started to show sanding marks on the edges first, meaning this heat sink was not all that flat.

After I sanded off all the black paint (damn stock HS), I changed to the 600 grit, again taping it to the flat surface and soaking it with water.
Here you can see in the center and just above it that the heat sink is still not flat; not to worry, as the 600 and 1000 grits of sandpaper will get this out. This one is after the 600 grit.

Lap

Again the same procedure for the 1000 grit…

Lap

Now here’s the point where I did something a little extra (Special Ingredient)…

After I used the 2000 grit for awhile, I was not satisfied with the finish (don’t get me wrong – it was good, I’m just picky). So what I did was used a product (2 actually) called Meguiar’s – it’s a high grade professional car finish product (I used them for years to give a mirror finish).

The first one I used is called Final-Cut Cleaner; I applied it with the cloth as per instructions on the bottle and then finished it off with the other one called Swirl Remover, again as per the instructions. I then cleaned off the heat sink with Isopropyl Rubbing Alcohol with a separate cloth. Making sure the heat sink is fully cleaned (until no more color appears on the cloth.) Allow for dry time before reinstalling.

The results, I’d say are outstanding.

Lap

I hope this helps even just one person. Good luck & Happy lapping!

Cheers

Goli

Leave a Reply

Your email address will not be published. Required fields are marked *

Discussion

  1. - What sort of thermal properties (negative or positive) would these two products have?

    Actually I would think that anything like that would impare the thermal transfer why? Well think about it, its a chemical thats creates a a finish thats resilliant to rain/sun.
    Well I can't put it into words now for some reason but you get the point.
    Wtf. This article seems really dumb and "duh" type to be posted on overclockers.com. Gah, I guess times have changed.
    Yes, shiny does NOT mean thermally efficient. Many polishing products just add an agent to make it look shiny in the little grits of space in the metal. These may be insulators and hinder efficiency. Just seems way too obvious of an article to be posted, and done in a weird manner.
    Buy the heatsink kits as posted in cyber-deals. Heatsink lapping kits.. I forget thier name, insulglass sells them. I've reviewed them and they're top notch.
    This article seems really to be posted on overclockers.com. Gah, I guess times have changed.

    Yeah, I think I'd agree with that. Feels like they're scraping the bottom of the barrel for subject matter these days. It's too bad. How long has it been since we've seen some good waterblock comparisons, radiator comparisons, etc.?
    Like everyone else has said, that review seems to be more about polishing a heatsink than lapping it. Also, is the bottom anodized? I don't think they'd actually paint it...
    There aren't any quantitative results either. *sigh*
    Yeah, I've never posted on a frontpage article before, but this one warrants it.
    Complete mis-information for obvious reasons people have noted earlier.
    The worst part is the guy's conclusion on the edge of the heatsink being worn away first:

    He assumes it's that the base isn't flat. WRONG! It's entirely what happens when you hand lap because hand lapping always makes a heatsink rounded.
    AGH!!! :bang head
    Seriously, I think the article should be pulled. Certainly not worthy of being on this site.
    If I had to polish a HS base, I'd probably use something made for it. (Or, at least, made to go on it!) AS Alumina, and AS Ceramique are both fine grained, and abrasive.
    They also have the benefit of not clogging the pores with anything that would harm heat transfer.
    I will agree, "shiny does NOT equal FLAT.
    steve
    Note the title to the thread: How(I)lapped my sink, not(you'r)sink
    I'm not forcing anyone to use my methods.
    As far as the "shiny doesn't mean flat", then tell me why is it that when I placed a flat edge on the sink the center was badly concaved and after almost 2 hrs of sanding the sink and using Meguiars for a final "CUT" polish....hmmm get it "CUT" polish, that it had raised my clocks by 5+ Mhz.
    Oh and 9mmCensor, the next time you quote me, please don't edit my post.
    I then cleaned off the heat sink with Isopropyl Rubbing Alcohol with a separate cloth. Making sure the heat sink is fully cleaned (until no more color appears on the cloth.) Allow for dry time before reinstalling.

    I clearly stated cleaning the heat sink after using the Meguiars product, so as far as contaminents go, there are none. The "cut" polish is used to get rid of fine scratches, IE sanding marks or machine grooves, the swirl remover is a follow up. Like I said, it's they way "I" lapped my sink,they are just "guidelines" and should be taken as such.
    Sorry you don't like it but my VGA card is just lovin it.
    Goli
    When you said the HS was stock before you modded, was that stock with stock thermal pad, and after AS5?
    Did you try removing and re-mounting the heatsink before and after the mod? (To eliminate the poor mounting variable).
    Did you test for flatness? Using even Cathar's DIY fine mesh meathod.
    Yeah, I've never posted on a frontpage article before, but this one warrants it.
    Complete mis-information for obvious reasons people have noted earlier.
    The worst part is the guy's conclusion on the edge of the heatsink being worn away first:

    He assumes it's that the base isn't flat. WRONG! It's entirely what happens when you hand lap because hand lapping always makes a heatsink rounded.
    AGH!!! :bang head
    Seriously, I think the article should be pulled. Certainly not worthy of being on this site.

    This is actually a perfect example of why you should not rotate the heatsink when lapping. When you sand anything, the front edge always tends to bite the hardest, which means that the front edge gets the most sanding. If you rotate, that means that each edge will wear like this and what you get is a convex base, which may pose problems for getting optimal core contact. After all, that is the intent of lapping - optimal core contact. I think lapping is over rated really, but if you are going to do it, you ought to do it right.
    A straight lift repeat method might have produced better results than what this person is seeing now.
    Note the title to the thread: How(I)lapped my sink, not(you'r)sink
    I'm not forcing anyone to use my methods.
    As far as the "shiny doesn't mean flat", then tell me why is it that when I placed a flat edge on the sink the center was badly concaved and after almost 2 hrs of sanding the sink and using Meguiars for a final "CUT" polish....hmmm get it "CUT" polish, that it had raised my clocks by 5+ Mhz.
    Oh and 9mmCensor, the next time you quote me, please don't edit my post.
    I clearly stated cleaning the heat sink after using the Meguiars product, so as far as contaminents go, there are none. The "cut" polish is used to get rid of fine scratches, IE sanding marks or machine grooves, the swirl remover is a follow up. Like I said, it's they way "I" lapped my sink,they are just "guidelines" and should be taken as such.
    Sorry you don't like it but my VGA card is just lovin it.
    Goli

    You need to chill out my friend. If you didn't want people to criticize your article, you should have either not submitted it which probably would have been the best solution, or you should have researched better so that you would have been aware and could have addressed the issues that people were obviously going to bring up like what has been discussed in this thread.
    As for your comments about the base being concave... I've never seen a heatsink base which was concave, and at a very basic level, I have a hard time seeing how the methods commonly used to cut the stock to create the base of the heatsink could even make the base concave, as a matter of basic physics. It was your lapping method.
    eeek, well i must contradict one thing. An HS that is shiney without the use of chemical additives, is flatter than one that isnt.

    Its not exactly true that a shiny heatsink is flatter than one that is not.
    eeek, well i must contradict one thing. An HS that is shiney without the use of chemical additives, is flatter than one that isnt.

    Like Censor alluded to, shiny and flat have nothing to do with each other.
    The inside of a spoon is nice and shiny. However, it isn't what I would define as flat.
    And, what do you mean by "chemical additives?" everything we use, except possibly electrons (or software), is chemical.
    steve
    And, what do you mean by "chemical additives?" everything we use, except possibly electrons (or software), is chemical.

    He is refering to the chemical polishers, like those used in the article.
    I couldn't resist jumping in. By all means, everyone consider contributing an article. It could get you a nice gift. It does not have to be earth shattering news. Lastly, a block only has to be flat in the small area that contacts the core or IHS. The rest of the face can look like a pizza pie for all that matters. Another misconception is having the entire face parallel to the top. It can have some tilt without it impacting the mounting effectiveness. That's why you can get pretty good performance using (carefully) a belt sander with different grit belts. The block face may be thinner on the leading end, but you can get darn good flatness since the sanding action is monodirectional. I use a table top 4 inch belt sander for my blocks. The critical component is the stop bar across the belt so you don't have to fight the block sailing off across the room. That leaves your attention to keeping the pressure even and stabilizing it laterally. Unfortunately, it's hard to find belts finer than 120 grit at the local hardware store, but they can be had down to 1000 grit from online suppliers.
    Just my own personal experience.
    Hoot
    As for your comments about the base being concave... I've never seen a heatsink base which was concave, and at a very basic level, I have a hard time seeing how the methods commonly used to cut the stock to create the base of the heatsink could even make the base concave, as a matter of basic physics. It was your lapping method.

    I'll have to disagree with you on that. The base on my XP-120 was concave. From what I've read, it comes from soldering on the heat pipes at the factory after the base has been machined. It took about 2 hours with 120 grit sand paper to get it flat, but when I was done, I could put a small drop of water in the center and it would spread all the way to the edges when I pressed a glass on top. Also, it looks straight when viewing the image of graph paper on it.
    CJ
    I never truly tested my xp90 after my lapping job. I probably should have checked it with the glass too, or graph paper... stupid me. Anyways, I don't have much of a so-called mirror finish, but my temps are outstanding now. They dropped by like 10C after my AC5 burned in almost perfectly.
    Like Censor alluded to, shiny and flat have nothing to do with each other.
    The inside of a spoon is nice and shiny. However, it isn't what I would define as flat.
    And, what do you mean by "chemical additives?" everything we use, except possibly electrons (or software), is chemical.
    steve

    ok if you don't use any polish or any type of liquid chemical added to the HS shiny will usually mean flatter. Copper will never look shiny unless you polish it, however when lapped it does begin to show some shine. A spoon is not made of neither copper or aluminum (which most HS are). It's made from stainless steel or silver, which is polished (in the silver's case). Therefore im not saying that a shine is what its all about, but if you grab 2 identical HS's and lap one not the other, the lapped one will appear more shiny than the other, the reason is that it is flatter.
    What im saying is, if you're only using sandpaper, and you see more "shine" out of your HS, its because it is flatter. Just so that you all understand that this will be without the help of any polishing agents, or other metals such as stainless steel.