First time lapping..

S

SanityAssassin

Limp Gawd
Joined
Mar 22, 2005
Messages
139
So, I'm getting a new zalman HSF in soon and decided I'm gonna lap it, any suggestions or tips that could be useful? Also, while I'm at it i figured i might lap my AMD X2 4200, this is the part I'm mostly concerned of. I'm somewhat familiar with lapping, however i haven't looked into CPU lapping much at all, figured i should come to the best for that ^_^. Any input on either subject would be great. I ordered a kit already that comes with 200, 400, 600, 800, 1000, grit sandpaper sheets. It also comes with 10, 15, 20, 25, 40, micron paper sheets, the piece of glass and some AS5. (Link)
 
back when this whole craze started, I lapped my p4 1.5 (this was around 2002) and held it up against the chinese instructions and took a picture.. it was featured on the [H] a LONG time ago...

anyways, I used 200, then 600, then 1000... just find a flat surface and go to work...

ain't hard, and you will see obvious flaws right off the bat
 
$13:eek: ....

I used 320grit to start and then 600.... anything higher than that is unnecessary......

At the beginning, I was doing up and down motion turning the cpu every 5-6 strokes, when the ICH is flat, I moved to 600 ans did figure8 motion..... after that i used Q tip and acetone to clean it up..... its pretty easy, took about 30mins or so.......
 
You'll find "how-to's" all over the net the biggest thing I've found is take your time; it will take some time depending on the level of machining marks on your HS. Take breaks while doing it, (if you chew off too much it won't be fixable.) ;)
 
Heh, I don't mind it being a bit excessive, after all since this is the first and last time I'll be doing it to the CPU I figured why not do it to the highest degree. Well, I've read a bunch of how to guides and such, but none for the CPU which is what I'm worried about doing somewhat, mainly cause if I do fuck it up, i wont have a spare CPU to fall back on. I figured i might do my first attempt on a 1.3ghz celeron hat I have laying around. Are there any inherent risks other than bending a pin or going too deep is there anything else that i should worry about with the CPU? Also, how well sealed is the IHS on the CPU? Is there any chance of water getting under it and not being able to get out due to bad ventilation?
 
Which model heatsink by Zalman are you getting? Because from my experience, Zalman's heatsinks already come with a mirror-like finish, so lapping won't do much for the heatsink, although lapping the processor will reduce temps.
 
Got a used CNPS7000B from FSBxtreme's FS page.

I'll see when the HS gets in, if its really almost a mirror I might just do it to as seriously mirrortastic as I can just for shits and giggles.
 
Get a nice camera too..... not mine but from Dani @ XS....

E6600_lapped_polished.JPG
 
mirror finish isn't always the best for heat transfer, other ppl will agree with me soon.
 
elyrly said:
mirror finish isn't always the best for heat transfer, other ppl will agree with me soon.
Click to expand...


i would agree in that a hill-valley ( /\/\/\/\/\/\ ) or grainy ( .-.-.-.-.-.- ) provides more surface area in which to transfer heat. a fine grain silver paste would provide the transfer agent
 
mitcity22 said:
have any before and after tests been done, that would solve the issue
Click to expand...

On the question of surface finish:

Like the AS vs white paste, about 5 years of discussion on the question of surface roughness/finish has been made and every now and then starts up again. The major crux is the size of the particals used in the paste compared to the finish on the mating parts with the idea being that the particles should "snuggle" into the scratches on the two surfaces to be mated. This gives an hypothetical ideal aproximate surface roughness with scratches about the width of the paste particles and a depth of 1/2 particle size. That was the major concensis back then with (of course) not all agreeing. Another school of thought is that with slightly larger scratches the parts with mate with a kind of "interlocking scratches" giving more surface area to transfer heat and these "scratches" should be larger than the particle size.



The trouble was, and still is, while it was fairly easy to create such a surface, one guy had industrial optical roughness measuring equipment that would measure in sub-micons, and gem lapidary equipment is easily adaptable to create the desired surfaces, and despite some really fancy jigs and holding devices, no test setup could be made repeatable or accurate enough to measure any difference/advantage with repeatible results and settle the question in a scentific manner. So it remains theory and debatible. If I may venture, because I am working from memory and all this stuff is long archived and hard to find, here were the major points that were brought out.

  • Flat is more important than anything else.
  • The mechanical mounting and clamping force had more to do with thermal performace than differences in surface roughness once the parts were flat and reasonably smooth.
  • Somewhere around 1400 grit or better it became impossible to measure any real improvements repeatiblely.
  • final finish should run in one direction, no swirling.

The one thing (only thing :rolleyes: )I can remember everyone agreeing on was that a piece of float glass was the most economical, easily obtainable surface to work with for casual lapping :eek: Everythng else was violently debated ad nausem.


edit, did I say 5 years ? Hell it probally was the 4th post made here and on XS and overclockers, how time flys when you are overclocked.

Here is an oldie, its anecdotal and by no means considered the "answer" but interesting.
http://www.overclockers.com/tips458/
 
Sounds like a rather brutal war, seems like every form of any activity has them really..go start a high pressure vs low pressure thread in any paintball forum and someone is bound to get stabbed. Personally I will be doing the lapping because i believe it will help, both sides do hold lots of weight to them however I believe i lean a bit more towards the mirror finish people.

Thanks for all the input but I'm still kinda worried about water + CPU. I mean if I actually did lap my CPU i would give it a thorough alcohol bath to dilute the water and have it evap quicker...maybe let it sit for a couple days to dry after comp airing it. Is there any worry of water getting under the IHS and sitting on the bare cores themselves?
 
SanityAssassin said:
Sounds like a rather brutal war, seems like every form of any activity has them really..go start a high pressure vs low pressure thread in any paintball forum and someone is bound to get stabbed. Personally I will be doing the lapping because i believe it will help, both sides do hold lots of weight to them however I believe i lean a bit more towards the mirror finish people.

Thanks for all the input but I'm still kinda worried about water + CPU. I mean if I actually did lap my CPU i would give it a thorough alcohol bath to dilute the water and have it evap quicker...maybe let it sit for a couple days to dry after comp airing it. Is there any worry of water getting under the IHS and sitting on the bare cores themselves?
Click to expand...

not really, I dont beleive the C2Ds have the little vent hole some P4 did, and actually the chip itself is embedded in a slug of solder under the IHS (intergrated heat spreader) that you will be sanding on. So its pretty well sealed. I would maybe set it in a sunny window for an hour after you are done or 5 min with a hairdryer set on low. These things are cleaned in the factory with all kinds of sprays and baths after assembly. Caution is good but unless you soak the thing in a glass of water overnight I would not worry too much about it, you asking the right questions and will I am sure proceede in a cautious craftmanship manner. dont sweat it. :cool:

Just handle it carefully, wet the PAPER not the cpu, you dont need a lot of water, and gently scrub the IHS on the paper, litte clean up with alcohol wont hurt a thing, coffee filters are nice source of lint free wipes.
 
BillParrish said:
Just handle it carefully, wet the PAPER not the cpu, you dont need a lot of water, and gently scrub the IHS on the paper, litte clean up with alcohol wont hurt a thing, coffee filters are nice source of lint free wipes.
Click to expand...

Is it okay to clean the non-IHS side with alcohol? And what about if this side gets wet?
 
its generaly ok to clean anything in a computer with rubbing alcohol. its safe on just about all plastics and metals. other alcohols and/or acetone or other solvents are probably not a good idea. if that side gets wet just let it dry, only thing you have to watch out for is the tiny caps on the underside of the chip in the middle, just make sure water doesnt get under them, if it does make sure you wait long enough for the liquid to evaporate. alcohol wont take nearly as long as water.
 
ghost6303 said:
its generaly ok to clean anything in a computer with rubbing alcohol. its safe on just about all plastics and metals. other alcohols and/or acetone or other solvents are probably not a good idea. if that side gets wet just let it dry, only thing you have to watch out for is the tiny caps on the underside of the chip in the middle, just make sure water doesnt get under them, if it does make sure you wait long enough for the liquid to evaporate. alcohol wont take nearly as long as water.
Click to expand...

Thanks a bunch! I think it'd just be best to not get water there in the first place- then i wouldn't have to worry about using alcohol or anything, but at least now i know what to do if i find myself in such a predicament. :)
 
I had my Zalman 9500 done by a friend that works as a welder at Lockheed Martin and he works within 1/1000 of inch and right off the bat the surface was concave or convex forget exactly, but when he had it so polished it was like a mirror and did the chip also a P-D 945 over clocked to 3.75 and runs cooll as a cucumber have no before temps but idle its about 18 -19 c with room temp about 20.8c mobo about 25c and the full temp the highest I have seen playing COD is about 42c and at first I thought these were not right to cool but they are inrange with the bios and other temp utilities. So I guess lapping does work.
 
Only lap it if it NEEDS lapping.

You could end up making contact surface unlevel and the results would be worse than before you began.

Most Zalman pieces are good enough fro the factory.

But if you are going to lap, be sure to use good sandpaper (3M), a flat and level surface, go in the correct order (low grit --> high grit), and take your time. A good metal polish is nice to finish it off with (like Mother's).
 
Mansize_tissue said:
Is it okay to clean the non-IHS side with alcohol? And what about if this side gets wet?
Click to expand...

k, I an NOT recommending this, but I just lapped my IHS, ripped the cpu out, made note of the markings in my motherboard manual for future reference, went to sink with a lapadiry gemstone faceting plate, this is a rigid alum plate with 2400 grit surface, ran the water on a dribble so it ran down to where I was scrubbing the cpu on the plate, and started lapping away, water got on every surface of cpu, however I tired not to drown the thing. Got it where I wanted it, dried it off with a coffee fillter, front and back, set it under a 60W work light for about 5 minutes, slapped it in.

almost crapped with computer started beeping continously on startup. Discovered I had dislodged the rmp monitor for a fan. checked temps, not good. Reseated my waterblock mount, all was well.

No temp improvement but my testing methods were primitive and the AS5 has probally not settled yet.

Proved nothing, no harm done, and the IHS was nowhere near flat, had bulge in middle. so I feel better even if temps did not change.

Gave me a chance to look at the chip again, its pretty sealed, I should have cleaned it with alcohol but didnt.

In short I treated it badly, got it all wet, didnt clean it well, took no extraordinaly means to dry it out, had no problem. Absolutely NOT recommended.
 
Why polishing compounds (like Mother's or Brasso) are a bad idea.

1] - as was said above, shiny mirrored finishes are not necessary. They are best for bragging rights, and to show off your lapping skills.

2] - polishing compounds (like Mother's or Brasso) are designed to leave a protective coating behind which should be removed.

If you lapp properly with clean water you should have no need for polishing compounds. Why add more steps to the process?
 
Susquehannock said:
Why polishing compounds (like Mother's or Brasso) are a bad idea.

1] - as was said above, shiny mirrored finishes are not necessary. They are best for bragging rights, and to show off your lapping skills.

2] - polishing compounds (like Mother's or Brasso) are designed to leave a protective coating behind which should be removed.

If you lapp properly with clean water you should have no need for polishing compounds. Why add more steps to the process?
Click to expand...

Yep. Now, I personally use brasso when I lap, but I always clean the base off with a mixture of 99% isopropanol alcohol and acetone.
 
ghost6303 said:
when bragging to people who dont know that a glass-like finish isnt a good thing, or to people who do know ; it will show off your low IQ
Click to expand...

A mirror finish (provided the block is flat) won't hurt transfer of heat. It won't help anything, but it won't hurt it.
 
Arcygenical said:
Yep. Now, I personally use brasso when I lap, but I always clean the base off with a mixture of 99% isopropanol alcohol and acetone.
Click to expand...
A waste of time IMO. Should be no problem getting a mirrored surface with just sandpaper, if that is your goal. But whatever makes you feel happy. :)
 
Thanks for all the input guys, I just received the kit yesterday so I'm starting now. Seems a bit tricky off the start cause its at a slight angle already, but i don't think it will me much of a problem to correct. Hopefully if all goes well i should have it done in about 30-45 and done well. Otherwise I guess ill be looking for a 9700 here. Hopefully it wont come down to that though. Post pics later.
 
Susquehannock said:
A waste of time IMO. Should be no problem getting a mirrored surface with just sandpaper, if that is your goal. But whatever makes you feel happy. :)
Click to expand...

Aha. Yeah. It is a waste... But things that reflect perfectly sell much faster on ebay than things that don't ;)
 
SanityAssassin said:
Thanks for all the input guys, I just received the kit yesterday so I'm starting now. Seems a bit tricky off the start cause its at a slight angle already, but i don't think it will me much of a problem to correct. Hopefully if all goes well i should have it done in about 30-45 and done well. Otherwise I guess ill be looking for a 9700 here. Hopefully it wont come down to that though. Post pics later.
Click to expand...

Can't wait to see the photos! I don't suppose you have any before temperatures to compare to, right?
 
First off, everyone, lets stay focused here. This isn't a "I DUN THEENK YER IDEA UV LAPPING IZ LIEK GOOD ITS GHEI!!!11ELEVEN" etc. Also, took a rather long break, caught a shot of ADD and wound up hanging around with my cousins. Just got back home and started up again. now I'm on the 800 grit and I can still see scratches, chiefly from the paper. I know its the paper because when i change directions by 90* the pattern of grooves changes with it. Is this supposed to be how it is? Do I go up to the next grit? need some info! Also along the sides of the middle I can see grooves that lead to the fins of the heatsink, I can only assume that I have gone too deep, even though the rest of the heatsink is flat and isn't this way (this obvious).

And yeah afterwards if it can still be used i will do temp comparisons.
 
SanityAssassin said:
First off, everyone, lets stay focused here. This isn't a "I DUN THEENK YER IDEA UV LAPPING IZ LIEK GOOD ITS GHEI!!!11ELEVEN" etc. Also, took a rather long break, caught a shot of ADD and wound up hanging around with my cousins. Just got back home and started up again. now I'm on the 800 grit and I can still see scratches, chiefly from the paper. I know its the paper because when i change directions by 90* the pattern of grooves changes with it. Is this supposed to be how it is? Do I go up to the next grit? need some info! Also along the sides of the middle I can see grooves that lead to the fins of the heatsink, I can only assume that I have gone too deep, even though the rest of the heatsink is flat and isn't this way (this obvious).

And yeah afterwards if it can still be used i will do temp comparisons.
Click to expand...

Go up to 1000!

You'll still see grooves, but they'll be very fine.
 
The whole point of lapping is to make the IHS as flat as possible, not to make the surface as smooth as possible.... The difference in temps will depend on the condition of the IHS, a concave IHS will benefit more from lapping than say one that has a bulge in the middle like Bill's..... and I agree with him that its still worth doing.....

As for mirror finish, It wont give you better temps but at the same time it wont hurt it either... I did the mirror finish the first time I did it, after that I never bothered cos Ive done it already..... If I was to lap for the first time, Ill make it mirror finish too, its just 20 more mins of your time and its BLING BLING.....:D
 
Alright, ill be moving up to 1000...but i dont think im gonna get a mirror finish..just from the grooves that I'm seeing that look to be leads to the fins themselves. Anyone else ever lap one of these heatsinks? I could use some help. Cause I'm pretty sure I wont be getting my mirror finish. However flatness i can do.
 
Dont know about the grooves but brosso will get that thing to mirror finish...... just make sure to clean it really good before mounting...
 
Well I'm moving onto the first sheet of micron paper, 1k made no real difference. Also ill only be using a polishing compound if all else fails..
 
Not perfect, and i killed those sheets of micron paper ~__~. Think next time that i do this, ill not only be ordering everything myself, but ill definitely make sure to get more micron paper.

Also my idle temp is about 35*c and 39*c under load. I'll be installing the HSF tomorrow and ill report back on initial #s and after the AS5 has cured.
 
You must log in or register to reply here.
Back
Top