What's the risk of lapping

T

tpfaff

2[H]4U
Joined
Jan 30, 2006
Messages
2,526
I want to lap my c2d and I was just wondering what risks i run doing so, are there any? I mean you just sand progressively with different grits until its all copper right?
 
You void you waranty obviously and if your over aggressive and slipshod I suppose you could burst through and do damage to the core but as you've said if you stop once its all copper and then switch to very fine grit (1000+) and get progressivly finer until you get that morror finish you should be fine. (though many argue the mirror finish is nothing more than aesthetic)

I think the grit you start on is important. I start on 600 or 800 grit, coarser grit than this, while quicker, increases the risk of you going right through, as I've outlined above, and not noticing until its too late.:eek:

I do about 50 laps check and so on.
 
You can go through until it's all shiny copper if you want to, but the important goal here is to get the bastard flat. For all intents and purposes, it's impossible to sand down the IHS through to the core with handlapping in one session. Your hands would cramp into dust before then, especially with higher grit paper :D There is a small risk of building up static charge which could be potentially dangerous in contact with the pads of 775 chips. To minimize the risk of ESD, use foam to pad the underside, wear a 100% cotton or other natural fiber t-shirt (or none at all), and perform the task on uncovered surfaces away from any rug.
 
Kinda bothers me coming from the machinist world.......all this discussion of lapping......to get a true surface....lapping is a three stage process....if you intend to get a truly flat surface.........you should use a surface plate and flatten the processor.....then flatten the heatsink...and then use lapping compound and work the heatsink and processor together. In fact...once two surfaces are lapped together....a very small droplet of light machine oil between the two surfaces should make them feel like they are magnetized.....it will develop a substantial amount of adhesion.

Using a pane of glass is a reasonable starting point....but trust me...it ain't really flat.....:D
 
Mad Machinist said:
Kinda bothers me coming from the machinist world.......all this discussion of lapping......to get a true surface....lapping is a three stage process....if you intend to get a truly flat surface.........you should use a surface plate and flatten the processor.....then flatten the heatsink...and then use lapping compound and work the heatsink and processor together. In fact...once two surfaces are lapped together....a very small droplet of light machine oil between the two surfaces should make them feel like they are magnetized.....it will develop a substantial amount of adhesion.

Using a pane of glass is a reasonable starting point....but trust me...it ain't really flat.....:D
Click to expand...

Sounds like what you say to be true, just using my common sense...

So here's a question...

What would you suggest for me, the average OC enthusiast, to do in order to get a true flat surface on my Q6600??
-V
 
Here's a risk:

If you don't check to see if your IHS is flat, and thus needs lapping...

... you might finish the lap just to find that it doesn't improve your temps at all.
 
My main concern is the oxidation of copper. Wouldn't that affect the chip surface over time? Although the surface under the HS won't be as oxidized, but I'd imagine the uncovered areas would get rusty.

Anyone check their copper after a few months of use? Does it require re-lapping to get rid of the rust?
 
Maxtr0sity said:
My main concern is the oxidation of copper. Wouldn't that affect the chip surface over time? Although the surface under the HS won't be as oxidized, but I'd imagine the uncovered areas would get rusty.

Anyone check their copper after a few months of use? Does it require re-lapping to get rid of the rust?
Click to expand...

Copper doesn't rust, rust refers only to the oxidisation of iron-base metals (rust is iron oxide).

Now copper does tarnish and this is a form of oxidisation. You may have noticed the green patina on old statues, roofs and even pennys but this is only on the exposed surface, it also requires the presence of carbon dioxide, moisture (water) and oxygen.

The base of your heatsink and heatspreader are more or less air-tight once mounted and covered by thermal compound. No air = no oxygen or carbon dioxide = no oxidisation and certainly no water! :D.

I also believe that the heat generated by your processor will help ward-off oxidisation of the copper as heat drives volatiles off the surface of a material.
 
Kinda bothers me coming from the machinist world.......all this discussion of lapping......to get a true surface....lapping is a three stage process....if you intend to get a truly flat surface.........you should use a surface plate and flatten the processor.....then flatten the heatsink...and then use lapping compound and work the heatsink and processor together.
Click to expand...

Mr Machinist is exactly correct and there are special machined surfaces made to exacting tolerances (the surface plate he refers to) for machining. Similar surfaces exist for use in gem cutting/polishing. They have specifications for both flatness and surface roughness. There are special rouges and polishing compounds as well. Far finer than any sandpaper you can buy.

Not to sure about the double pane glass as a flat surface. Probally fine as it just float glass in a frame but finding small sections would be difficult and I dont see the point. Classic advice and the one I give (I have access to lapidary/gem tools and use them) is to use single pane float glass. You can "borrow" the glass from a picture frame in a pinch as long as it is really glass. Because of the way it is made, it is a very cheap, reasonably flat and smooth surface more than sufficient for what we are doing.

There are reams of old posts on this with the bottom line being that the application of thermal paste and the mounting method inconsistencies make the point moot. In other words the variations that naturally occur when remounting the CPU and applying the paste obscure any minor benefits to "extreme" lapping. It never hurts to do the best job you can but most of the time the extra effort does not translate into better temps.

Also reams of old posts on surface finish with most now conceding that a mirror finish is not desirable. IBM recently released studies showing that a particular pattern of surface roughness is desirable but a lot depends on the size of the particles in the paste used. A perfectly (unattainable) smooth flat surface would cause the heatsink and CPU to be separated by the diameter of the particles in the paste where uniform small "valleys" that were optimally as wide as the particle diameter and having a depth of 1/2 particle diameter would allow the larger particles in the paste to migrate into these depressions and reduce the spacing while completely filling the voids.

All nice BS.

Use a pane of glass and go for flat, polish it shiny if it makes you happy as long as the polishing process does not affect the flatness. Dont loose any sleep over it. I also agree, lapping the heatsink without checking and doing the IHS as well, is a waste of time unless the heatsink is horribly out of flat. (my waterblock was in good shape, the C2D IHS was awfull with 2 distinct 'hills" quickly observed with the razor blade test and confirmed with just a couple of passes while sanding) After doing both the IHS and waterblock to a 50micron finish on gem polishing equipment (Master 8825 gem fauceting machine) I could tell no difference in temps. Probally due to the inaccuracies in temp measurement and not having a good software testing enviroment (core temp under windows, and its hard to make windows run exactly the same , its always doing something in the background and I was too lazy to force it to stop).



---------------------------

I just recently took the maze 2 all copper water block off my 5+ year old NF7-s box and while the exposed copper was dark the area in contact with the cpu was as nice as the day I put it on. I should have just left it alone LOL.
 
Vengeance said:
in order to get a true flat surface on my Q6600??
Click to expand...

Strictly speaking, you don't need a flat surface. You need the CPU and heat sink to have the same surface profile. The description above, where you finish by lapping the CPU and heat sink together with polishing compound between them, ensures they will mate. They can be flat, cylindrical or spherical as a result, but if they mate, they will have good contact. If you change to another heat sink, the process has to be done again to get the new heat sink to mate. If they are truly flat, then they mate easily - hence flat is a good option.
 
BillParrish said:
Probably fine as it just float glass in a frame but finding small sections would be difficult and I dont see the point.
Click to expand...

You can get flat float glass in reasonable sizes at any wood-working shop that sells materials for "Scary sharp" sharpening of cutting tools (planes, chisels etc.) This won't be single-diamond glass either - it will be a thick, stiff piece fo glass.
 
bail_w said:
what is consider "flat" to you? since you are a "pro" why dont you make some suggestions..........
Click to expand...
Actually MSC ( www.mscdirect.com ) has a small surface plate...granite, which is perfect for what ya'll are trying to accomplish ( http://www1.mscdirect.com/CGI/NNSRIT?PMPXNO=4762560&PMT4NO=28929475 ) I use a larger one for reconditioning end plates on hydraulic pumps and air motors......but the small ones are sufficient for lapping a heat sink and processor. You can also get adhesive abrasive disks form an auto body paint supply store that are perfect for this. Finish by using ultra fine valve lapping compound between the processor heat spreader and the heat sink.

I have to agree 100% with BillParrish......mirror finish really isn't the way to go for this purpose......for the reasons he stated.

If you know someone that works at a monument company......see if they can score you a piece of polished granite......it ain't perfect....but it's way better than a piece of glass

If you are really anal....you can make a surface plate from three pieces of steel or cast iron....the pieces are first surface ground and then worked against each other with lapping compound in rotation and checked with prussian blue......after you get 'em close....then you get a carbide scraper and work down the high spots......you can get it to .0001 over the area of your plate with that method.....but wayyyyyy too much work for this application.:D

And BillParrish......do you have the goodies for facet cutting stones......????I come up with some good to great rough stones occasionally....up in North GA....rubies....garnets, citrine, amethyst saphires......and the occasional emerald....maybe you need some stock ......and as my wife loves jewelry......we could come to an arrangement.
 
Or you could just use a scrap piece of double pane glass. It's typically flatter than a single pane. A hardware store might charge $2 for a scrap piece large enough for your needs.

Unless you are going to make a career out of lapping, the machined flat surface jigs are overkill for CPU/HS lapping.
 
A glass covered 8x10 picture frame worked just fine for me.

Took the massive concave dip out of my e4400, got a 10C drop in temps.

Flat enough.
 
You must log in or register to reply here.
Back
Top