lapping the cpu= fried cpu???

Sanding the cpu itself I don't think i have heard of anyone doing, and most definitely wouldn't try it myself....

Lap your heatsink....

Wet sand with a mineral oil or clear mineral sprits. No ESD, no water hazard.

i lapped mine. nothing happened except better results with temps.

This guy did it not to long ago.
http://forums.techpowerup.com/showthread.php?t=32771

If you have no power going to the chip, you could throw it into a bucket of water and it'd be fine as long as you allow it to dry before powering it up (and maybe clean off any mineral deposits from evaporation).

please remember, silicon chips without current flowing through them are VERY heat resistant, and can withstand heat FAR in excess of the T junction thermal breakdown limit 125C. So heat killing a non powered cpu would be difficult unless using a blow torch lol

panchoman said:

lots of people have been getting their cpu's fried due to lapping the cpu. people that dry sand get their cpu fried due to esd generated due to the friction. and wet sand posses a huge risk as it cause of the water...

is it really safe to lap the cpu?

Click to expand...

Well if you are afraid that you can burn your cpu then just unclip the metal over it, lap it, clip it back and you are done

mandelore said:

please remember, silicon chips without current flowing through them are VERY heat resistant, and can withstand heat FAR in excess of the T junction thermal breakdown limit 125C. So heat killing a non powered cpu would be difficult unless using a blow torch lol

Click to expand...

What about a super heated TEC?

those grits are perfect. just make sure to go to the roughest first and work towards the smoothest (which is the 600 to the 2000 direction)

ESD is like a game of luck. I never used any special anti-esd stuff and even used my vacuum to suck some dust out of the case and nothing went wrong, but I might have been just lucky...

panchoman said:

where can i find some mineral oil?...

and so i just take the mineral oil [and spill it all over the sandpaper????], which will be taped to a piece of glass and then i take my hsf and move it in a figure 8 till the sandpaper wears out or till i'm happy, and then i take my ihs and move it up and down?

Click to expand...

http://forums.techpowerup.com/showthread.php?t=31579&highlight=d44ve

Same for cpu

a111087 said:

ESD is like a game of luck. I never used any special anti-esd stuff and even used my vacuum to suck some dust out of the case and nothing went wrong, but I might have been just lucky...

Click to expand...

From me and my electronics degree;

Yes that is more or less true, and it is not so common as it sound, it is more precaution.

The biggest risks of esd come in December-February when people are using heat and it dries the air. The more moisture in the air the less likely esd will occur. Another factor is the clothing you wear, in fact if you are naked you are less likely to discharge Sorry I could not help myself. I avoid carpeted areas while installing pc components, because carpet creates static. The room I build in has a hardwood floor and white walls. And yes believe it or not white walls do decrease static. The last and most important thing to prevent cmos damage is to handle circuit boards by the edge. That is the most full proof protection from static you can get, far exceeding grounding straps. You can't discharge into the circuit if you are not touching the circuit!!

I've sanded and wet sanded all my CPU IHS and HSF's and never had any problems *touchwood*

When i lapped my cpu using water, it pretty much got wet everywhere. After letting it dry and some cleaning with isopropyl alcohol, it was fine.

I did it all on a table standing on laminated flooring.

panchoman said:

and also what grit papers do i need for cpu lapping? i think 3x sheets of 600, 800,1200 and 2000 should be enough to lap and shine both the cpu and the thermalright ultra 120 extreme.

Click to expand...

I did mine with 400 grit wo/water and then 800 grit wet sanding paper with plain old water. Put some blustickystuff (don't remember the English name) in the hole and taped over, to keep copper water from getting in the core.

Why I didn't need a 1200 sheet, is because the sanpaper wears out, making it smoother=1200. I Polished a bit with a strip of 2400, not too much, as it would make it harder for thermal paste to find grooves.

Did my Thermalright Ultimate-90i base with the 400 grit I used for my CPU, but wet and it worked fine by it self, as it was worn and only got smoother while lapping.

I'd say buy 2x600, 1x1200, 1x2000 and it's plenty for the CPU and TRUE.


400 frip starting out ..... finished with 800 grit ... finished after 2400 grit (not really so smooth )

edit: and yes it's just a inch thick laminated table, will do just fine without any glass, if you happen to have something similar Just noticed my CPU turned 9months today being lapped, still not fried.

panchoman said:

lots of people have been getting their cpu's fried due to lapping the cpu. people that dry sand get their cpu fried due to esd generated due to the friction. and wet sand posses a huge risk as it cause of the water...

is it really safe to lap the cpu?

Click to expand...

if done with mineral oil you reduce the risk of static electricty AND water penetrating the chip.

the nickel coating is to stop oxidation of the copper as well as prevent the softer copper from being scratched and deformed, copper is a very soft metal. once you remove the protective nickel covering you allow oxidation to start and it can never be stopped. oxidation blocks good heat transfer so over time you will need to relap it to get that smooth polished surface again.

all this is to gain a minor difference in cooling so if you are willing to damage your cpu to get better temps it will work but i just dont think its worth it.

nothing i mentioned even addressed the possible bad lapping job you could do and you may end up with worse temps after.

panchoman said:

where can i find some mineral oil?...

and so i just take the mineral oil [and spill it all over the sandpaper????], which will be taped to a piece of glass and then i take my hsf and move it in a figure 8 till the sandpaper wears out or till i'm happy, and then i take my ihs and move it up and down?

Click to expand...

NO! do not go figure 8 it will round off all 4 corners of the cpu, you go in one direction only then rotate 90 degrees then sand again doing this untill all the scratches from the other direction are gone then turn and repeat. to test if your finished use a sharpie marker to rub all over the surface completely then make a pass or two on the finest sand paper you used and it should be clean, if not, continue the process until it is.

My current CPU is lapped, the IHS was simply warped, quite significantly. When lapping and getting to the copper core of the IHS, I quickly saw an outline of the raised sections of the CPU, the outer edges and the center with 2 little bumps.


My previous CPU was also lapped, both using water in the process. As long as you keep the pins clean, dry, and unbent, and make sure to not get metal dust on sensitive components, it should work fine. Provided of course that you stop sanding before you actually hit the die.

mandelore said:

please remember, silicon chips without current flowing through them are VERY heat resistant, and can withstand heat FAR in excess of the T junction thermal breakdown limit 125C. So heat killing a non powered cpu would be difficult unless using a blow torch lol

Click to expand...

Just because each individual transistor can take a momentary burst of more than 200C doesn't mean the temp limit is 125C. The tJunction does NOT measure the temperature at the actual transistors (would be millions upon millions of measurements anyways) and is simply a representative temperature of an area of the die which a sensor can be fit.


It is possible for a CPU to operate over 125C and it is possible for one to be destroyed ad 75C, as the tJunction temp really is an estimate. 125C throughout the chip (submerging an inactive chip in 125C oil or something) would not damage it in the slightest, you are correct, but this never actually happens in a computing environment. If the 'ambient' die temperature (approximated by the tJunction readings) is 125C and you still have current traveling through the gates, it really only takes nanoseconds for a gate to be activated, heat up beyond the melting point in a very small localized area, and destroy the chip. Just as providing too much voltage and suddenly loading a gate can make a comparatively cool chip fry in the same manner, as the tJunction temp would not register the temp at the one transistor that fails, and that's all it takes.

By keeping your temps down, you reduce the risk of damaging your CPU significantly, as lower ambient temperatures lead to a lower instance temperature and a quicker recovery (making repeated gate opening not build up additional heat).

people do some stupid things in pursuit of negligible if any gain..

definitely a case of fools rushing in where angels fear to tread..

trog

trog100 said:

people do some stupid things in pursuit of negligible if any gain..

definitely a case of fools rushing in where angels fear to tread..

trog

Click to expand...

Depending on the chip, sometimes the gain isn't negligible at all. You'll see quite noticable gains if you have a severely concave IHS, like Intel sometimes sends out.

mandelore said:

please remember, silicon chips without current flowing through them are VERY heat resistant, and can withstand heat FAR in excess of the T junction thermal breakdown limit 125C. So heat killing a non powered cpu would be difficult unless using a blow torch lol

Click to expand...

I dunno, i killed a T-bird 1400B with a cigertte lighter cause i was bored a few months ago.