Who has lapped Their CPU?, Info Options

[mtotho]

Hey, i was just wondering how many of you have lapped your CPU.

Here is a pic of what it looks like
 805D_lapping_guide_008.jpg ( 150.41K ) Number of downloads: 146


This has helped my q6600 (pre g0 steppings) at 3gz go from 75 full load to about 62 celcius. I am also using a thermalright 120 extreme HS with a 120mm Fan.

Here is what you do. Take your CPU out and protect the back of it(i used the original casing that it came in)
Find a very flat surface (i used a big piece of glass) and tape atleast 400-600 grit sand paper. You can move up the grit.. but to start out at 2000 would take to long.

These CPUS are very concave, you can see it if you put it on an angle.

To sand it, you want to dab a little amount of water on the sandpaper to keep any static shocks away. Sand until you have what looks like a flat surface. Make sure your moving the CPU on the sand paper, not the sand paper on the cpu. also make sure everything is extremely flat. Dont worry about a mirror finish, you just want to make sure that you cpu is not concave.

Ive done this with a q6600 and e6600 both with great results. You can also do this to the Heat sink fan your using... but depending on the shape of it, this could be more difficult.

[apple apple]

Good info. May have to give it a try sometime.

[BashfulTux]

Hey, i was just wondering how many of you have lapped your CPU.


This has helped my q6600 (pre g0 steppings) at 3gz go from 75 full load to about 62 celcius. I am also using a thermalright 120 HS with a 120mm Fan.

Your in need of a serious upgrade in the case cooling department with temperatures that high my G0 I run now which I bought pre-lapped used runs at around/below 30C idle and barely managed to hit 50 running prime95 small fft's. I have the extreme version of the same HSF as you do with a 1400rpm fan on it. Now it has a VID of 1.2125 which is what I run it at with the vdrop on my board it runs at 1.17v 3150mhz under load since I have the C1E enabled most times it runs at 1.09v idle. Even my previous q6600 with a VID of 1.3250 which I again ran with the C1E enabled so it usually ran at 1.15v idle, ~1.26v under load same setup with HSF ran at around 40C idle mid to high 50's load at 3000mhz. One more thing you should try to get those scratches polished out mine has a mirror finnish on it that way you will get better contact and check to make sure the bottom of the heatsink is flat as well.

[mtotho]

My q6600 runs sub 30 idle. It doesnt have g0 steppings which means it uses alot more energy and isnt as good thermally. I have the thermalright 120 extreme, that is what i meant. I get 62 at most with prime on for 2+ hours. My CPU is rated to be able to stand 75c so i have plenty of space to work with... besides .. nothing taxes my cpu as much as prime does.. and not for 2 + hours. Also That picture i posted is not of my CPU. I am not about to open my case, take out my motherboard, unscrew the 120, clean of the Arctic silver, and then take a picture. I believe i lapped mine further than that.. i did it with 400 grit, then 800, then 2000 for a mirror finish. Which btw, it doesnt matter if you have those scratches really.. what matters is you dont have a cpu with the concavity that even the hubble space telescope would be jealous of.

I am not in need of a serious upgrade in the cooling department... my cpu runs way cooler than anyone else with my same cpu/steppings.. and im not about to spend 280 bucks for another cpu of the exact same specs just so i can get g0 steppings. g0 Steppings make a beautiful difference.. i wish i had it. You should Overclock your higher... thats the point of having it lapped. You can easily get highy clock on Air.

[ocZZZ]

I did on one of my past q6600 and it lowered the temps by about 4c.

[bonestonne]

i've lapped a CPU, but not what i run in my workstation. i lapped a Celeron D 2.53GHz 533MHz CPU, but i didn't notice much difference with it.

but i also think that its very important to get at least a near mirror finish. not having that kind of finish means that the CPU has a larger surface area [comprised of the dips and valleys in the IHS created from lapping]. so if you do plan on not lapping it to a mirror finish, i'd suggest Arctic Silver 5 thermal compound [preferrably not thermal tape, or cheap-o thermal paste] so that you can fill the gap [of course be very modest in the amount you put on though].

other than that, great info for anyone. I run a Scythe Ninja on a Pentium D 940, and it keeps it tame and tolerable by me until i go C2D or C2Q.

[mtotho]

Well about the mirror finish.. yea of course i made mine have a mirror finish but really.. there isnt much of a difference. I mean stock cpu doesn't have a mirror finish and its concave.. so having a flat.. more shiny/mirror finish than what you started out with is better than nothing.

[wilstar]

I still think those temps are a bit high I'm using a B3 Q6600 at 3.2 G with a Thermalright Ultima 90 (both lapped) and the hottest core is at 61-62 after prime for a couple of hours. Did you lap the heatsink as well?

[FabricioGS]

I did it.

[Azurael]

I daren't lap my Q6600, even though the IHS is seriously concave... It means my load temps, even on my new water cooling setup at 3GHz are over 50c

[olinboy1]

I've considered lapping my chip, but I'd have to take my motherboard back out if I wanted to because of my HSF. The tuniq is a f*cking beast. Might try it next time I teardown and clean inside.

I just ran prime 95 (x4) last night for about an hour on my Q6600 (B3) OC'd to 3ghz

my avg. readings:

core 00 - 52c (maybe lapping would help normalize the hotter cores?)
core 01 - 51c
core 02 - 48c
core 03 - 47c

I'm pretty satisfied for a B3, and I had just cleaned and re-set the tuniq, so hopefully once the arctic silver 5 bonds a bit more the temps will be even better.

I also had all four cores under 50c doing an TMPGEnc encode on a 55min video (done in 11mins ) since I figure that's more of a "real world" usage test. I understand testing all 4 at 100% for stability sake, but what else is really going to tax the CPU that much?

This was all in Vista Home Premium.

Does anyone think Temperature Monitor in OS X might read temps a little bit on the high side? I'm going to have to do some more tests with 10.5.2

[capt_cope]

The finish DOES matter, but mirrored isn't "better" it's just different, and requires different TIM. According to the guys who make it, AS5 is intended to work with less than smooth surfaces. It's too viscous to work well between two mirrored surfaces. If you lapp your IHS and polish it to a mirror finish you'll be better off lapping and polishing your HS, and using a thinner TIM. And if you don't polish your IHS, use AS5 or some other thicker TIM. (and lapping your IHS is kinda a moot point if your hs isn't flat, so you might want to go ahead and check that while you're at it. My Ultra 120 Extreme was anything but flat.

[Listed1st.com]

My MacBook Pro was running really hot, practically scalding (I thought the 45nm was supposed to be cooler) But aparently not! I read this thread on lapping the CPU and I figured it's worth a shot. Problem is I can't get it to boot now! It won't even turn on. I followd al the steps, right down to wetting the sandpaper to prevent static. It was pretty polished before I proceeded, I was thinking perhapse if the surface was a little more coarse then it would have a larger surface area. Any ideas on why it won't turn on? I didn't think lapping it would ruin it!

Attached File(s)

 Before.png ( 158.07K ) Number of downloads: 31
 After.png ( 163.48K ) Number of downloads: 37

 

[DiaboliK]

My MacBook Pro was running really hot, practically scalding (I thought the 45nm was supposed to be cooler) But aparently not! I read this thread on lapping the CPU and I figured it's worth a shot. Problem is I can't get it to boot now! It won't even turn on. I followd al the steps, right down to wetting the sandpaper to prevent static. It was pretty polished before I proceeded, I was thinking perhapse if the surface was a little more coarse then it would have a larger surface area. Any ideas on why it won't turn on? I didn't think lapping it would ruin it!

you dont lap a laptop cpu!

you are basically polishing the core itself, not a good idea.
you most likely have to shell out some $$ to get yourself a new cpu, sad to say but......ah screw it HAHA!

-D-

[capt_cope]

My MacBook Pro was running really hot, practically scalding (I thought the 45nm was supposed to be cooler) But aparently not! I read this thread on lapping the CPU and I figured it's worth a shot. Problem is I can't get it to boot now! It won't even turn on. I followd al the steps, right down to wetting the sandpaper to prevent static. It was pretty polished before I proceeded, I was thinking perhapse if the surface was a little more coarse then it would have a larger surface area. Any ideas on why it won't turn on? I didn't think lapping it would ruin it!

I really hope you're joking. They are all talking about lapping the IHS, the metal covering found on most modern cpus that you can buy as components (like a C2D e6750 etc. etc.) The pictures you're showing are of the actual core. If you sanded that... you're SOL.

[teknojunkie]

true sorry but that aint a desktop cpu, hopefully you can get that fixed.

[Listed1st.com]

Okay, Okay, I'm just messing with you! That was an old Core Solo I screwed with just for kicks and giggles... But I do say, I wouldn't put it past some people!

Right now I'm hoping to find anyone with some info on a Mac Pro Logic Board I got... It appears to be a prototype and it has jumpers on it that I have no idea what they do (Other than make the system say there is 1 cpu with 7 cores when I have 2 quad core cpus in it)... so if anyone has some idea on any info about the board, I'd love to hear it!

Attached File(s)

 Picture_2.jpg ( 340.13K ) Number of downloads: 44
 Picture_3.jpg ( 350.87K ) Number of downloads: 33
 Picture_4.jpg ( 306.03K ) Number of downloads: 30

 

[teknojunkie]

could be that since its a developer bored it combines both cpu as one, just a wild guess not sure, will look into this

[fandel]

I lapped my processor before even putting it into my box so I would not know the difference it made in termperature. I got a Q6600 B3 and I'm not really too satisfied with it since I was only able to overclock the thing to 3.2Ghz but I turn it down since it required a huge amount of more voltage. I'm currently running it at 3Ghz with watercooling and its around 52-55 on full load which is not bad but I will be upgrading it to a B0 soon.

[~pcwiz]

I've considered lapping my chip, but I'd have to take my motherboard back out if I wanted to because of my HSF. The tuniq is a f*cking beast. Might try it next time I teardown and clean inside.

I just ran prime 95 (x4) last night for about an hour on my Q6600 (B3) OC'd to 3ghz

my avg. readings:

core 00 - 52c (maybe lapping would help normalize the hotter cores?)
core 01 - 51c
core 02 - 48c
core 03 - 47c

I'm pretty satisfied for a B3, and I had just cleaned and re-set the tuniq, so hopefully once the arctic silver 5 bonds a bit more the temps will be even better.

I also had all four cores under 50c doing an TMPGEnc encode on a 55min video (done in 11mins ) since I figure that's more of a "real world" usage test. I understand testing all 4 at 100% for stability sake, but what else is really going to tax the CPU that much?

This was all in Vista Home Premium.

Does anyone think Temperature Monitor in OS X might read temps a little bit on the high side? I'm going to have to do some more tests with 10.5.2

I get temps like those with my Q6600 G0 with a STOCK COOLER that is NOT LAPPED