Lapped the i7

on the topic of warranty voiding. I can't remember a time at all, ever, that I've ever seen a processor need to be RMA'd. It's always the motherboard, or the ram. The only real way to bust a processor is to drop it/physical damage or overclocking it. Both void the warranty anyway.
 
Would you lap the i7 the same way you would with core 2 quads?

Is there anything different?
 
I'm about to put my i7 build together and the water block I bought is lapped and polished to a mirror shine, and I was considering lapping my CPU. Mind sharing your method for those of us who have never tried OP?
 
I said nothing about lies nor called anyone a "liar". I am an engineer, I say what I mean and mean what I say. Untruths most certainly do not equal lies. There is a huge fundimental difference.

This entire subject "is it worth it" has also been beaten to death many times. But whatever, have at it. I could ask Kyle if he will change the name of the site to havenotesticalesOCP but I after seeing his method of erasing data off of old hard drives I do not think I will.

http://www.hardocp.com/news/2009/03/21/50_bmg_api_vs_18_hard_drives/
Ok, Mr Engineer, graph the data ... should be very easy given irrefutable, unarguable evidence as you are suggesting. I also have seen many claims on the net with some obviously made up by people that just want to see havoc. Which I do not believe the OP is guilty of.

Have yet to see a review for lapping. AND, since packaging processes change, prior efforts with older CPUs are not necessarily valid to the i7s. I will admit that I do not know those details else I would post a link ... just to cut thru the BS ... but I don't where that comparison might be.

I have my opinion differences with some of Kyle's opinions, but we are all privileged to have to have them. On the other hand, when I read (& especially watch) a "Kyle review" ... I am into it. I'll credit him with the choice of RAM I bought for my system ... per his video review. KUDOs to Kyle ... nothing like being able to make semi-intelligent product choices based on at least one reasonably well documented test. It is still "anecdotal" as it was just 1 test ... versus 3 or what ever; as we engineers would need if performing tests that our job was dependent on. But this was not my situation ... so 1 was good enough AND ALL THAT THERE WAS. If we had access to Dell or Intel or whomever internal test results things might be different. Since I bought the same RAM as his review & had similar results on a similar m/b, maybe we have *2* good tests?:cool:

This is the difference and what I was encouraging the OP to do ... put out as well as he remembers his process. I was interested in precisely what *he* did. Mind you I am not negative about lapping as I have seen it commented about for years. The total effect as well as comparison to the latest TIMs would be of interest tho.

Ok, clearly I am skeptical.
 
I've lapped many a sole of hand planes (hobby). The process takes hours and the ability for a novice to do it wrong is high. Generally most people attempt to lap something and only make the object go from concave to convex. Manufacturers purposefully produce them convex because that ensures a stable mating surface.

When you start, the ability for you watch the process is easy. However, once finished, if you screwed it up, the tried and true razor blade method on a convex surface isn't perfect.

This same tendency can be seen with people who use a belt sander. It's natural to put more pressure on the edges as your body stretches. When you lap your parts, make sure they don't wobble. Make sure you stand in a way that doesn't stretch your arms causing you to put more pressure on one edge as you extend in one direction and another when you retract. Be gentle! A light, but firm pressure is ideal. The ability for our fingers to apply exact pressure on a 2inch square is hardly ideal. Pushing hard so you can get it done faster doesn't help, it's like trying to push hard with a hacksaw. Let the sandpaper do the work.

Here is a link to a Fine Woodworking video - Flattening a Handplane Sole with a small scratch in it. This is a good starting place to get a feel for how you should lap. Check out the grits he uses just to get a scratch out. I myself start out with DuoSharp Bench Stones to asses the work, then either move down to sandpaper, or work through 4 different grits on these stones. Finally, I'll follow up with a very fine grit.

Hopefully this goes to show that there is a wide range of quality and interpretation of what a successfully lapped cpu might entail, and how damn easy it is to actually make things worse.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
wow, this thread has turned into quite the debate

ok, let me give you the low down on exactly what i did

i wet sanded the entire time with plain water

i started with 600 grit and spent probably 45 minutes on that grit and went through about 8 sheets of narrow paper

i held the cpu with my hand, fingers on opposite edges and went back and forth on a piece of glass on my kitchen counter, i rotated my body from one side to the other in order to give the cpu a 180 degree turn about every 10 passes, and while i was turning made another couple passes, would stay still for 10, then turn and repeat, about every 100 passes i rotated the cpu clockwise 90 degrees in my hand and continued

once i had almost all of the nickel removed i moved up to 1000 grit, i performed the same procedure for about another 45 minutes until i had the beginnings of a nice mirror shine, this is where i took the photo i posted

at this point i cleaned the cpu extremely well with alcohol and q tips, being very careful not to leave any little cotton fuzzies behind

i decided i wasn't quite happy with the finish yet and moved up to the 1500 grit, i followed the same process for another half hour or so, then cleaned it up well again, at this point i really liked the finish and felt that it seemed much better than stock to me, so i cleaned it thoroughly, allowed it to dry for about 15 minutes(and sped up the process with a hairdryer on cool setting, then installed it into the board and popped my heatsink back onto it and fired her up

i hadn't changed any bios settings since the last shutdown where i had decided that i had hit a temperature wall and noticed that my idle temps had dropped approximately 5 degrees from what i had been seeing previously from the last cold start, i then loaded it up with intelburntest and was pleasently surprised that after 20 runs it was now appearing to hover at 70C rather than the 80C it was at previously

i then started tinkering with bios settings again

so i can't say one way or the other whether just the lap was responsible for my temp drops, i didn't remount the hsf several times, i didn't repeat any testing with the exact same settings, i saw lower temps and was pleased so i continued tinkering

if anyone wants to follow what i did, that is the best write-up i can give you...this wasn't intended as a study, i simply wanted more headroom to push more voltage and hopefully attain higher clocks

good luck to anyone who attempts this, i hope your results are as good as mine :)
 
anyone else that does this should post their results...perhaps in a more scientific fashion than i did lol
 
I'm tempted to do this to my i7, I've had decent results with a few P4's and my previous 6600. Only thing is time lately. The last p4 I had saw the biggest drop while my lapped q6600 + lapped TRUE gave around 5C on average.

If the HSF is really not flat, I can see 10C. Most incorrect mountings and high overclocks usually make getting an acceptable stable state for the system tough to do.
 
WELL!!! We will just see what my minions have to say about this!!!

Aren't all forums pretty powered by opinions Bill? What exactly is the point of your post?

After Minion Semenary in High School I Pursued a degree in Ass-Kissery with a minor in Yes sirery
In Community College. From there I could only dream of Hobb knobbery and might I dare, Keeping up with the Jonesery. I feel qualified to say "Yea Bill what is your point?" What he said.
 
Back
Top