Lapping Q6600

[petercintn]

Nice job annaconda. That's how this community likes it, lots of pics. Though I've personally given up lapping, mostly, that's how you do it. But the Advanced Auto here sales 2500 grit. :>

Welcome to XS.

[annaconda]

Thanx guys, it would have been nice if i could find 2500 grit.

[ls206]

I think I might give my E6300 a go...the temps on the cores is sometimes 6C different! and runs pretty hot.
nice work there time to get some grit!

EDIT:
Goat Eater, can you post a pic of the Tape idea you mentioned, can't quite see it in my head

[GameGuru]

Napa Auto parts carries everything up to 2000 grit FWIW

[it_burns_when_i_pee]

nice job man

i lapped my cpu but not to that extend. i was nervous to go any further. although seeing yours im considering whipping it out and attacking it to see if it makes a further improvement.

its good to see you had plenty of praises, i posted my results to share and i just got some some abuse!!

http://www.xtremesystems.org/forums/...d.php?t=156216

[sobol]

From a theoretical standpoint of course its not impossible. From a realistic standpoint, highly unlikely. A processor is really densed already crammed with transistors. Putting some kind of implementation will increase their engineering costs and put more investment into protecting their product, instead of producing the best damn product they can (sounds familiar Sony?). Plus, honestly if Intel did not want people to overclock, I'm sure they could find a way so that a processor can not exceed over x fsb. But then they would pretty much lose the ethusist market.

Although losing the ethusist market would not hurt them so much because we area minority, we are also the most vocal and annoying ones (from a company standpoint ). So Intel rather just give us the RMA, instead of having to spending more money on an issue that is rarely going to happen.

I won't agree with the above opinion.

Forum people are the first to test the products, and most competent to do it. They are creating opinions, whether something is good or bad.
Its spread up from here to news papers, web portals or computer builders.

And just from my personal experience. I've got friends and most of them are coming for advices when they need help with buying or upgrading computer. They know what I know from the forum.

There are also people here who working in big companies and deciding which brand to use.

This is like pyramid, we're at the top. In fact I think that we're most influential group on the market.

[timber715]

has anyone with temperature difference between cores of the quad(usually 4-5C idle and 3-7C during load) been able to correct it by lapping?

[ls206]

Before lapping as i stated @ 3.51

Idle: Core1: 52C, Core2: 50C, Core3: 46C, Core4: 49C,
Load: Core1: 62C, Core2: 60C, Core3: 56 or 57 C, Core 4: 60C

After lapping @ 3.51

Idle: Core1: 40C-42C, Core2: 42C, Core3: 36-37C, Core4: 38C
Load: Core1: 50C, Core2: 50C, Core3: 46C, Core4: 48-49C

hmm...results don't quite suggests that it will "even up" the different core temps, they all dropped together..

[annaconda]

hmm...results don't quite suggests that it will "even up" the different core temps, they all dropped together..

To be honest i was thinking the same thing, and i think i need to lapp the D-tek Fuzion too. For some reason Core0 & 1 runs hotter than others, i think because all the programs go to Core0 & 1.

[hifiking]

That Q6600 looked just like mine, I lapped it yesterday too, but I used 240 paper only, and 15 min done. The temps dropped nearly 10C aswell. Had to do my waterblock too.

[petercintn]

What little bit I got to play with my Q6600 before the 680i board gave up the ghost, I had 3 cores within 3c of each other and one core 10c higher than the lowest core. I had used the big o-ring on the Apogee GTX and bowed it for better temps on the E6600 I have. Not pleased with the temps with the quad, I put the little o-ring back in the block and used an extra 2mm plastic washer on top of the springs just to get a little more pressure on the quad. The flimsy pcb of the 680i couldn't take it I guess, the mobo never worked right since and then died. So I did get some 2500 grit and lapped the Apogee just to take a little of the bow out. When I pick up another mobo, if I find a 10c difference in the cores I guess I'll lap the quad's IHS. Shame I can't cut the lid off these Intel chips, wouldn't have no worries about lapping IHS's.

And Annaconda, I think your results wuold have been no better with the 2500 grit, you did a good job. I was just teasing.

[annaconda]

Peter you motherboard DIED, sorry to hear.

[TheGoat Eater]

Best place (AND EASIEST) to find high grit sand paper - you local model/hobby shop as they sell it usually sells by the sheet and have a good selection of grits

[TheGoat Eater]

EDIT:
Goat Eater, can you post a pic of the Tape idea you mentioned, can't quite see it in my head

I will get my E6850 lapped soon and will post pics for sure but am testing the CPU on air for now w/o lapping as I am waiting on my watercooling supplies and will post after I get it and the board (IP35 pro)... Basically you get some painters tape and depending on the width overlap the tape - for narrow tape - on the bottom of the cpu pad and press it down good for a good seal and trim the overhanging tape with a razor blade. then when you lap w/o water or alcohol you don't risk getting dust in the CPU - I figured it seals well when I do volt mods with conductive pen so It won't F w/ the bottom of the CPU

[annaconda]

Any picture of that would be appreciated.

[vocch]

You guys looking for lapping kit....here you go from 400 - 2500.

http://cgi.ebay.co.uk/Lapping-kit-co...QQcmdZViewItem

[Morais]

Very good post dude, i enjoyed all the pix

[ak_47_boy]

The worst concave i have has was on a celly D and it only took me a half hour to get it to a mirror.

[nealh]

dont lap the fuzion..it is desigend to be slightly convex

[TheGoat Eater]

Here is quick snaps of my tape method for lapping - 3M painter's tape for best results, it is not the widest form so I used two overlapping layers - no pics of the finished product as I was in a rush to install it with my new watercooling.

[vocch]

I use the chip protector backing then seal the sides up with masking tape...my Q6600 looked nice and flat until i started sanding it lol.

[scaramonga]

I just got myself this lapping kit in UK (400-2500) and comes with a sheet of glass.

http://www.watercoolinguk.co.uk/cata...roducts_id=560

Gonna start the work tonight

[annaconda]

Good to see people are lapping. It's worth it my friends, trust me.

[tastyratz]

The tape idea seems good, but the only thing that concerns me is conductivity and residual electricity in the chip (if any?) I mean remote possibility but its something.

Lapping should even out temps because your creating an even contact surface for cooling. Because theres 4 cores in different locations you may then see cooling differences based on heatsink/waterblock design. If 1 of the cores was in a bowed location then it will be reflected in its temps.

Goes with a lap that is practical. Making it a mirror finish wont truly net you any real gains, its the making it flat that will. Just hitting it with some 180 to take down to the copper then finishing with some 400 will probably net you 95% of the gains you would have gotten going to 2500. Considering its usually only a 5-8% gain in most scenarios its hardly worth it. going to 2500 is just bragging rights.

Lapping is a really great idea though for people who really are at the edge.

[nugzo]

The tape idea seems good, but the only thing that concerns me is conductivity and residual electricity in the chip (if any?) I mean remote possibility but its something.

Lapping should even out temps because your creating an even contact surface for cooling. Because theres 4 cores in different locations you may then see cooling differences based on heatsink/waterblock design. If 1 of the cores was in a bowed location then it will be reflected in its temps.

Also you must remember that the contact between the core and underside of IHS may not be good. I've seen that a few times. If you lap it perfectly straight and still have big fluctuations, it's probably the core/IHS contact, not IHS/HSF/Waterblock contact.