PDA

View Full Version : Lapped Ultra 120 - super cool & ultra quiet


it_burns_when_i_pee
08-22-2007, 06:30 AM
Thought i'd share with you the success i have had with the goal of a quiet cool gaming pc.

I've done it before with great success so there was no question that my new build was to lap all the heatsinks within the machine - CPU, motherboard and video card.

I've read a few people on here have used 800 grit or even lower to lap the heatsink, this is far too abrasive and will erode too much too quickly. I use 2400 silicon carbide which is almost as smooth as paper but leaves a mirror like finish. Cleaning either surface i use acetone with a good quality kitchen roll (cheap ones tend to break up and leave fibres). Lastly everything would be re-applied with arctic silver.

The ultra 120 has had considerable success even though the cpu contact has ridges which can be felt with a finger nail. For comparison here is a picture (this picture is from a website) - i didnt think of taking pictures until half way through the last heatsink.

http://www.gwynfryn.eclipse.co.uk/hs5.jpg

My method of lapping is to put a few drops of water on a flat surface like your kitchen worktop, place the sandpaper ontop, wet the top of the sandpaper with a few more drops and begin lapping. Moving in circular motions with a bit of downward pressure, I rinse away the black debris build up every few minutes. Rotate the way its held then repeat. After 2 hours of elbow grease, this is the result:

http://www.gwynfryn.eclipse.co.uk/hs1.jpg

http://www.gwynfryn.eclipse.co.uk/hs2.jpg

Neat eh? Well I decided to lap the processor too. I did this slightly different because I didnt want to get the CPU wet. Instead of putting water ontop of the sandpaper i did it dry and every minute wiped the processor and sandpaper with acetone to clean the debris build up. Several minutes later its smooth and mirror like, streaks can be seen but cannot be felt. I was happy with this.

http://www.gwynfryn.eclipse.co.uk/hs3.jpg

I had already lapped the heatsinks on the video card and motherboard by the time i thought of taking pictures. But will say that the heatsink on the video card was practically mirror like anyway. I think I hardly improved on it. The motherboard heatsinks were also very good but both were done regardless.

All surfaces were thoroughly cleaned and then installed with arctic silver. From experience the better the surface contact a little less paste is required than normal because there are shallower valleys and pits on either surface for the paste to gap.

So whats the temperatures?

http://www.gwynfryn.eclipse.co.uk/temp1.jpg

Pretty cool! The Nvidia Ntune wont run on my machine because the motherboard is a P5b Deluxe (although the website says you need a nvidia motherboard to run it). Downgrading to an older display driver i had, it stated the video card temperature was sitting at 38 doing nothing. According to the bios, on idle the motherboard sits at 29 (CPU reads as 22.5 to 23 degrees in bios)

Under load the cpu reaches 41 degrees (via intel thermal tool), the motherboard 38 (bios) and the video card sees 62 (older display driver).

The system has a total of 4 x 12cm fans all of which are whisper quiet and cannot be heard unless you put your head up close to the tower. There are 3 x nexus fans plus 1 x 12cm fan in the seasonic power supply.

Here's a pic of the layout. You will notice blue electrical tape and that the retention bracket by the video card is un even and a bit buckled. This is because the courier dented it recently in transit. I've straightened it out for now and I am awaiting the compensation payout. When it comes I'll be replacing the case. Any recommendations of a case welcome.

http://www.gwynfryn.eclipse.co.uk/pc.jpg

The rear exhaust fan is mounted externally to see if it improved temperatures thinking of drag, it dropped the motherboard temperature by 1 degree but nothing else.

The front and back fan grills have been dremelled out and the fan guard on the PSU removed. Notice that there is a cable tie holding the heatsink; this is simply to stop it sagging whilst upright as it can move a few mm and I am uncomfortable with this.

It's several months old now (dent happened recently) and although its doing everything I want it to I am looking into overclocking for the first time in 4 years! Will report any success, or failures.

Before you ask, spec is

E6600
2Gb GEIL 4-4-4-12 DDR 800
Asus P5B Deluxe
MSI 512mb 7950GT
Audigy 2 Platinum
Seasonic 550Watt PSU
2 x 160b Maxtors in RAID 0

I did have 2x 150gb raptors in RAID however one died and has been sent back under warranty (is a year old). Given the performance of some perpendicular SATA 2 drives I will be selling the raptors and changing. I had 2x 74gb raptors in the past and one of them died too.

Any comments or questions on the above welcome!

jellysandwich
08-22-2007, 07:36 AM
nice temps. what is ur ambient?

Morais
08-22-2007, 08:23 AM
Very good lapping job dude! The pictures are very cool

it_burns_when_i_pee
08-22-2007, 01:32 PM
cheers guys

ambiet? well its summer here so roughly between 18-29 degrees celcius. today i was particularly hot and it didnt affect the temps.

i've been pointed out that the pic of the temperature isnt quite accurate. i have azureus downloading, several IE windows, outlook and winamp playing.

so here is another screenshot with nothing running except essentials. with the room on temperatures i should be ok for some overclocking but yet to get down to it.

http://www.gwynfryn.eclipse.co.uk/temp.jpg

Vapor
08-22-2007, 03:22 PM
snipDude....your name....

Little inappropriate, dontcha think? Would you like that changed? (I'd recommend it...)

SP1
08-22-2007, 03:45 PM
18 to 29 degrees!
Where are you getting those teperatures?
It's cold, windy and wet.

it_burns_when_i_pee
08-23-2007, 05:33 AM
at the moment its 25 degrees outside, celcius

when i had my shuttle sn95g5 and a 3800 i lapped that to kingdom come and i had temps of 29 to nearly 50. proper directional ventilation and removal of air plus good heatsink finish will give you good temperatures. but bear in mind some people are running overclocked chips at just 10-15 degrees higher, when i overclock mine it will be interesting to see how much higher it gets. the fans i have to dont push that much air as they are of low rpm but when i built it a while ago my intention was cool n quiet. i think i achieved that

as for my name i have already been contacted about it. been told to simply leave, been told to re-register, and one person stated they will report me to a higher authority. no sooner had i had that pm, admin has advised he will change my name. i think its a bit obsurd really, then again i have a sense of humour but not everyone is the same

Vapor
08-23-2007, 11:47 AM
I'm an admin as well and the name isn't about having a sense of humor. I've talked to other admin, we both agree the name should be changed.

If someone put that in their post, we'd kindly ask them to edit or delete it...or if it were a continued problem with the member, we'd simply do it ourselves. With a name it's attached to every post, and should be changed.

Please contact me with a different name of your choice (so long as it's appropriate). It's very easy to change a name and you probably want to pick it yourself before we do.

Nice CPU temps, btw.

joshd
08-23-2007, 12:49 PM
between 18-29 degrees celcius

TAT seems well out then. Use Core Temp

it_burns_when_i_pee
08-23-2007, 12:57 PM
the temps dont decrease in the winter either because of course i have the heating on

originally with the temps i left the cpu heatsink without a fan and was a 2 or 3 degrees higher, however since they are all the same noise level and can hardly be heard i put it back on. the case has sound proofing on the panels and i put some around the hdd's as the original raptors were easily denoted which isnt something that bothered me cos its when it works but since i had some foam to spare i stuck it anyway.

as for my name, i am speaking to njkid32 to change it.

on another note, i mentioned above i am going to be changing the case when i get paid out by parcel force ... like the idea of BTX style however the heatsink on the video card faces the wrong way for that. Any suggestions on a case?

it_burns_when_i_pee
08-23-2007, 01:03 PM
TAT seems well out then. Use Core Temp

heatsinks are cold to touch, except video card

asus probe (program sux) said the same, bios reads a 2-3 degrees higher but i read thats common on the deluxe board

also tried speed fan program to verify way way back but as it used to cause an exception when i closed it i just removed it from the machine.

my friend is lapping is next week when he wont need the machine. he has a skythe ... the finish will be as good as mine. will take pics and post results, although i doubt he will be bold enough to lap the processor

Jodiuh
08-23-2007, 02:47 PM
Nice job...and hysterical name. It fits the avatar perfectly. ;)

I need to do a better job on my TRUE, but it's such a pain to hold. Unless it's moved VERY slowly, it's bound to skip on the paper.

evoic
08-23-2007, 02:54 PM
Very nice job, OP

Here's my lap job on the TRUE, for the sake of comparison. ;)

http://i58.photobucket.com/albums/g263/Evoic/eBay%20Pics/Forum%20Sales/2-1.jpg
.
http://i58.photobucket.com/albums/g263/Evoic/eBay%20Pics/Forum%20Sales/1-1.jpg

lad
08-23-2007, 04:07 PM
I find it funny that you think 800 grit is too abrasive, but aren't worried about lapping on a rough countertop...

evoic
08-23-2007, 08:19 PM
I have to agree here.

When you say, "it will erode too much, too quickly", I kind of have to laugh.

If your project is starting with something that needs to be flattened, erosion is the only way to get there. ;)

I go 400, 800, 1000, and finish with 2000.

I think my result above speaks for itself.

it_burns_when_i_pee
08-24-2007, 05:28 AM
think about it logically

your roughing it up, then smoothing with finer and finer paper. your desired goal is a smooth finish

well im just skipping straight to the smoothest paper. which is what your going to end up doing anyway. in the meantime i'd be smoothing, you would have spend time just grating away metal until you decide "right now i will smooth"

i've done it with the past 3 machines and had excellent results. if you compare the finish on mine to your pictures you can argue the finish on mine is better and even from edge to edge. until you try it yourself i wouldnt slate it, first lapping job i did was with my thumb and palm of my hand ... had terrible results, because it was an uneven surface. the worktop i have is level and not course, the pictures i took were in the utility room which worktops above the washing machine is just a cheap n nasty one to cover the tops of the machines and sort clothes.

the reason i lapped it like this is because a friend of mine worked in a place where they machined down piston heads for very precise pumps. they put the round sheets on a turntable like unit and the heads are placed against them and spin whilst water trickles over them. they are sanded for best part of an hour. the finish on the metal was like a mirror or the platter of a hard drive, so smooth its unbelievable.

he aquired the silicon carbide sheets and we both attacked our own heatsinks. try it yourself before you slate it because its different to what you have tried. i think the results i got speak for themselves and im not going to try and convince, im simply sharing my experience for others to try.

evoic
08-24-2007, 02:54 PM
A few things:

1. I agree with you, I believe your lapped result is shinier than mine, and probably a "cleaner" finished product.
Beautiful is a word that comes to mind.

2. No one is saying that your way is wrong, simply a different cat that you're skinning.

3. If you want shiny - your way appears to be superior.

4. If your heatsink or whatever you're lapping is concave.....I doubt it's effectiveness compared to the method that I outlined.

it_burns_when_i_pee
08-24-2007, 03:04 PM
the heatsink i lapped that was concave was a aero cool ht tower. but thats because i used my hand and thumb. the athlon was sitting 60 degrees idle and crashed under load - live and learn. one i have now i was maticulous, maybe a bit too careful with? just after christmas the E6600 wasnt exactly cheap and i didnt want to cook it

if the surface is mirror finish there are less pits and valleys which is what you want - the whole point of paste is to bridge this gap to conduct heat to the heatsink. get some 2400 paper and give it a whirl, lots of elbow grease and time required though but i bet you will be happy with the results

anyway i have had no recommendations on a case. i liked the one i had as it has 12cm in and 12cm out fan slots which is something i want on the new case. not after a window or shiny piano like sides. anyone own or can recommend something?

technodanvan
08-24-2007, 03:28 PM
think about it logically

your roughing it up, then smoothing with finer and finer paper. your desired goal is a smooth finish

Technically your goal should primarily be a flat finish, then having it smooth is secondary.

Stome sinks and processors are curved enough it'd take a real long time at 800 grit to get her flat.

Of course if yours was already flat it doesn't really matter, looks pretty good. :)

MaxxxRacer
08-25-2007, 03:16 PM
think about it logically

your roughing it up, then smoothing with finer and finer paper. your desired goal is a smooth finish.

NOPE. You're desired goal is a FLAT finish. FLAT is MUCH more important than smooth. This been proven in the watercooling world a number of times.

In one test that Cathar (arguably the most well known watercooler in history) performed, he found that 800grit paper actually provided the best thermal trasnfer performance. Granted, this may differ with different TIM's, but with the TIM he used, 800 grit provided the coolest temps.

So no, shiney is not what you want. It may look cool, but doesnt do you much good. Your time would be better spent brining the block and CPU to a facility that does precision grinding. They will grind/sand down the CPU and block to a level of flatness that is not achievable by hand.

Your last option is to use the abrassive slurry used to create handmade telescope lenses. You put this slurry on the heatsink, turn it upside down and rub the CPU on the heatsink so that the two surfaces are rubbing against eachother. Doing this will cause both surfaces to be the same curvature (inversly) and thus will have VERY little air gaps between them.

MaxxxRacer
08-25-2007, 03:21 PM
anyway i have had no recommendations on a case. i liked the one i had as it has 12cm in and 12cm out fan slots which is something i want on the new case. not after a window or shiny piano like sides. anyone own or can recommend something?

Lian-Li PC-07B covered in Dynamat Extreme on the inside of the case walls and with a HDD Silencer in the 5 inch bay to reduce vibration.

I have this setup on my girlfriends PC and It is completely silent with a A64 X24200 @ 2.6GHz and a X1900GT overclocked.

If it wasnt for the power light you wouldnt know the system was on. When I was building the system I had the front panel led unplugged and freaked out a few times because I couldnt tell if the system was on. You can barely hear the PSU fan if you put your ear right up to the PSU, but thats about it.

it_burns_when_i_pee
08-25-2007, 09:37 PM
have already looked at that case but there is another lian li i quite liked the look of, asthetically ... cant decide

as for the results ive had, well you cant argue it doesnt work and its been with the last few machines i have had. just thought i'd share it, if people wish to believe otherwise im not going to try and convince them. i did it for my benefit, sharing it was just an extra task i could have done without, and im not that charitable a person!

lad
08-26-2007, 12:39 PM
as for the results ive had, well you cant argue it doesnt work and its been with the last few machines i have had.

Actually, we -can- argue, and we -will-, to benefit the lurkers who might be misguided by your misunderstanding of the facts.

Here's the facts:
One: On two microscopically flat adjoining surfaces, there is very little actual contact. The peaks from one surface contact the peaks from the other. We use thermal interface material (TIM, thermal grease, etc) to replace the insulating air molecules in the valleys with heat-conducting TIM.

Two: Even the best TIM does not conduct heat nearly as well as metal. TIM still -insulates- from heat transfer. So we want the absolute thinnest layer of TIM between each surface as we can reasonably get. The best way to get the -most- area in a thin layer is to lap -as flat as possible-. As the other posters have mentioned, MANY tests have shown that using 800 grit paper is the sweet spot for flatness. Going to smaller grits (1000, etc) has little effect on temperatures.

Now, I will add one fact that I found important for -me-...
-I- found that using 1000 grit, and then 2000 grit (after lapping as flat as possible with 800 grit) allowed -me- to get a more -reliable- flat surface, because of -my- difficulty with hand-lapping. -I- found that I could more easily see when I was lapping with offset pressure, ruining my 'flatness'. This may well have been due to my NOT using water on the sandpaper, and YMMV.

But the bottom line is that 'we' want -flat-, mirror-smooth does not help with heat dissipation, and many of us do not need epeen points...

evoic
08-26-2007, 02:07 PM
I for one keep an epeen points scoreboard next to my computer, specifically for threads like this.

At the moment, I am trailing it_burns_when_i_pee, but I have a healthy lead over lad.....for what it's worth.
.
.

lad
08-26-2007, 04:01 PM
I for one keep an epeen points scoreboard next to my computer, specifically for threads like this.

At the moment, I am trailing it_burns_when_i_pee, but I have a healthy lead over lad.....for what it's worth.
.
.

:ROTF: :clap: :up:

it_burns_when_i_pee
08-27-2007, 12:33 PM
wow i guess the laws of physics work differently where you live. is the sky blue there too?

i will have to inform toyota and the remainder of the industry, who no doubt spent millions in research, that they got it all wrong and the finish they strive for is just a huge waste on resources and a needless bigger carbon footprint

also i'll start informing other forums and communities that they got it all wrong too.

lastly i need to get in touch with intel and ask them whats going on ... despite their thermal tool working fine on everyone elses machine, it saw that my heatsink was mirror like and gave me false low low readings for the fun of it!! duping me into beliving what i've read and what my friends particular industry specialises in. oh what fools we have all been.

it_burns_when_i_pee
08-27-2007, 12:35 PM
oh and what the phuck is epeen?

lad
08-27-2007, 05:36 PM
wow i guess the laws of physics work differently where you live. is the sky blue there too?
< plus lots more off-topic ranting >

Dood,

You spent all that time and energy, and completely avoided all the points I made.

It's fine that you don't intuitively understand heat-flow, or the physics of TIM, really! But to think that you can just waltz on the scene and start spouting nonsense, and actually insist that everyone assumes you know what you are talking about... is insane.

I -don't- claim to be an expert... but I've spent easily 400 hours in the last year studying and learning and testing. The spreadsheet where I track my tests and experiments over that time has more than 260 columns of tests. I've seen, very clearly, firsthand, the effects of rough and smooth lapping, on 3 different HSF's.

The temps I achieved on those 3 sinks, allowing for vcore and ambient and fan flow, were always near the best temps reliably reported. I have first-hand proof that flat is more important than shiney.

Now you can continue to act like a whiney little brat, or you can suck it up like a man, admit (at least to yourself) that you don't grasp the topic, and start learning. I hope you chose the latter, cause there's hundreds of people here who want to share their knowledge with you. You aren't going to lose a skirmish with me, 'cause that's not what's motivating me. I want to teach others what I know. I think that's a cool thing.

And, honestly? Lose the stupid nick. The lowest form of humor is scatalogical humor.

Epeen? Google, child.

Scyth3
08-27-2007, 06:43 PM
There is definitely something wrong with those temps because it is under ambient. Unless you are blasting your A/C.:)

PyroFire
08-27-2007, 07:47 PM
this is offtopic but what is that passive cooler ( i think its passive ) that you have on your gfx card.

Also your cpu temps are great especially for a passive cooler

and whats so offensive about the name it burns when i pee, i thought it was funny :x

it_burns_when_i_pee
08-28-2007, 02:05 AM
Dood,

You aren't going to lose a skirmish with me, cos i am the internets greatest keyboard warrior. i may be a 38 year old ginger loser, living in my parents basement in the real world, but online and behind my keyboard, i am a god, believe what i say is fact or else



so your basing that fact that i only recently joined means i know nothing? theres logic.

epeen - i see, sorry but i have a girlfriend and have other interests, so forgive me for not being an internet nerd with profound l33t vocabulary.

HA this is great!!

i dont doubt you keep spreadsheets, sounds like you have way too much time on your hands (or incredibly anal retentive) or most likely both. the fact you take everything so seriously backs this up. i can imagine you updating your precious sheets after you spent another hour looking at your bios, another at forums, updating todays fan rpm, another ..... oh my god, get a life

using google i found like 3 pages of lapping guides all of which, are the same as what i did. strange that! also they are reporting the same or even better temps than i got. if you care to read the ones that use lower grit rating, they say its because its quicker to remove any paint / plating. its assuring though that the laws of physics work the same where they live.

i think you should google lapping guides and techniques and inform the 100's and 1000's that despite their success, what they've done is completely wrong and a waste of time, you are right and everyone else is wrong. dont forget to mention that your not a newbie to a forum therefore you MUST know more than them. also mention that the thermal transference isnt working and that their bios, thermal program etc is just tricking them into believing what they did works.

its same as it has with me, even though my temps are just a degree off from one another depending whether its speedfan, asus probe, intel's thermal tool or the bios - its a conspiracy pc's have against people who lap heatsinks to a better finish than yours.

lastly before i go to my girlfriends, everyone else seems to like my nickname, you don’t, whats that say? maybe you should get a sense of humor, but avoid budget or store brands, they just dont cut it. you'll have a better chance at fitting in with the rest of the world if you could losen up a bit

Alik4041
08-28-2007, 02:46 AM
I thought it was better to lap in one direction only and use isopropyl instead of water. Just what I've heard. I'll be lapping my cpu (q6600) a couple weeks after it arrives. BTW, nice job OP.

it_burns_when_i_pee
08-28-2007, 08:18 AM
cheers!

i used pure acetone because of its properties (and i already had some). only problem is that acetone evaporates in literally 2-3 seconds when applied and wiped but takes gung with it.

my mate lapped his heatsink just today and he used the dry lint free wipes you get for cleaning a tft. he said they were better than using kitchen roll. i got a box of them here and compared the two and i think he is right, so you may wanna try them???

Alik4041
08-28-2007, 12:32 PM
cheers!

i used pure acetone because of its properties (and i already had some). only problem is that acetone evaporates in literally 2-3 seconds when applied and wiped but takes gung with it.

my mate lapped his heatsink just today and he used the dry lint free wipes you get for cleaning a tft. he said they were better than using kitchen roll. i got a box of them here and compared the two and i think he is right, so you may wanna try them???

Thanks for the tip. I'm still wondering though if lapping in ONE direction only (sliding the cpu/hs all the way to the other side and picking it up and sliding it the same way) is better than the circular motion you performed.

it_burns_when_i_pee
08-28-2007, 01:09 PM
i did the circular motion because i read a few guides. the only difference to my method is i that i went straight to the smoothest stuff you can get.

the 2400 silicon carbide sheets my friend got for me, the acetone i had. i thought higher grit would be a waste to begin with because your only smoothing / flattening it later with finer stuff. if you think about it, very very fine layers i know, but your eroding more surface by grating away metal with more course sheets. so i just went straight to the finer stuff and ended up with a good surface. that was several pc's ago and did it ever since. im not the only person thats done this either. there is another who did it with the scythe.

i joined here initially to look into overclocking - havent done it since the early athlon xp and had a costly experience. whilst here i thought i'd contribute and share the success i had.

are you getting that new zalman cooler in your other thread? it looks like it would be really effective.

PyroFire
08-28-2007, 04:43 PM
still didnt tell me what that passive cooler on your gpu was ;x

Alik4041
08-28-2007, 09:27 PM
i did the circular motion because i read a few guides. the only difference to my method is i that i went straight to the smoothest stuff you can get.

the 2400 silicon carbide sheets my friend got for me, the acetone i had. i thought higher grit would be a waste to begin with because your only smoothing / flattening it later with finer stuff. if you think about it, very very fine layers i know, but your eroding more surface by grating away metal with more course sheets. so i just went straight to the finer stuff and ended up with a good surface. that was several pc's ago and did it ever since. im not the only person thats done this either. there is another who did it with the scythe.

i joined here initially to look into overclocking - havent done it since the early athlon xp and had a costly experience. whilst here i thought i'd contribute and share the success i had.

are you getting that new zalman cooler in your other thread? it looks like it would be really effective.
You make a very good point about the sanding part. But here is what I saw is a disadvantage with the circular motion which shouldn't affect performance noticeably but it might.

When it's done in a circular motion, I would think that molecular rings would be made that can not be seen with the human eye. These rings would trap hot molecular gas in a way.
http://img209.imageshack.us/img209/8426/77348411qm4.gif

But when it's done in one motion, the lines that are made allow the air to relieve these gases of their heat.
http://img519.imageshack.us/img519/9782/a2tv2.gif

If I am wrong, please do tell because I really wanted to do it the circular motion way which should be easier :D I don't think that this will affect performance at all maybe 1C max.

About the Zalman, I dont know yet :p: . Price is more important than its performance currently IMO simply because when zalman made this thing, they knew it would be much better than the stock cooler (and that's all that matters to me). If it's expensive, I'll wait for G92 and see if it fits on that and make my judgment then ;)

still didnt tell me what that passive cooler on your gpu was ;x
http://www.digital-daily.com/video/7950gt_chaintech_and_msi/index02.htm

EDIT: 100 Post Count :woot:

fireice2
08-28-2007, 10:25 PM
You make a very good point about the sanding part. But here is what I saw is a disadvantage with the circular motion which shouldn't affect performance noticeably but it might.

When it's done in a circular motion, I would think that molecular rings would be made that can not be seen with the human eye. These rings would trap hot molecular gas in a way.
http://img209.imageshack.us/img209/8426/77348411qm4.gif

But when it's done in one motion, the lines that are made allow the air to relieve these gases of their heat.
http://img519.imageshack.us/img519/9782/a2tv2.gif

If I am wrong, please do tell because I really wanted to do it the circular motion way which should be easier :D I don't think that this will affect performance at all maybe 1C max.

About the Zalman, I dont know yet :p: . Price is more important than its performance currently IMO simply because when zalman made this thing, they knew it would be much better than the stock cooler (and that's all that matters to me). If it's expensive, I'll wait for G92 and see if it fits on that and make my judgment then ;)


http://www.digital-daily.com/video/7950gt_chaintech_and_msi/index02.htm

EDIT: 100 Post Count :woot:

I dont think that would be a problem considering you will be putting some thermal grease anywayz. :D

evoic
08-29-2007, 12:04 AM
Figure-8 method, for the win.

it_burns_when_i_pee
08-29-2007, 01:49 AM
hmmm i see your point with the sanding method. figure of 8 seems logical too but i think ultimately it would be the same as circular just different and probably easier method of doing it (my arm was aching afterwards). being copper i would imagine the transfer of heat through the material would be even regardless of method of lapping, but there is only one way to find out.

ideally somebody needs 3 untouched identical heatsinks and spend an equal amount of time lapping on each and test them all. maybe someone would be willing to give it a try? 3 different people wouldnt cut it because one may be pressing harder or finish sooner. whichever method you try i recommend you stop every minute or so and clean off the residue build up as you dont want to work that back in, it will slow the progress of your desired finish

i choose circular because of the guides i'd seen. thinking of when you sand down filler after patching up a car using rough to erode the excess then smooth to get it neat as possible for painting was why i decided to use only smooth. as i said it may be tiny layers but there are few people on here concerned about the heatsink having enough pressure on the processor, taking away more heatsink wont help that issue.

check this : http://www.xtremesystems.org/forums/showthread.php?t=156774&page=2

its an awesome job lapping a q6600 in a straight line method. the finish was like my heatsink. i didnt go further on my processor as i didnt want to kill it, however, seeing his finish im thinking of taking it out sunday and lapping it. if i do i will try the straight method and post back with results.

as for my video card, sorry PyroFire. its an msi NX7950GT-VT2D512EZ-HD and came with that pre-installed. i got it at the beginning of the year when they were a bit expensive. i removed the heatsink to improve the finish and upon cleaning the heatpad they had on it, found it was practically mirror like anyway. i did a bit and would say i improved it by a fraction. but i think the arctic silver made the biggest difference on temperature over the heatpad.

you can see the card here:

http://msicomputer.co.uk/index.php?func=proddesc&prod_no=1045&maincat_no=130&cat2_no=136

it_burns_when_i_pee
08-29-2007, 02:28 AM
i've posted a new thread about a case i may purchase but will require some modding. check it out and tell me what you think.

http://www.xtremesystems.org/forums/showthread.php?p=2399184#post2399184

sobol
09-13-2007, 03:51 PM
Well I'm in trouble now. I've done 6h lapping job to HS. The temperature raised up. They used to be 68,68,66,66 with 1.45v @3.6ghz while now I can't even get that stable and on 1.40v@3.4 my temps are 78,78,69,69.

I can't really tell what is wrong. I'm sure that TR's mounting is very tight ( with my mod). Amount of AS5 is right ( used my own method again, but this time even less compound. After taking HS off, sometimes TIM is equally spread up sometimes not ( I've remounted everything 4 times since lapping).

it_burns_when_i_pee
09-14-2007, 03:56 AM
have you cleaned both surfaces? too much or too little paste will produce bad results.

many guides out there say to apply half the thermal paste recommended for un modified heatsinks to one that is lapped to such a finish.

since your temps are really high, touch the cpu heatsink - if you can hold your hand on it then the heatsink isnt reaching those temperatures meaning you have a contact issue. if the heatsink has good contact then the temps of cpu and heatsink will be just a few degrees off from one another.

if your heatsink isnt as hot its possible either chip or heatsink is concave

sobol
09-14-2007, 04:30 AM
I would rule out the clip ( mounting ) issue here, I'm sure you've seen photos of my TR's mounting.

have you cleaned both surfaces? too much or too little paste will produce bad results.

many guides out there say to apply half the thermal paste recommended for un modified heatsinks to one that is lapped to such a finish.

since your temps are really high, touch the cpu heatsink - if you can hold your hand on it then the heatsink isnt reaching those temperatures meaning you have a contact issue. if the heatsink has good contact then the temps of cpu and heatsink will be just a few degrees off from one another.

if your heatsink isnt as hot its possible either chip or heatsink is concave

GAR
09-16-2007, 12:06 PM
nice job lapping, i can never get my laps to look mirror finish, maybe i need a higher grit sandpaper.

it_burns_when_i_pee
09-16-2007, 12:10 PM
deffinately!

theres a guy on here that got a mirror finish like my heatsink on his Q6600

awesome! i intend to pull mine out and do the same. lost my nice temps since the overclock though, im getting 25 ish idle @ 3.2ghz & 1.45v core

cant get the thing to load windows at higher speeds than 3.2ghz, crashes. at least its free speed but still jealous of many other peoples overclocking results