Printable View
Lapped my CPU+Heatsink with grid 400-600-800-1000-1200.
Temps before lapping :
[picture]
heatsink before lapping :
[picture]
Heatsink after lapping :
[picture]
Cpu before lapping :
[picture]
Cpu after 5min lapping :
[picture]
Cpu after grid400 :
[picture]
Cpu after lapping :
[picture]
Temps after lapping :
[picture]
looks good but would have thought you would have had a bit better temps since you have the TRUE
did you reduce the amount of paste on re-application? and whats the mounting like? loose, tight or in between?
Difference between core 0,1 and 2,3 are almost 10C, maby bad contact between core & ihs :(. I remounted the HS serveral times but it stays like that.Quote:
Originally Posted by it_burns_when_i_pee
The mounting is tight enough i think, i can move the heatsink a bit left or right still though. Maby i can try to put a penny between it for more pressure.
it looks more like 2 cores are being loaded, 2 are not, and are receiving residual heat from the other two. im assuming @ idle the temps are close to each other? otherwise ignore this post :)
i was meaning overall temps not individual cores
you got a pic of the internals of your machine? maybe improved airflow will help those temps a bit further?
Well, it's not exactly scientific testing methodology, but I just used some MX2 and found its thinness to be very beneficial to getting that super thin layer you want. It's cooling better than the last few applications of AS5, and I applied the pastes the same way each time (AS's line-of-paste across cores method) and always use Articlean. The MX2 squished out into a translucent layer whereas the AS5 was more an opague layer that didn't spread as well.
With AS5 I was seeing 70-72C on one of the dies. This was about the same as before I lapped, btw. With the MX2 it's holding at 66-67C. That die sure does run hotter than the other. Lapping didn't really help with that difference. I kinda suspect that one die has higher leakage power than the other... Could be more solder on one too, I guess. Who really knows....
I'll probably use MX2 from now on after getting this improvement and really liking its ease of application. It's cheaper too I believe.
AS5 may possibly work better if you spread it thinly yourself. But I've used up so much paste and pulled my heatsink on and off so many times in the past week that I'm a bit burned out. lol
have to ask, are you seeing that your paste has mushed out by lifting the heatsink after applying it?
the as5 squished out nicely for me - noticed after remounting the heatsink when i changed case layout.
your temps seem very high if you have lapped. im getting 25 celcius on idle @3.2ghz 1.4v ..... 44ish CPU load or 51 if in a game - extra heat contribution thanks to a large GPU heatsink next to the processor!!
Hmm...I can get 3.2 @ 1.300v :cool: I'm @ 3.6 right now w/ 1.45v. After I lapped, I got these temperatures. This is @ 3.2ghz 1.300v:Quote:
Originally Posted by it_burns_when_i_pee
Idle Unlapped - Idle for lapped and unlapped was unchanged
http://i11.photobucket.com/albums/a1...anUnlapped.jpg
Load Unlapped
http://i11.photobucket.com/albums/a1...anUnlapped.jpg
Load Lapped
http://i11.photobucket.com/albums/a1...lmanLapped.jpg
Oh and just for :banana::banana::banana::banana: and giggles I tried 3.8 but it wasn't stable.http://valid.x86-secret.com/show_oc.php?id=248132
i havent tried my cpu at anything lower than 1.4 yet
i tried 400x9 and it loaded windows but wasnt stable :(
someone said i might want to drop FSB voltage to try x9 multiplier. on the weekend im going to try dropping the cpu voltage and see if its stable at 8x multiplier.
still new to the overclocking of C2D, am i right in thinking if i go 9x 400 i might wanna try upping CPU voltage?
Yeah I've checked a few times. The AS5 spread out, but not nearly as well as the MX2 does. Currently application is the best temps I've seen yet out of this setup.Quote:
Originally Posted by it_burns_when_i_pee
When I Prime 3.5 GHz at about 1.36v (measured by Everest) 1.45v BIOS I see almost 80 C on one of the cores. There can be about a 10C difference between cores. This hasn't changed with lapping or different paste. Primeing at 3.2 GHz 1.23v (Everest) 1.325v BIOS or so, one will sit at around 69C while the other is at 63C. This is with Prime95 25.x running small FFT (really bakes like nothing else lol). When idle at 3.2 GHz / 1.23v, I see 26C or so on that strange non-core temp reading while the cores themselves are at 33-35C or so.
I'm recalling these readings from memory cuz I'm at work :). They should be close tho.
This board has some crazy Vdroop combined with maybe poor voltage readings (no way lol). When I set 1.325v in BIOS, I get 1.23-1.24 at load in Windows. To get up to a 1.4v reading, I have to run about 1.5v in the BIOS.
Here it is:Quote:
Originally Posted by annaconda
http://img406.imageshack.us/img406/3...2977fr7.th.jpg
http://img406.imageshack.us/img406/7...2975tz9.th.jpg
I improved the thermalright mounting even more, i put a washer under the "X" bracket. Now there is more pressure on the CPU.Quote:
Originally Posted by it_burns_when_i_pee
I just tested an 30min OCCT test 3250mhz @ 1.27v and it passed. temp on warmest core didnt went higher then 57C. This b3 is not a sucky clocker though ;p
http://img65.imageshack.us/img65/416...occtry9.th.jpg
I've considered chopping off the pathetic plastic Intel-style mounting pins on the Scythe Ninja Plus and using bolts. But it wouldn't be possible to get the cooler off without taking the mobo out if I did that. Ick... And there's no backplate anyway so pressure increase would be limited.
I'm sure the fairly low pressure is why the AS5 doesn't spread well.
lapping my q6600 made 0 difference...
lapping my e6600 made 0 difference...
in my experience.. lapping doesn't work
same here, no difference after my lap.Quote:
Originally Posted by the highlander
i've had bad results after lapping too. saw a temperature increase but it was because i didnt mount it properly. upon removing the heatsink i saw the paste had spread unevenly to one side. re-applied properly i saw excellent results.
the as5 spread out fine for me. after putting half the recommended thermal paste on arctic silvers website, i positioned the heatsink on top, pressed down firmly with a fair amount of pressure rotating just a few degrees 2 or 3 times. not wanting it to pop up and allow air in, i kept the pressure on the top of the heatsink and screwed it in place.
oh and if you want the best chances / ability to mount it properly, remove the board from the case, its small compromise
Lapping had close to 0 difference for me on idle. But for load temps, my 4 cores were all within the same range as the other (check above post).
Question: So when I lapp my CPU, does it make sense to lapp the heatsink as well?
Greetz
Ganzir
Definitely, except if your heatsink is already flat and the surface has no scratch.Quote:
Originally Posted by ganzir
even if its flat a smoother surface has lets pits and valleys (the microscopic imperfections)
all these "holes" allow air to be trapped, even though are very small, it all adds up in preventing the best possible heat transferrence. the idea of thermal paste is to fill these gaps. anything is better at conducting heat than air but the thermal paste's properties is the reason it is so widely used.
but remember if you have a smooth finish you will need less paste
Working on my big project a watercooled SLR system. Going all out, decided to lap by CPU. Was going to be smart, just get it flat don't worry about shine. Was going so well I saw the shine coming, got excited, was now into the shine. Must be careful, never did this before. Got the shine! Clean up and re-install my CPU ready to see that 10C drop in temps. ......machine would not boot. Clear cmos etc, etc still no boot. What is this. Take CPU out, inspect CPU .. something on pins, clean, clean and clean. Re-install CPU, no boot, try try and try no boot. Remove CPU inspect, clean, re-install -no boot. Money money money, time to check ebay for new CPU. Went to bed. Next day inspect CPU, clean, install no boot. Make bid on eBay. Inspect with microscope, more cleaning needed. Install ---BOOT. Am I happy. Still winning bid on ebay. Temps look good, now running ORTHOS 22 minutes Core Temp reading 35C on air. Uguru (Abit) reading 45C. Good work but ah the sweat and tears.
I'm glad you finished it with a good ending. I was gonna be crying for you if it didn't :p:
Just got outbid on the new CPU that I no longer need. Someone paying too much.
I lapped my CPU and installed a D-Tek Fuzion (from a regular Apogee) and my temps improved a little bit, but now I can't even come close to the overclock I had before. Any ideas???
I did notice that the mounting hardware for my D-Tek fuzion seems to be bowing my motherboard a little bit, do you think a backplate would help? I do beleive im getting a good spread cause when I remove the block, I can remove the retention plate and the block actually stays attached, a little twist and it comes off.
I did remove the NB/SB heatpipe to clean the stock putty and apply AS5. temps went down about 3 deg there. I honestly don't think that's the issue though.
@ deathspine - nightmare with the cpu not firing, care must be exercised at all times, glad it was sorted (and your not forced to buy another cpu). also those are excellent temps for 3.6ghz, would love to get mine upto 400x9 but it's not stable. i was getting 19 degrees idle @ stock and 25 @ 3.2ghz but your load temps there are better than mine and your running 400mhz faster. im jealous!
@ laydback - back plate should help but i have no idea why its not clocking anymore. when i took my cpu out the bios settings reverted to defaults because it claimed i put a new processor in!! perhaps the same has happened with you and your overlooking a setting? hopefully thats all it is