lapped my IHS on a Q6600 (pics and results)

[bing]

One question I believe there hasn't been an aswer for yet.

...Were you guys never worried about breaking off (via coming loose) the IHS due to applying pressure to it during sanding/lapping?

Is it glued on so well that this becomes a non-issue?

Don't worry about that IHS epoxied to the substrate, the bond is "very-very" strong.

 

[GFORCE100]

Don't worry about that IHS epoxied to the substrate, the bond is "very-very" strong.

OK thanks, I'll do it this weekened then for both it and the IFX-14.

I've already bought 500 x 3, 800 x 3, 1000 x 2, 1200 x 1, 1500 x 1, 2000 x 1 and 2500 x1 grit sandpaper.

 

[bing]

OK thanks, I'll do it this weekened then for both it and the IFX-14.

I've already bought 500 x 3, 800 x 3, 1000 x 2, 1200 x 1, 1500 x 1, 2000 x 1 and 2500 x1 grit sandpaper.

Wow .. the monster IFX-14 !!!

If this is your 1st time, I suggest you check this & this . Good luck !

 

[graysky]

@toeknee and GFORCE - be sure to post your results, take some before/after pics AND BE SURE to record some before/after temps noting what room temp is for your runs. Post the results to this thread or make a new thread and post a link to it here

 

[lil_falco]

Well, finished lapping my AMD 3800+. Find the results here: http://www.ocforums.com/showthread.php?p=5187203#post5187203

Next up is my E6700

 

[ToeKneeATX]

so i lapped my cpu tonight. first off i would like to apologize for the crappy pics, but my digi cam is broken and i only had my phone to snap pictures with. my u120 extreme was lapped before i installed it the first time.

here is what it looked like:

what it looked like after i removed it:

what the cpu looked like after removal:

what the cpu looked like before lapping:

cpu after lapping:

idle before lapping:

load temps before lapping:

idle temps after lapping:

load temps after lapping:

sorry for no inbetween shots, i was busy doing other things as well. also i have not collected as much data to do averages and such, but my initial results seem to agree with graysky's (~6c lower across the board). it was an extremely hot day today here in houston and my load temps were the highest i ever seen them (even with the ac at 74f!). i find it kind of odd how the temps at idle are all different, but at load cores 1&2 are the same and 3&4 are the same (albeit 2c different between the pairs). i am going to try reseating the heatsink again if i have time later this weekend. i am off to bed now, have to wake up in 3 hours to get ready for the drift event, heh...

 

[graysky]

How long did you allow prime95 to run before taking the screen shots? Those are pretty good temps considering you're running @ 8x400... what is your vcore in the bios? If 1.213V is indeed correct as reported in your screen shot, I'd say have a damn good chip; I can only do 9x329 @ 1.208V as reported by CPU-Z (which is 1.2375V in the BIOS).

 

[jason4207]

The pics of your HS removed look like you may have used too much AS5. I don't think it should spread out to the extreme edges of the IHS if done correctly.

You do have a nice drop there after the lap job, though! Good job!

 

[majinwar]

*Takes a deep breath* Ok big time noob question. I have never built a computer from scratch or tried lapping or anything, but I am going to soon when the new AMDs come out. So I was wondering in all the pictures of your CPUs, what is the small piece of motherboard, it looks like, attached to the back? and where does it come from?

 

[Big Mike]

I agree about the thermal paste, you should put a rice grain sized blob right in the center and typically it should expand out to form a nice symetric circle when you put the HS on. Spreading it around just gets air into the mix which lowers performance.

And Majin, the green part is just a circuit board that the chip itself is mounted to. Most processors have a board of some sort on them, intels just happens to look like a standard printed circuit board like your motherboard. It serves the same purpose as the ceramic packaging on older processors, its probably cheaper to make though.

 

[Peepaw]

Nice work Toeknee. I'm going to do my cpu and heatsink sometimes next week. I hope it comes out that nice.

 

[jason4207]

Yeah the small piece of MoBo (PCB board) is what the actual CPU die, and various resistors/capacitors are mounted. The bottom of the PCB has all the pins, or in the case of 775 socket, all the gold pads that connect the CPU to the rest of the system. The top is then covered with the IHS which is a large piece of metal and it is typically welded on nowadays on these mid-higher end chips. It helps to transfer the heat to the Heatsink, and also protects the CPU die from damage.

 

[shirker]

The top is then covered with the IHS which is a large piece of metal and it is typically welded on nowadays on these mid-higher end chips. It helps to transfer the heat to the Heatsink, and also protects the CPU die from damage.

I read an article somewhere regarding cpu heat spreaders, and the author of it claimed that the cpu IHS isn't designed with heat dissipation and transfer in mind, but simple to protect the cpu cores and components on the substrate. If you think about it, cpu heat spreaders are relatively inefficient at transferring heat, since the heat from the cores isn't evenly distributed across the IHS to allow it to dissipate evenly.

 

[jason4207]

I read an article somewhere regarding cpu heat spreaders, and the author of it claimed that the cpu IHS isn't designed with heat dissipation and transfer in mind, but simple to protect the cpu cores and components on the substrate. If you think about it, cpu heat spreaders are relatively inefficient at transferring heat, since the heat from the cores isn't evenly distributed across the IHS to allow it to dissipate evenly.

Yes...my wording was wrong. The IHS transfers heat to the HS, but it really interferes w/ optimal heat transfer. It mainly protects. Welding the CPU die directly to a Copper HS would be ideal I think, but very impractical.

 

[graysky]

I agree about the thermal paste, you should put a rice grain sized blob right in the center and typically it should expand out to form a nice symetric circle when you put the HS on.

For a single core CPU that may work well, but I'd encourage anyone to read the instruction over at AS5's website when working with dual or quad cores as they are different and require another technique.

The red triangle I drew shows where that tag is on the CPU, remember that on quad core chips, the dies are placed in a different location relative to a dual core, see the instructions on AS5's website for more on this.

 

[jason4207]

^^ What he said.

I could have sworn it was a blob for single and dual core, and a line for quad-core. The AS5 website says different, though.

 

[ToeKneeATX]

i just tried the line application method and my temps went back up to 67c load! i just reapplied it by spreading it out with a cc and they are back to 60c now. so appearantly the line method does not work too well on lapped cpus...

edit: one thing i forgot to mention is that my cpu looked exactly like graysky's while lapping - it took awhile before i was finally able to remove that strip of silver (nickel?) on the ihs.

 

[graysky]

@toeknee: yeah, seems as though these q6600s are particular bowl-shaped! I dunno what to say about the method you're using to spread out the TIM. I think the key is good contact however you manage it. I will also say that AS5 may take several application-trys to get it right.

 

[ToeKneeATX]

well i have always used the cc method of applying tim since i can remember, but since everyone recommended the line i thought i would try it out. just reporting my results, like always every case can be different.

 

[Audioaficionado]

I spread my AS-5 with a cc and my C2D runs relatively cool ~50°C under full SMP_FAH load. I also do a small amount of rotation of the mounted heatsink before I lock it down to help even it out.