Lapped core 2 duo IHS pics

[Daveb2012]

what are you talking about? I put my motherboard in the dishwasher once per month, no problems yet. I will even put my memory in the silver wear basket.

[little_scrapper]

This is a 600 grit finish after a little Weimens.



BTW: It took 8C off my load temps and I have before and after lapping screen shots of TAT and SpeedFan to prove it. AND, The paste hasnt cured yet. I expect another 2C or more drop in temps once the paste has cured!!! I have been reading the others posts, argueing that 10C isnt possible from lapping, and Im wondering "where do these people get thier ideas." People have been reporting good reduction in temps from lapping for years and suddenly someone is going to try to say it doesnt help, maybe 2-3C???? WTF!!! I got 8C drop, this was my first lap.

[laziepig]

It will take out another 1-3°C if I open up my side case... Dax5 must be living outside the sun, that's why he don't get 10°C drop lol

[Dax5]

You guys rely too much on tat and other on board monitoring crap. get an external temp monitor and see for yourselves.

[BNDMOD]

Here is my zalman CNPS 9500 LED lapped.


I have done this job with 100-350-500-700-800-1000-1200-1500-2000 and repeat the progress twice from 800 to 2000...

[Eastcoasthandle]

This is a 600 grit finish after a little Weimens.



BTW: It took 8C off my load temps and I have before and after lapping screen shots of TAT and SpeedFan to prove it. AND, The paste hasnt cured yet. I expect another 2C or more drop in temps once the paste has cured!!! I have been reading the others posts, argueing that 10C isnt possible from lapping, and Im wondering "where do these people get thier ideas." People have been reporting good reduction in temps from lapping for years and suddenly someone is going to try to say it doesnt help, maybe 2-3C???? WTF!!! I got 8C drop, this was my first lap.

Is it me or does one end look higher then the other Looks like the end where the ram rest is lower. Look at the reflection of the letter "P" on the HS surface. Then look at the other end.

[Daveb2012]

http://www.easypckits.com/

I have used them every single time. They have a great kit for $6 that includes a brick of glass and several different grits, they also ship out same day most of the time, these are the grits in a kit:
* 400
* 600
* 800
* 1000
* 1200
* 1500
* 2000

I will post pics of a Scythe Ninja I just finished tomorrow, I'm doing an E6600 when it arrives also.

[celemine1Gig]

temps may improve but not that drastic! where talkin about 10 degrees here, 3 is feasible. but 10? whew! besides the convex part you're talkin about is covered with thermal gunk! its job is to facilitate the transfer of heat on the not so even areas.

I doubt that you read and thoroughly understood my previous posting.

do some reading first before you blurt out anything will ya.

And please for gods sake, don't talk to me like that. I don't need no extra reading on this matter. You ever heard about "thermodynamics"? I for one had exams and practical courses covering this at university.

a 10 degree difference using a hi end hsf versus a watercooling setup is sometimes even hard to reach. if you're idea's so great howcome the boys in intel havent thought about this? lolz

First, I doubt that Intel hires boys to work on the thermal specifications of their products, but rather some highly educated engineers.
Second, there is a major difference between enthusiasts lapping their CPUs to get an extra MHz out of it and engineers planning a production line that earns money instead of eating it. When running the CPU at stock, a lapped IHS and/or heatsink wouldn't help you much, as the thermal dissipation isn't high enough in most cases, to let the difference show. It would be wasted money to get the surfaces all to nearly perfect flatness.

The point that you totally missed is the following: I never said and never will claim that you'll always get at least a 10°C drop after lapping your IHS and heatsink. That would be as much bull as in previous postings of certain persons here.
I did say that in specific situations, where you have a for example concave suface and/or a convex or also concave surface combined in one thermally stressed system, you'll most likely get horrible temperatures, unless the convex and concave part match each other. And the temps will get worse for every more Watt that you pump through the system, when overclocking it for example. Now let's say you have a Core2Duo running at 3.4GHz with 1.5V VCore, cooled by a concave heatsink and covered by an also concave IHS. This would create a "gap" right in the middle of the IHS, directly above the core. The most thermally stressed zone, or in other words, the area where the thermal dissipation per cm² is at its max. The gap has to be filled with thermal material (compound, paste whatever), but you already know this I suppose. The important thing is that the needed thickness of the thermal-compound layer can be for example 4-times higher than in a system where both surfaces are nearly flat (comparing 0.05mm thickness, to let's say 0.2mm). It's just an example, but I hope you get what I want to emphasize with it. This difference can indeed be responsible for delta T's of more than 10°C in systems with a high thermal dissipation. For example you could drop your load temps from 75 to 65°C by lapping both surfaces. 10°C may sound high at first, but in certain component combinations, it is possible.

You can never say in general that something is absolutely impossible, if you haven't thought of any potential circumstances that could perhaps make it possible.

[jws2346]

jws2346,

BTW... What is the "ketchup glass" technique? Googled it but didn't find anything...

Sorry, it's a pretty "low tech" procedure and it probably works with with any cheap TIM or fluid that's got a "thick" viscosity. I use ketchup because it's handy, easy to clean off and relatively cheap. What you do is spread whatever (ketchup, etc) on a piece of glass, mount your chip to it, turn it over and you can see any non contact spots on your chips heat spreader. That way you can see if your chips heat spreader is defective enough to warrant lapping. (ie concave, convex or generally fubared) I still stand on the statement that a 10c drop in temps due to lapping is pretty unbelievable unless 1) your heat spreader is defective (ie convex, concave) 2) you need to reseat your HSF because you blew it the first time, 3) if it's close enough you probably need a better TIM, etc Now, this is not only my opinion, but Google around the "Net", it's the opinion of many computer enthusiast's So the whole deal of one person being right and the whole world being wrong just doesn't wash with me.

[BulldogPO]

There is my way to clean CPU after lapping:


Yes, glass is full of idustrial grade ethanol. I let CPU to lay there for an hour and lean it properly.

[Dax5]

celemine1Gig,

There are alot of other factors that may have affected the result of your study. have you taken into consideration the ambient temp, the amount of load and voltage that you put into the cpu before and after the lapping process? what sort of device or software did you use for this experiment of yours? because as what i have mentioned in my post earlier, on board monitoring sucks. and i say this, because i have used each and every software known to man,,,,, the whole nine yards of temp monitoring if i may call it - and i was unsuccessful. why? because they all rely on a diode, embedded on your motherboard, and temps may vary from board to board. asus being the most nasty.

do you know the difference in temperature reading i get from the on board crap versus the external temp device? you wont belive it.... 25 degrees! yes! and for TAT or coretemp to detect 75 degrees on full load, shouldn't my cpu be shutting down by itself already? i once did an experiment to check my thermal sensor accuracy, i ran my watercooled system and unplugged the water pump until it reached its maximum threshold of 70 degrees, and it never failed to shut down by itself each and everytime before or after reaching 70 Degrees.. This only goes to show that thermal specification of 70c degrees is indeed the maximum temperature at the maximum Thermal Design Power (TDP) value that the processor can handle.

please go out of your way to get an external monitoring kit and post something to support or even give us a glimpse of what you have achieved. belive me, you'll thank me for it

[Funky]

So finally lapped the hell out of my E6600 :P

Used 3M sand paper.
320-400-600-800-1000-1200-2000 grit dry + wet.
No polishing done.

Here is the final result.


E6600 @ 3.24Ghz 1.44v
Temps with Ultra 120 extreme.
With WCG 100% load



With Orthos small fft 100% load ( temps does not change after this )

[celemine1Gig]

celemine1Gig,

There are alot of other factors that may have affected the result of your study. have you taken into consideration the ambient temp, the amount of load and voltage that you put into the cpu before and after the lapping process? what sort of device or software did you use for this experiment of yours? because as what i have mentioned in my post earlier, on board monitoring sucks. and i say this, because i have used each and every software known to man,,,,, the whole nine yards of temp monitoring if i may call it - and i was unsuccessful. why? because they all rely on a diode, embedded on your motherboard, and temps may vary from board to board. asus being the most nasty.

do you know the difference in temperature reading i get from the on board crap versus the external temp device? you wont belive it.... 25 degrees! yes! and for TAT or coretemp to detect 75 degrees on full load, shouldn't my cpu be shutting down by itself already? i once did an experiment to check my thermal sensor accuracy, i ran my watercooled system and unplugged the water pump until it reached its maximum threshold of 70 degrees, and it never failed to shut down by itself each and everytime before or after reaching 70 Degrees.. This only goes to show that thermal specification of 70c degrees is indeed the maximum temperature at the maximum Thermal Design Power (TDP) value that the processor can handle.

please go out of your way to get an external monitoring kit and post something to support or even give us a glimpse of what you have achieved. belive me, you'll thank me for it

Sorry, but this is getting ridiculous. Either you are not able to read and understand, or you just refuse to.
I never talked about software monitoring tools. Hardware monitoring that is used on most motherboards might not be very precise, but the error is consitent. That means if it shows you 10°C lower temps after lapping, then it is 10°C lower than before, even if the temp is not correct in itself. For example if your mainboard reports a CPU temp of 75°C before lapping and in reality it was ~65°C, then after lapping it would show ~65°C and in reality it would mean ~55°C. The error won't change (not much at least). So the temp drop will stay the same!
And BTW, core-temperature readings are much more accurate than you might think. The internal diode is calibrated and in the core2 cpus, the thermal data is processed digitally (less error, due to less influence from other sources; interference!).

Have a short look here for example:

http://www.silentpcreview.com/forums...ic.php?t=34022

Much more information on the net and from intel directly, if I remember correctly.

The reason that your 70°C measured with the external temp probe matched the intel spec is just because the spec says that the surface temp (measured on the IHS, if I remember correctly, and not the chips internal temp) must not pass the specified temp. BUT, this has nothing to do with the internal temp, which is much higher.
Please, don't ever confuse internal with surface temperature! The internal temperature is measured in the very core (embedded in the architecture). Then there is a resistive thermal junction between the core and the IHS and another one from the IHS to the heatsink. Furthermore I think you didn't measure in the center of the IHS. Or did you drill a small channel for the probe?
I don't think you did and that makes it even worse. How should it show an accurate temp? Copper does have a specific resistance like any other material and the core only contacts the IHS in the middle, from where the heat radiates to the outside and through the thermal material transfers into the heatsink. But the max. temp will always be directly above the core and not 1 or 1.5cm next to it. And due to the thermal junction between core and IHS, the temp in the core itself will be even higher.


This is all just physics. It applies to all of us and to all systems, at any ambient temp, anywhere here on earth. I never heard of a single person who really tricked the laws of physics. I never said that I'm referring to special measurements that I did take on a special system. Why? I don't need to. If you think that you can break or bend the laws of physics, go out and publish your results to let experts have a look at them. I really doubt that you'll ever do.

I won't waste more time on this matter, as I'm not a professor or something similar and honestly I doubt that you'll ever want to understand what I wrote. Perhaps others will.
Don't get me wrong, I'm not perfect and I don't know it all. I have no problem when people who know better than me correct me. I even appreciate it. But seeing someone who obviously knows less than me on a special matter, who still doesn't want to understand or learn, but rather wants to insist on being right, is kinda sad.
I mean this attitude might make your world look great, but reality is different.
I'm out. Won't clog up this thread with this senseless discussion anymore
If you have any real, plausible corrections on what I wrote, feel free to post them right away. No complaints from me.

[little_scrapper]

celemine1Gig,

There are alot of other factors that may have affected the result of your study. have you taken into consideration the ambient temp, the amount of load and voltage that you put into the cpu before and after the lapping process? what sort of device or software did you use for this experiment of yours?

I would venture to say that he infact did take ambient into acct, the loading programs he used dont change because the CPU is lapped(thats just dumb to imply), nor do the programs used to monitor the temps.

celemine1Gig,

on board monitoring sucks. and i say this, because i have used each and every software known to man,,,,, the whole nine yards of temp monitoring if i may call it - and i was unsuccessful. why? because they all rely on a diode, embedded on your motherboard, and temps may vary from board to board. asus being the most nasty.

You seem to think that the actual amount of temp drop achieved from lapping a cpu and HS has something to do with a diode??? Thats some crazy logic...hehe. And/or your saying that temp monitor software sucks because of the diode on the mobo??? This is also some crazy logic. What does ASUS's thermal diode have to do with how crappy the programmers were at writing a temp monitoring software??

Your obviously not an engineer, not even a technician I would bet. The diode is a very simple device. It simply allows more or less current to pass through it depending on its temp. The monitors generally measure the voltage, Vbe(voltage base-emmiter), which is based on the collector current of the transistor diode. Being someone who has personally worked with a dozen different kinds of thermal diodes, I can say that series resistance is the most relavent factor in determining how acurate the readings are. Because when measureing then directly I can say that most diodes current at a fixed ambient was within .2C and only around .5C when monitoring tools were not properly accounting for the series resistance, but thats an engineering fault not the diode.

Feel free to correct me if im wrong here its been many years since I had solid state class.

celemine1Gig,


do you know the difference in temperature reading i get from the on board crap versus the external temp device? you wont belive it.... 25 degrees!

What, where is your monotoring device measuring ??? Are you somehow able to measure the cpu die directly???

celemine1Gig,

and for TAT or coretemp to detect 75 degrees on full load, shouldn't my cpu be shutting down by itself already? i once did an experiment to check my thermal sensor accuracy, i ran my watercooled system and unplugged the water pump until it reached its maximum threshold of 70 degrees, and it never failed to shut down by itself each and everytime before or after reaching 70 Degrees.. This only goes to show that thermal specification of 70c degrees is indeed the maximum temperature at the maximum Thermal Design Power (TDP) value that the processor can handle.

Your saying because you cpu didnt throttle at 70C that this proves that this is its Tj Max and its shutdown temp??? All I can say is "wow". you need to reread what you wrote and think about it.

the Tj Max (where throttling kicks in) of my E4300 is 85C. I killed the fan and blocked my HS and let my temp climb until the throttling kicked in. Using speedfan to measure temps and voltage. At 83C the multiplier dropped from 9 to 6 and the Vcore dropped way down. Thats within 2C of where Tj Max is supposed to be. "Check Mate!"

[little_scrapper]

Is it me or does one end look higher then the other Looks like the end where the ram rest is lower. Look at the reflection of the letter "P" on the HS surface. Then look at the other end.

Its just an optical illusion. The base of the HS has equal elevation above the mounting bracked front and back. The front and back edges are very curved, maybe thats why it looks that way.

[joebuffalo]

the Tj Max (where throttling kicks in) of my E4300 is 85C. I killed the fan and blocked my HS and let my temp climb until the throttling kicked in. Using speedfan to measure temps and voltage. At 83C the multiplier dropped from 9 to 6 and the Vcore dropped way down. Thats within 2C of where Tj Max is supposed to be. "Check Mate!"

How do you know it was 83 when it started throttling? (hint: you don't know, you assumed)

[Daveb2012]

Here is a Scythe Ninja I lapped, decreased Idle temps about 3*c:

[little_scrapper]

How do you know it was 83 when it started throttling? (hint: you don't know, you assumed)

I know because TAT gives me a bright red thermal monitoring warning, CPU-Z shows the multiplier AND the voltage drop and speedfan says 83C when this all occures. But you are correct. It is an assumption after all. There is no way to read the temp of a C2D via software. The chip doesnt have the capability. BUT!!! The engineers at intel, people far smarter than you or I, have set them up to start to throttle at 85C. And since my temp monitoring software (SpeedFan) reads within 2C of that target when throttleing actually kicked in I would venture to say that it(speedfan) is roughly acurate.

You can believe me on this because I say I know what Im talking about hehe

[little_scrapper]

Daveb2012,

is that an earlier version of the ninja? I just got mine not long ago and I do NOT have the lower corners cuttouts like your does. Mine has the fins squared off all the way to the bottom.

[fatfry2k]

I finished lapping my heatsink and IHS on my Dell 8400 and now the temperature controlled fan is alot more behaved. Before it would keep spinning up and down making me go nuts. This is a known problem on the Dell forums. I have 100 windows open in Maxthon right now with 30-50% cpu utilization and the fan is steady.

Also, when I was finished I still had a lot of marks left on the heatsink and IHS but people in this thread don't seem to have them. Is it just the pics or do I have to lap longer?

Big thanks to everyone here.

[celemine1Gig]

I finished lapping my heatsink and IHS on my Dell 8400 and now the temperature controlled fan is alot more behaved. Before it would keep spinning up and down making me go nuts. This is a known problem on the Dell forums. I have 100 windows open in Maxthon right now with 30-50% cpu utilization and the fan is steady.

Also, when I was finished I still had a lot of marks left on the heatsink and IHS but people in this thread don't seem to have them. Is it just the pics or do I have to lap longer?

Big thanks to everyone here.

One tip for less scratchmarks: Take a bar of soap and rub it onto the sandpaper before or after putting the water on it. That doesn't really matter. With soap 600grit lapped surfaces look better than dry lapped or only wet lapped 1200 grit surfaces.

[fatfry2k]

I read a guide that used dishwashing liquid. Would that be ok?

[celemine1Gig]

It will perhaps depend on the special kind of dishwashing soap. I doubt that you can generalize the qualification of a whole product-group for this special purpose. I for one found normal bar soap (FA) to work best for me.

[Frostbyte]

I am waiting for some float glass to come in that I ordered from Rockler, and have the sandpaper ready, up to 3000 grit! Can't wait! (funny, I used Norton waterstones to hone my chisels, but it took xtreme computing to get me to the "ScarySharp" method!

[hotrippr]

There is my way to clean CPU after lapping:


Yes, glass is full of idustrial grade ethanol. I let CPU to lay there for an hour and lean it properly.

Doesnt the fluid get trapped under the IHS? Or is the seal totally liquid proof?