about how thick is the aluminium/nickle layer on a TRUE?

[faxon]

so in building on a thread a friend posted a while back about TRUE vs V8, we got the TRUE Black, and i am currently lapping it due to it having a visible curvature/wobble of the heatsink base in the shape of a road speedbump more or less. going to lap it to the copper underneath, and i'm curious how much metal is in between me and my goal. i have lapped the nickle IHS of an athlon 64 before, and that took me 4 hours to lap through and flatten out to a shiny flat finish at 1500 grit. so, lapped TRUE owners chime in! should i go and pick up some rougher grit paper than the 600 im currently starting with and rough it up a bit before putting the mirror finish on it when i hit copper, or is it not worth the hassle. i have already been lapping it for about an hour now

 

[faxon]

nvm just hit copper. will post a pictures log once i'm done

ed: it's 6:30 AM and i'm really tired. uploaded the photos with the comments i will be posting here on my facebook for those who have it. will reupload them to post here later for those who dont.

ed: removed FB link now that i have everything up on imageshack.

 

[RallyMaster]

Thanks for linking us to your facebook.

 

[zagood]

Someone give me his facebook...I have a multi-million dollar proposition for him. My uncle, the crown prince has been imprisoned, and...well I don't want to get into it here. We'll talk directly.

 

[Rubycon]

Is nickle really a pickle with an abundance of electrons? 😉

 

[faxon]

Here's the best of the photos i took of the entire process. I took about 40 in all but most of them didnt come out very well due to bad light conditions and having to take pictures 1 handed using an exilim camera with no tripod.

So, this cooler was purchased for a friend (Shmee here on AT), as a trade for his 9800GTX. His birthday had just past, and since he wanted the best cooler possible for his i7 920 overclock, I offered to lap it for him as a present, since I had lapped coolers in the past and I still had plenty of left over materials.

The issue at hand is simple. Thermalright does a craptastic job of sanding the bases of their coolers, despite making some of the best coolers on the market. For those new to this field, the purpose of flat lapping a cpu cooler like this one is to increase surface contact area between the cpu and the cooler base. it also serves well to sand through the nickle, since nickle has about 1/5th the thermal conductivity of copper. Bellow is an example of what the cooler's base looked like to start with. Full credit to www.benchmarkreviews.com for snapping this image, as i forgot to take one myself before i started. their cooler actually looks a tad better than the one i started with did.

http://img172.imageshack.us/im...120extremebaseclos.jpg




So i started at 600 grit, because i wanted to get this job started before work and see how far i could get. Ended up striking copper after about an hour of dry/wet/dry sanding on each piece of paper i went through
http://img524.imageshack.us/im...043558220282156542.jpg

After an hour at 600 grit i discovered i had one last piece of 220 grit in my modding supplies box, and this is the result. I left to go to ace hardware and then work after this point http://img209.imageshack.us/im...043558260283156542.jpg

After i got home from work, i immediately got out a sheet of 120 grit paper and decided to go to town. this is where i was after another hour with the rough grit. the copper is coming through nice and evenly, but there are still visible disk grinder scours along both outside edges in the nickle. the surface is still uneven as i havent sanded it down far enough to be completely flat yet
http://img300.imageshack.us/im...043558300284156542.jpg

I wanted to save some of the supply of 120 grit paper i purchased for next week, as i was planning on ordering 2 thermalright northbridge coolers for my own case projects, and the finish on the base of these is apparently just as bad as on the TRUE. So i went at it until the single piece of 120 grit paper i used beat it, and took it down to 220 for the rest of the way. all in all this took about 3 hours from when i got back to when i lapped away the last of the nickle, spending about half the time at 120 and the other half at 220
http://img264.imageshack.us/im...043558340285156542.jpg

After i lapped through the nickle entirely, the task at hand became much simpler. the cooler base was more or less flat, and the over all surface area that would come in contact with the CPU was increased dramatically, but there was still room to improve. So i got out the 600 grit paper again, and began the 2 hour process of polishing this baby up nice and pretty. the difference at this stage is that instead of continuing at a certain grit until i reach a goal, i only need to go at the grit im using until i no longer gain any additional smootheness, finish, or surface area. this generally only takes half a piece of sand paper to do at 600 grit. the result is bellow.

http://photos-f.ak.fbcdn.net/hphoto...1043558380286_1565421730_514477_4256242_n.jpg

as you can see, the surface is becoming reflective, but there are still clearly visible signs of surface abrasion. i demoed my Benchmade 930 with the logo reflected to show how clear the surface reflection is at this stage.

After i was done at 600 grit, i opened the lapping kit i had purchased from frozenCPU to test out the effectiveness of their 1200 grit paper at this stage. it worked quite well, but they only gave me a single quarter sheet of paper, so i went until the paper was clogged with copper dust before taking it to 1500 grit. there was little difference between the results at 1200 and 1500 grit, so i didnt bother taking 1200 grit photos. at this stage i decided to grab one of my empty G33K B33R bottles to show how much more reflective and smooth/clear the surface had become.
http://img201.imageshack.us/im...043559100304156542.jpg

After 1500 grit i again went for the FrozenCPU lapping kit, since the hardware store i go to didnt sell anything finer than 1500. the kit had a quarter sheet of 2000 and 2500 grit paper which i used until completely clogged by copper dust. after clogging the 2000 grit paper, the surface was beginning to look pretty well finished, though signs of scratching remained. The surface reflectiveness was clear enough to make the logo work in my Spyderco Delica 4 knife (i collect knifes if you didnt notice yet) clearly legible. the logo is about 3/4ths of an inch across at the longest point.
http://img258.imageshack.us/im...043559140305156542.jpg


After burning out my 2000 grit i took it to the final step. this didnt take more than 5 minutes because the sheet was so small and of such a fine grit. There are unfortunately still signs of surface marring but the surface is 100% flat on my word. i tested it for visible light being let through using the flat side of my Delica 4, which has a perfectly flat mirror lapped finish itself. I brought out the 930 again to show the difference between before and after.
http://img382.imageshack.us/im...043559180306156542.jpg
close up http://img79.imageshack.us/img...043559220307156542.jpg

the mess i made. pretty straight forward. i ran out of space for the old shitty and dying thermaltake fans in my trash. no photos of that project yet, it wont be done until i recieve the IFX coolers im ordering on saturday

http://img300.imageshack.us/im...043559260308156542.jpg
http://img144.imageshack.us/im...043559420312156542.jpg
http://img511.imageshack.us/im...043559460313156542.jpg


All in all the job took me about 9 hours. I would like to thank Fry's for funding the project, as well as Ace Hardware and www.frozencpu.com for supplying the required materials, and Gatorade, Bawls, and Pizza-My-Heart for fueling me through the night! Happy birthday Shmee!

 

[videogames101]

Nice work!

 

[Rubycon]

9 hours! Wow! Before and after temps? Find a machinist that can CMM it for you then you'll know for sure if your TRUE base is really in fact, true. 😉

 

[BonzaiDuck]

Originally posted by: Rubycon
9 hours! Wow! Before and after temps? Find a machinist that can CMM it for you then you'll know for sure if your TRUE base is really in fact, true. 😉

I didn't read through all of his text, but I ran through the pics.

I had ordered the custom-lapped TRUE from SVC when they offered it, calling them to discuss it. I said I thought they must've used a custom-built jig to do it to reduce the time and labor, and they say they did.

I later did the same thing myself with an Ultima-90. My "jig" was very crude: a couple oak blocks cut to size and a small "quick-release" clamp.

The only reason for using the jig is to get the flat surface started on the purposely-designed "roadbump" of those TR coolers -- you don't want the cooler sitting cockeyed. Once you see bare copper along the area where the roadbump had been, you don't need the jig anymore.

From experience firsthand and from the talk with the SVC people, you eventually grind off less than 1mm at the highest point along the "roadbump" until the entire heatsink-base is flat and "all copper."

I think it took me more like 4 or 5 hours while watching Daffy Duck cartoons with my friend Kim-Sung-[mentally]-Ill.

Half the improvement derives from "flatness," and the other half (as Ruby-doo may have implied) from getting rid of the nickel-plate.

 

[faxon]

Originally posted by: Rubycon
9 hours! Wow! Before and after temps? Find a machinist that can CMM it for you then you'll know for sure if your TRUE base is really in fact, true. 😉

no before temps unfortunately. i might be getting another TRUE for myself though this summer, at which time i will do a before and after. given the performance of the prolimatech cooler though i will probably end up getting one of those by the time i get anything which would warrant more than an OCZ Vendetta 2 or an S-1283 (pictured, look for the orange 120 fan).

as for after temps, well shmee sort of went to sleep before i got off work, so now i dont have any pants besides my dress slacks i wear at work, or some shorts that i need to dig out of a box. we were gonna push his OC up from 3.06 GHz to 4 tonight when i got off, but given that he was asleep when i got there, will have to wait until tomorrow. i will let him post the results when were done, since im not gonna wait around for 3 hours while OCCT runs lol


in response to bonzaiduck, the method i use is a bit more careful and meticulous. as you can see from the pics, i do all my sanding on a piece of 12"x12" glass, and i use the fact that sandpaper grit wears out over the period of it's use to help polish off the base later in the process. i generally start by checking if the cooler base is flat enough to just start sanding it, and if not i start it out really slow, letting the glass do the job of keeping everything flat and even. the best way to check for flatness over a 3 dimensional surface area is actually not by using a razorblade (this only lets you know if the surface is tilted inward or out), but by using another flat surface. the delica 4 knife in one of my pictures above is flat lapped at the factory to a pretty nice shine, and the surface area of the blade that's flat is great enough to gauge whether the area is really flat or not. the BEST way to tell though is if you have a small piece of mirror. usually mirror glass is only reflective on one side, but it's flat as the glass on the side that's non reflective. you can simply put the cooler down and check for light showing through. alternatively, a dark polished rock (Granite or look alike) countertop or slab will also get the job done, so long as the chinks in it dont extend from one end of where the cooler is set down to the other.

if the surface of your cooler starts out so rounded that it rocks in all directions, the best way to go about getting the lap started is by finding a way to mount it so the cooler doesnt move around at all, using a laser level to check and make sure it's level, and then running the sand paper + flat surface it's attached to underneath the base of the cooler until you create an area flat enough to keep the cooler from rocking while you sand it by hand. this generally isnt very common though, and the only coolers where this would be an issue are usually ones which you also dont want to lap in the first place because they are already finished properly, and the curvature of the base is intended to put light pressure in the center of the cpu IHS to make it bend inward, spreading the solder under the IHS across the CPU die more evenly, which increases cooling performance

 

[Shmee]

after final installation with 2 fans, and upped to 3.6 ghz, idles between 22 and 27 as of now. Vcore =1.3 aprox. now for some stress testing.

so far I am pretty impressed!

 

[faxon]

wow nice dude. load temps any higher at 3.6 vs the 3.2 you had it at when i left?

 

[Shmee]

lower actually, now that both fans are there and properly. about 43 load with small fft's. will up oc more later.

 

[faxon]

holy shit lol. 43c load at 3.6ghz? thats drool worthy. you will probably break 4GHz with temps like that. you're making me want to get my quad before i can actually afford it 🙁. my poor e5200 is suffering at 1.39v (1.36 CPUz) to hit 3.2ghz at 55c and you're getting 10c lower temps on a quad with a similar fan configuration at near stock voltage :|

 

[BonzaiDuck]

in response to bonzaiduck, the method i use is a bit more careful and meticulous. as you can see from the pics, i do all my sanding on a piece of 12"x12" glass, and i use the fact that sandpaper grit wears out over the period of it's use to help polish off the base later in the process. i generally start by checking if the cooler base is flat enough to just start sanding it, and if not i start it out really slow, letting the glass do the job of keeping everything flat and even. the best way to check for flatness over a 3 dimensional surface area is actually not by using a razorblade (this only lets you know if the surface is tilted inward or out), but by using another flat surface.

I do all that. 1/2"-thick glass-plate. For the worn-out sandpaper, though, I just keep washing the particles off it. You could almost start with 240-grit; by the time you're finished, the worn-down sandpaper is probably more like 400 or 600.

Very fine scratches don't matter. I use nano-diamond paste, and particles fill those scratches with no ill effects -- possibly, even improvement.

As I said, you could start the sanding on a glass plate as we do, but with the convexity, the flat spot might develop at the wrong angle. But with only 1mm to grind down, I'd think that you can see how to keep it on an "even keel."