If you must use a polish to get the contact surface to a mirror finish, you should really use a cleaning product to then remove the polish to prevent it from harming the thermal dynamics of the system. Remember, polish is not designed for it's ability as a thermal transfer medium
Not only that, but depending on the polish, some of the chemicals can break down any thermal interfaces like AS5 that you use
Originally Posted by Saracen
Good god god , you dont't need it as smooth as that , if the surfaces are polished to a 'real mirror finish then you will find that after polishing you temps will increase ' gleaned of x treme' correct me if I am wrong . \\
By the way --great thread - I spent a lot of time on lappitary
Yes that is certainly true. The Polish can be highly volatile and mixing the Polish with vodka or any other alcohol/Germans is not advisable, it's an explosive mix.If you must use a polish to get the contact surface to a mirror finish, you should really use a cleaning product to then remove the polish to prevent it from harming the thermal dynamics of the system. Remember, polish is not designed for it's ability as a thermal transfer medium
After polishing I cleaned the surface with isopropyl alcohol and then used a nice trick*. I only did the polishing on the chipset heatsink to see what sort of finish I could achieve with non-industrial products such as the Sainsbury's Silver polish I found in the cupboard under the stairs whilst visiting my mum.
*For the heatsink I used a metallurgist's trick for cleaning. What you do is run the hot tap as hot as you can stand it, hold the heatsink in the palm of the hand - surface up, pour a small amount of detergent onto the surface, then use the thumb of the same hand to gently rub the detergent in a circular motion, after that put it under the hot tap and gently rub until the detergent is gone ( couple of seconds) then quickly hold the surface up to your mouth and blow (about the same weight as if you were trying to gently blow on a small camp fire to get it started) and the water will run off leaving the surface unblemished.
All of you have no idea how much it pains me to to see a "flat" surface that has scratches in -eg WannaPiEcE pic. Yep, it's more than enough to get cpu/heatsinks to reduce the temperature by a few °C and a "mirrored finish" is not needed, however, it's just one of those things that as a professional (and sad perfectionist) I just can't look at without thinking /joke "ffs they could of at least have done a 1micron polish"/joke
My first job was in a lab where I was quickly introduced to the deity Micraw - the God of Flat and Scratch Free Surfaces. I had to rely on his/her/it's favour and use a sacrifice to this mighty Deity to ensure that my microsections** (and my colleagues) remained free from the dreaded surface imperfections.
Each year a sacrifice was made, which was in the form of an effigy of the lab manager constructed from all the materials a metallographer (yeah look that one up!) uses for microsections. Lollipop sticks, plasticine, bakelite , plastic resin and hardener, little loops to hold small pieces of pcb - seen in the pic below, plastic clips, double sided sticky tape etc. This effigy then resided in the lab, in peace for one month, until it was brought out from the darkness and sacrificed, dying from a huge hammer blow administered from a copper/horse hair head Thor hammer.
**
And yes we were bored sometimes!
"Reality is what it is, not what you want it to be." Frank Zappa. ----------- "The invisible and the non-existent look very much alike." Huang Po.----------- "A drowsy line of wasted time bathes my open mind", - Ride.
You are right, the best temps will be achived with a rough contact area.Good god god , you dont't need it as smooth as that , if the surfaces are polished to a 'real mirror finish then you will find that after polishing you temps will increase ' gleaned of x treme' correct me if I am wrong . \\
By the way --great thread - I spent a lot of time on lappitary
not sure about the rough contact area.
the way i see it is you need to be able to use as little thermal compound as necessary, the thinner it can be spread out the better so the heat has less distance to travel up from the cpu to the cooler.
the rougher the surfaces the bigger the distances will be between both, kindoff like valleys \/ so to me smoother and flatter the better.
i also wouldnt use any polishing compound that wouls leave a residue on the lapped surface. minimise the layers.
thanks guys, i'm learing nice new techniques as well!
Eer, no.
I have 3 qualifications in electronics and i can assue you it's far better to have a rough area between the 2.
Gives you a better surface area and means the compound stays in place easier.
This is why Cpus come with a rough top to them.
Rough is a relative term. However a mirror finish on both surfaces will be detrimental to heat transfer with thermal compound because it will form a layer between the two surfaces. Have a slightly rougher surface is better because there will be more metal to metal contact points between the two surfaces - the gaps between adjacent contact points will be shallow so the compund will be very thin there. Certainly the smoother the surfaces, the less thermal compound you need.
(\__/)
(='.'=)
(")_(")
Been helped or just 'Like' a post? Use the Thanks button!
My broadband speed - 750 Meganibbles/minute
It don't matter if you have 30 qualifications, its wrong
2 flatter surfaces = more direct contact area. Less space for air molecules and other impurities.
If you lap both surfaces properly, they will 'hold' together when you pick one up; briefly at least, in the same way that its hard to pick a CD up off a mirror.
When you have rough surfaces, you cause more of these pockets of air and other impurities to sit between the heatsink and the CPU.
Yeah, TIM helps stop this by filling these, but its basic physics that every time you transfer heat from one object to another, it wont be a 100% efficient transfer (and more importantly, different materials have different thermal absorption and thermal radiation rates) . The more items the heat has to be transferred through, the more efficiency that is lost.
Lapping Guide
Explains it in more detail.
CPU's come with a rough top for a lot simpler reason: Cost. There is no point lapping a CPU when its not going to make a difference for 99.9% of users.
Now, moving air over a surface to dissipate heat is an entirely different ball game, and is almost always what people get confused with
I have electronic qualifications too, and sadly thermal transfer is rarely touched on in detail (atleast in the ones I did)Heck, it wasn't until some of the more recent stuff that Fourier's law and Newtons law of cooling was covered
No, having rough surfaces causes the exact opposite of this, even if its on a microscopic level. See the link I gave for picture examples.Have a slightly rougher surface is better because there will be more metal to metal contact points between the two surfaces
Last edited by Agent; 03-11-2007 at 09:27 PM.
It don't matter if you have 30 qualifications, its wrong
2 flatter surfaces = more direct contact area. Less space for air molecules and other impurities.
If you lap both surfaces properly, they will 'hold' together when you pick one up; briefly at least, in the same way that its hard to pick a CD up off a mirror.
When you have rough surfaces, you cause more of these pockets of air and other impurities to sit between the heatsink and the CPU.
Yeah, TIM helps stop this by filling these, but its basic physics that every time you transfer heat from one object to another, it wont be a 100% efficient transfer (and more importantly, different materials have different thermal absorption and thermal radiation rates) . The more items the heat has to be transferred through, the more efficiency that is lost.
Lapping Guide
Explains it in more detail.
CPU's come with a rough top for a lot simpler reason: Cost. There is no point lapping a CPU when its not going to make a difference for 99.9% of users.
Now, moving air over a surface to dissipate heat is an entirely different ball game, and is almost always what people get confused with
I have electronic qualifications too, and sadly thermal transfer is rarely touched on in detail (atleast in the ones I did)
No!
You really are wrong, and not sure if you noticed but when purchasing metal it comes flat so they have actually done out of thier way to create the rextured top.
Wheather it's rough or not when done properly thier shouldn't be any air between the cpu and the heatsink, thats what the compound is for.
A 'fine' abrasive serface will be the best for heat transfer, thats the way it's taught in colleges and i'm sure my electronics tutor knew more than you. I think i'll take his word.
I think the purpose of lapping is to make the surface of the CPU heat sink as flat as possible and not concave.
As you can see a decent job of lapping the CPU will leave minute scratches across the surface which is normal and indeed what the thermal interface material is designed to be used for however the entire purpose of lapping the CPU heat sink of the heat sink Fan is to make the surface as flat as humanly possible thus eliminating any differences in height which can lead to a bad conductivity between the CPU heat sink and the heat sink fan.
I'm sorry, but I entirely disagree. Each to their own though, it seems obvious you're not even willing to debate it
Purchasing what metal?
Im totally lost by what you've said here!
Impossible. Compound or no compound and lapped surfaces, either way you will never get rid of air molecules between 2 surfaces with the methods we are using.
Its really not the best way. If you are so confident, read the link and explain to me the reasoning.
Also the "i'm sure my electronics tutor knew more than you" is incredibly pompous and immature. For all we know, your electronics tutor could have not known his subject.
You don't know my education, you don't know where I work and you don't know my background in this area, yet you're willing to tell me that you know someone who knows more? If you did have a point, right or wrong, you've already gone half way to loosing it.
Just for your reference, I'm a teacher too, and I teach above college level - but I have no intention into getting into a debating match on who has the most education behind them. Its Irrelevant, the only issue here is the question in hand. If the people with more education never made incorrect statements, we wouldn't have politicians about.
Electronics isn't thermal physics, and most electronics teachers usually have a strong understanding in electronics.
Like I said though, there is a difference between dissipating heat via air based cooling, and heat transfer between different materials. Getting the two confused is very easy.
Now, if you want to make a point / debate which reason is right, lets try and do it with a bit more theory behind it than "I was told this and he knows more than you, ner ner ner".
Regards
Agent (04-11-2007)
Good sound theory > bits of paper (*), as any sensible academic knows
And I've just did some checking and reminding to confirm my suspicions. Maximising contact area is good, but it has to be a perfect contact. Two rough surfaces created through semi-random grinding are highly unlikely to touch directly, requiring transfer through the medium between the surfaces. The more different mediums of different types heat has to get through, the worse the transfer.
It's not a simple matter to understand, not by any means. Especially when maximising raw surface area is normally a good thing o.O
(*) That's all qualifications are, when all is said and done
Agent (04-11-2007)
There are currently 1 users browsing this thread. (0 members and 1 guests)
SCAN.care@HEXUS
Posting Permissions
LinkBack URL
About LinkBacks
Reply With Quote
