Dont know how it would work. But how bout some liquid metal blocks?

Originally posted by Waddles
I've seen direct-die wc setups, where the water was touching the die directly, and if my memory serves the temperatures and heat transfer was astounding
Click to expand...

Cool. Now give me one of those! Lol

Heh, seriously though, I guess there are troubles since water tends to mess up quite a lot of materials, especially metals. Anyway, according to that chart, water is 5.61 Watts/cm-°C versus silver being 4.18 Watts/cm-°C, which sounds like a small difference, but on such a heat sensitive thing, would have a huge effect, plus it adds up in a hurry anyway. (This is bearing in mind that the whole setup you mentioned was water cooled, where I'm just considering what would happen if you had water simply acting as the transfer from core to heatsink rather than paste, obviously in a full water cooled system you get better results.)

One of these days I have to get around to getting a water cooling system. I'd just love the extra ocing leeway you can get with that. Unfortunately, with this stupid palomino, it's just not worth it.
 
Originally posted by Nazo
EDIT2: Btw, Kadarom Douhrek, you said that weight and density are not related, but that isn't by any means true. The reason that they are is that you will tend to have to have the substance be a certain size, so that means that you will have more mass when you have a denser substance. Effectively, when density goes up, weight goes up unless you decrease the size exactly proportionally. Did you perhaps mean to say density and conductivity are not related?
Click to expand...
Nope said what I meant, and I said atomic weight, not weight.
AggieMEEN mentioned using Ununhexium because it has the highest atomic weight, I was correcting that.
 
there is no real reason to use water as a means of attaching a heatsink to the die. If you're gonna put water on the cpu, may as well skip the heatsink all together and do direct die water cooling.

Obviously this makes the entire discussion of what metal transfers heat the best because the point of the metal is to get the heat from the die to the water inside the loop. If you put the water directly on the die, that work is already done.


Obviously this type of setup will outperform any other possible method. The only problem you face is sealing up the die and the area around the die so that water cannot escape. Not really all that difficult, and certainly not impossible, but probably a little tricky and definately pretty scary! :D
 
Well, part of the point being that a water cooling system requires more power, maintenance, etc as well as being more expensive, but if you just had a setup where you simply used water as thermal paste for a more normal heatsink, then it's a very cheap solution. Actually, something about what you said made me think up yet another even more interesting idea. A sort of heatsink designed to mix the normal heatsink method with watercooling. Cheap to buy (may be cheaper than a normal heatsink with copper,) no more power than a pure aluminum/copper heatsink (watercooling has pumps/etc) and I'd imagine would cool better by far.

Image removed to save web space

Here's my idea. Kind of protected on the bottom with some rubber to keep it tight. The screw at the top is surrounded by rubber so there's no leaking (I've seen things that worked like this) and you can pull out the screw to put in more water. Other than that it works like a normal heatsink minus the thermal compound. Ok, it's just an idea, probably wouldn't work on normal motherboards since it has to go over the cpu and all, but, I just thought it was an interesting idea I wouldn't mind seeing around. I suspect this might cost less than normal heatsinks in fact since it has less metal.

EDIT: Oh, and normal water cooling has a metal thing that attaches with thermal compound, so that's still a bit of a weak point. Though like waddles says there is apparently one or two exceptions to that (I haven't seen any, so I think not many.)Image removed to save web space. Image removed to save web space.
 
Originally posted by AggieMEEN
Some of this stuff is related to what I'm doing my graduate research on. There were some questions about why this liquid metal material would be harder than regular materials. If memory serves, the Hall-Petch equation, which relates grain size to yield stress, tells us that

Sy = So + Ky / ((Do)^(1/2)) Where Sy is the yield stress, So is the friction stress, Ky is a material constant, and Do is the average grain diameter. As grain diameter / size decreases, yield stress will increase. This is a good thing, as it could mean bullet-proof aluminum and other novel materials for industry. The problem is that when grain sizes get to the nanometer scale (well...below about 20nm), the Hall-Petch equation begins to break down, and quantum effects pick up. This may well be Coble Creep, atomic diffusion. In this way, it is possible to create "superplastic" metals. But getting back to grain size vs. hardness: dislocations have a harder time moving across grain boundaries, especially when the grains are oriented differently.

I don't do much work with amorphous metals, but my guess would be that, with no grain structure at all, dislocations would have one hell of a time moving about. This may be the source of the enhanced hardness, but I could be wrong.

Now, as for thermal properties, I'm not as up to date on. Someone in materials confirm the below, please?

This "liquid metal" material is amorphous, in that it has no grain structure, hence no crystal structure. Assuming that the atoms are closer together than in a normal crystal structure, the melting point may be lower than a standard BCC, FCC, or HCP structure. (Talking about the bond energy curve; energy on the Y axis, radius / distance on the X axis. The curve starts high at low radius, dips to a minimum -- the optimal atomic bond distance where energy is minimized -- then increases due to attractive forces). And since thermal and electrical properties are also dependant upon grain structure (and alignment, I think) this material may prove to be a poor conductor of both heat and electricity.

If I remember correctly, amorphous metals are commonly used in the electrical equipment found on power transmission equipment. Transformers, I think. They make good dielectric materials.
Click to expand...

I am doing some work with bulk metallic glasses right now in school, making a ceramic/BMG composite material. The Hall-Petch equation, as you stated, is not valid here. Since there are no grains or dislocations in the metal, strength is dependent solely on interatomic properties. But the stuff is definitely not "bullet-proof" considering the way that it deforms. Plastic deformation is a result of formation of slip planes, sometimes as few as one, that propagate throughout the part until it basically tears apart.

About the thermal properties... I don't have any numbers, but I know that even copper-based BMGs have relatively low thermal conductivities. One grad student has actually been testing thermal conductivities of a copper BMG, so I can ask him for the data he has gotten, if I remember. There are other BMGs that are iron, titanium, and zirconium (the one I'm working with) based, which I'm sure have even lower conductivities just because of the base metals.

And yes, some transformer cores are made of amorphous materials, but not because of dielectric properties. The benefit is their magnetic properties which result in more efficient energy transfer (smaller B-H hysteresis loop) compared to standard iron-based or ferrite cores. Check out MetGlas , they sell transformer cores made of several amorphous alloys.


Finally, it was mentioned before that the alloys sold by LiquidMetal are used in devices like cell phones. It's used in the metal covers which are very thin. Golf club faces are also very thin. This is a necessity because there is a critical cooling rate that is needed for the liquid to solidify as a glass, which cannot be achieved in thicker parts. So even if there was a BMG with suitable thermal properties for a waterblock, the processing would still be practically impossible.
 
quick question... how dose haveing a non crystalin structure apply to a material holding a magnitic field???

dose melting a material in a magnitic field , and letting it cool in the same field allow the molocules to arrange with polerization?... and if so... how would the arrangeing the molocules as such affect heat transfer?

thore
 
Originally posted by zer0signal667
I am doing some work with bulk metallic glasses right now in school, making a ceramic/BMG composite material. The Hall-Petch equation, as you stated, is not valid here. Since there are no grains or dislocations in the metal, strength is dependent solely on interatomic properties. But the stuff is definitely not "bullet-proof" considering the way that it deforms. Plastic deformation is a result of formation of slip planes, sometimes as few as one, that propagate throughout the part until it basically tears apart.

...
Click to expand...

DIdn't mean for my previous post to be so confusing. The work I'm doing right now is in severe plastic deformation of bulk materials, not BMG. Sorry if my statement about Hall-Petch sounded as if I was making a generalization to all materials:eek:

But the Hall-Petch eq. does apply in bulk materials that have crystal structures, and it is possible to modify the crystal structures to yield extremely interesting properties (hence the bullet-proof aluminum statement). (edit: Gah. modify the grains. not crystals!)

Hall-Petch breaks down at very small grain sizes, though. And, as you very correctly pointed out, is not to be used to describe amorphous materials.

Edit: Here is some information about some of the interesting properties materials display at the nanoscale.
 
Amorphous solids would decrease heat transferring performance if anything. The stuff may be harder but hardness has nothing to do with heat transferring capabilities.

Originally posted by thore
quick question... how dose haveing a non crystalin structure apply to a material holding a magnitic field???

dose melting a material in a magnitic field , and letting it cool in the same field allow the molocules to arrange with polerization?... and if so... how would the arrangeing the molocules as such affect heat transfer?

thore
Click to expand...

Yes, that is how permanent magnets are made, and no, it really shouldn't, at least not significantly.
 
hey...wanna take a look at this and some research before you try and rubbish me?
Click to expand...
Well, diamond is a rare-earth material. I didn't say it was something else entirely, i just forgot what it was, and was asking if you were kind enough to support your claim with evidence... sheesh.
 
Originally posted by M4d-K10wN
Well, diamond is a rare-earth material. I didn't say it was something else entirely, i just forgot what it was...
Click to expand...

Reference.

Sorry, diamonds may be rare (in the earth), but they are not rare-earth materials.
 
Originally posted by Nazo
Well, part of the point being that a water cooling system requires more power, maintenance, etc as well as being more expensive, but if you just had a setup where you simply used water as thermal paste for a more normal heatsink, then it's a very cheap solution. Actually, something about what you said made me think up yet another even more interesting idea. A sort of heatsink designed to mix the normal heatsink method with watercooling. Cheap to buy (may be cheaper than a normal heatsink with copper,) no more power than a pure aluminum/copper heatsink (watercooling has pumps/etc) and I'd imagine would cool better by far.

Ultimate_in_Cheap_Cooling.png


Here's my idea. Kind of protected on the bottom with some rubber to keep it tight. The screw at the top is surrounded by rubber so there's no leaking (I've seen things that worked like this) and you can pull out the screw to put in more water. Other than that it works like a normal heatsink minus the thermal compound. Ok, it's just an idea, probably wouldn't work on normal motherboards since it has to go over the cpu and all, but, I just thought it was an interesting idea I wouldn't mind seeing around. I suspect this might cost less than normal heatsinks in fact since it has less metal.

EDIT: Oh, and normal water cooling has a metal thing that attaches with thermal compound, so that's still a bit of a weak point. Though like waddles says there is apparently one or two exceptions to that (I haven't seen any, so I think not many.)
Click to expand...

Water is not a very good thermal conductor compared aluminum or copper. I suspect you would probably get some localized boiling in that chamber, or just extreme overheating since the water may not boil in a sealed container. I think the only way to implement direct die cooling is with flowing water.
 
Originally posted by M4d-K10wN
Yes, a polymer metal would simply own! You could even use it in mixture with a glass metal; to go accross the structure, and prevent it from shattering.

And no, I wasn't talking about heat. I was talking about kinetic impact. I don't see why a glass metal would be any better or worse at heat conductivity than a crystallized counterpart.

Say, you make a small plate out of a glass metal. Then drop a steel ball on it. It would bounce as if it was made of rubber.
Click to expand...

What exactly do you mean by "polymer metal"? By definition, a polymer is a material consisting of chain-like molecules of repeaitng units. There's not really a way to create something like that with a metal, since metallic bonding is not really conducive to chain-like molecules.

The mixture you are thinking of is a composite, and it's been done, and that's what I'm working on creating right now as a senior design project. The glass will be reinforced with a ceramic however. And to make it clear, metallic glasses do not shatter like regular silicate glasses do. Metallic bonding gives it much better elastic behavior compared to brittle ceramics which are full of defects that open up and propagate easily.

The glassy metals are worse at thermal conductivity mainly because of the materials they are made of. Simply stated, there aren't any alloys that conduct heat well that can also remain amorphous upon solidification at any reasonable rate. In order to solidify copper or aluminum as a glass you would need cooling rates on the order of trillions of degrees C per second. Some of the better bulk metallic glasses can be cooled at rates of around 10C/s or lower.

And yes, a steel ball does bounce very well on metallic glass due to its unique elastic properties. My professor/advisor has a demo model of exactly that, she loves to show it off. She drops the ball on a variety of plates- steel, aluminum, titanium, and a metallic glass, and it bounces so much longer on the glass, it's pretty cool.
 
Originally posted by zer0signal667
Water is not a very good thermal conductor compared aluminum or copper. I suspect you would probably get some localized boiling in that chamber, or just extreme overheating since the water may not boil in a sealed container. I think the only way to implement direct die cooling is with flowing water.
Click to expand...

According to previous posts, water is a better thermal conductor than silver. And boiling? It would have completly crashes and melted the CPU long before it hit boiling temperatures (100C) anyway (AMD says to go no higher than 70C or so, and good luck getting your CPU that high without it crashing first anyway.) But, hey, don't take my word for it, look for yourself at the URL posted earlier:
http://www.mse.arizona.edu/classes/mse222/1997_diamond/thermal.htm

Look at the top of the list. Water is the fourth item, two above silver which is above copper.

Oh, and if it weren't a good thermal conductor, then why would people be using it for cooling? Obviously it wouldn't work very well, now would it? Since it has to be air tight and all that, if water were so bad, it would make more sense to use air or even some liquid metal like mercury. Logic.

EDIT: I'll go ahead and give you the numbers just to save some time and trouble. Water has a thermal conductivity of 5.61 W/c^3, silver 4.18, and copper 3.80. So water is 34% more conductive than silver, and 48% more conductive than copper. I'm not even going to bother with aluminum. There's a reason that the calorie is based on water. Anyway, if you read over the part about why diamond is apparently sometimes called "ice" due to it's thermal conductivity, you'll understand why water always seems to feel at least a little cool even on a hot day.
 
Originally posted by zer0signal667
There's not really a way to create something like that with a metal, since metallic bonding is not really conducive to chain-like molecules.
Click to expand...
Five years ago, a something similar was said about glass metals. It's not impossible, we just don't know how to do it yet :D. Oh, and yes, the composite is a glass metal with some crystallic structure reinforcing it. And just glass metal would shatter on impact, if the impact carries enough power. Where a regular metal would just bend, this stuff breaks.
 
Lol, well, my way maybe costs the end consumer $15 or 20 is the difference. I don't want to even IMAGINE what that much silver would cost.

However, having a silver heatsink sure would be an interesting subject for discussion at lan parties, ne? ^_^ Oi, I can just imagine....


BTW, about the whole metals bouncing off the glass metal, that website for the liquid metal (see first post) has a demo video of just that for those interested.
 
I skipped a couple pages in this post, but has anyone mentioned the idea of doing liquid convection (I say convection because you'd need a pretty strong pump) using mercury or galladium?

Mercury of course is liquid at room temperature. Galladium isn't, but becomes liquid at about 35 degrees celcius.

THAT would be pretty neat :). Maybe not efficient, but neat.

Erik
 
Originally posted by Nazo
water is a better thermal conductor than silver.
Click to expand...
I know you're just quoting from someone else's link, but you should always double check...
http://www.visionengineer.com/ref/physical_properties.shtml
http://www.hukseflux.com/thermal conductivity/thermal.htm
Water's thermal conductivity is 0.6 W/mK, a tiny fraction of silver's or copper's.

Oh, and if it weren't a good thermal conductor, then why would people be using it for cooling?
Click to expand...

Water is used because it's cheap, non-toxic, and has a high specific heat (it can absorb a great deal of heat without its temperature rising significantly). That fact that water isn't a good thermal conductor is why water block makers put so much time and effort into increasing turbulence and surface area - the heat needs to be physically moved with water flow instead of being conducted through the water.

if water were so bad, it would make more sense to use air or even some liquid metal like mercury. Logic.
Click to expand...
Mercury has a low specific heat (can't transport much heat), a high viscosity (can't be pumped well), and is poisonous and expensive.

I'll go ahead and give you the numbers just to save some time and trouble. Water has a thermal conductivity of 5.61 W/c^3, silver 4.18, and copper 3.80.
Click to expand...
Again, it's best to double check your sources. Oh, and the units are W/mK (watts per meter*Kelvin), or similar (power per lengthtemperature).

you'll understand why water always seems to feel at least a little cool even on a hot day.
Click to expand...
Water feels cool due to evaporative cooling.

Originally posted by Dem
has anyone mentioned the idea of doing liquid convection (I say convection because you'd need a pretty strong pump) using mercury or galladium?
Click to expand...
You mean gallium? Anyway, won't work, specific heat is too low, viscosity is too high.
 
There has been some talk about water and its thermal conductivity.

It's early, and I need some coffee, but I seem to recall that we use water instead of air not because of water's thermal conductivity, but because of its convection coefficient.
 
Originally posted by HeThatKnows
I know you're just quoting from someone else's link, but you should always double check...
Water's thermal conductivity is 0.6 W/mK, a tiny fraction of silver's or copper's.
Click to expand...

I was just thinking of an idea and throwing it out as it came to me. Why would I sit there going around and checking everyone's links ultra-carefully to make sure they weren't wrong.

Again, it's best to double check your sources. Oh, and the units are W/mK (watts per meter*Kelvin), or similar (power per lengthtemperature).
Click to expand...
Incorrect. It's whatever works. You seem to think science is a very linear specific thing where nothing can be done differently. Basically all that the other site I cited the numbers from changed was the size measurement. Kelvin is based on celsius, merely adjusting for absolute zero being 0 rather than a huge negative number and being better suited to work with gasses and other things. And I hope I shouldn't have to tell you that the centimeter IS the meter, merely scaled down. After all, we are talking about heatsinks, not giant blocks of metal, so I thought that the unit there was more fitting and didn't feel the need to convert or find another source just to change that.

Anyway, so that idea won't work. We'll just stick with the silver alloy heatsink then. It was just an idea.


BTW, a material that melts at 35 just won't do. It needs to be liquid at room temperature or you have all sorts of problems. For one, it would have to melt almost all of it before any sort of flow can get started while in the meantime it would have issues with pooling/etc causing heat to not be transfered far. For another, there's no reason it's impossible a person could want to get their CPU below 35. Heck, I did with just my cheap HSF when I don't overclock so much and turn all my fans on.
 
Originally posted by Nazo
Incorrect. It's whatever works. You seem to think science is a very linear specific thing where nothing can be done differently
Click to expand...
You can use whatever units you want, but they need to be the right kind and in the right places. Hence my saying "or similar (power per length*temperature)". The "W/c^3" that you used doesn't correspond to thermal conductivity - it would be sorta like stating your CPU speed in cubic-yards or your age in Btu/furlong. :)
 
Originally posted by HeThatKnows
... it would be sorta like stating your CPU speed in cubic-yards or your age in Btu/furlong. :)
Click to expand...

I hear the Millenium Falcon did the Kessel Run in three parsecs!:D
 
Originally posted by M4d-K10wN
Five years ago, a something similar was said about glass metals. It's not impossible, we just don't know how to do it yet :D. Oh, and yes, the composite is a glass metal with some crystallic structure reinforcing it. And just glass metal would shatter on impact, if the impact carries enough power. Where a regular metal would just bend, this stuff breaks.
Click to expand...

Metallic glasses were discovered more than 5 years ago, first of all. And do you understand how metallic atoms bond to each other? And how polymeric chains are bonded?
Where are you getting your information that metallic glasses shatter whereas a "regular" metal bends? There are crystalline metals that are very brittle, and amorphous materials that are quite ductile.
 
Originally posted by HeThatKnows
You can use whatever units you want, but they need to be the right kind and in the right places. Hence my saying "or similar (power per length*temperature)". The "W/c^3" that you used doesn't correspond to thermal conductivity - it would be sorta like stating your CPU speed in cubic-yards or your age in Btu/furlong. :)
Click to expand...

W/c^3 IS W/mK! Just on a different scale!!! Sheesh. Ok, I missed the m, but you don't have to be a brain surgeon to guess what unit of measurement I was using there...... I mean, really.

BTW, fault picking and general criticism to such extremes is NOT conducive to coming up with new ideas like I was simply trying to do. Remember the original spirit and point of the thread.
 
Originally posted by Nazo
W/c^3 IS W/mK! Just on a different scale!!! Sheesh. Ok, I missed the m, but you don't have to be a brain surgeon to guess what unit of measurement I was using there...... I mean, really.
Click to expand...
Relax, I'm not pickin' on you or anything. Just trying to toss some useful info out, to correct the mistake early in the thread about water's thermal conductivity. Can't design a water block from amorphous metals if we don't understand water. Speaking of which, did anyone find any numbers on the thermal conductivity for the stuff?

By the way, how do you convert W/c^3 to W/mK?
 
Originally posted by HeThatKnows
By the way, how do you convert W/c^3 to W/mK?
Click to expand...

Well, the first time I listed it correctly as "Watts/cm-°C", but then I decided to abbreviate more, and apparently forgot the C. Beyond there, it's just simple math.

1 meter = 100 centimeters, which is commonly abbreviated as cm (I accidentally dropped the m the second time and further when just repeating. I did, however, get it right the first time.)
^3 means to the third power. In other words, cubic centimeters.
0K = -273.15C because Kelvin is absolute zero for better physics calculations, but it's still based on the Celcius measurements, merely scaled. This is especially used when working with measurements involving gasses (look up Boyle's law and you'll find it rather abundant.)

If you truly want to convert, I'm just going to let you do the math because you'll pretty much have to write it all out and everything. Wasted effort if you ask me. All that matters here is which is more than the rest, and regardless of the measurement scale proportions should be the same or measurements are wrong somewhere. I did check and everyone else seems to list water as being lower, so you were right that the person's earlier source was wrong in that respect, but they weren't necessarily wrong for using different measurements, which are, in fact, far more fitting in purposes like these (especially since it was meant to compare to diamond and I have to say it's pretty rare that you find a diamond that is 1 meter or more...)

EDIT: Accidentally said 1K. 1K = -273.15C + 1. Like I said, it's based on absolute 0, so 0 Kelvin = absolute zero which is -273.15C. Oh well, simple mistake, sorry.
 
Originally posted by Nazo
Well, the first time I listed it correctly as "Watts/cm-°C", but then I decided to abbreviate more, and apparently forgot the C. Beyond there, it's just simple math.

1 meter = 100 centimeters, which is commonly abbreviated as cm (I accidentally dropped the m the second time and further when just repeating. I did, however, get it right the first time.)
^3 means to the third power. In other words, cubic centimeters.
Click to expand...

but there is no cubed length unit in a thermal conductivity number, it's W/m*K, or using centimeters, W/cm*K

they weren't necessarily wrong for using different measurements, which are, in fact, far more fitting in purposes like these (especially since it was meant to compare to diamond and I have to say it's pretty rare that you find a diamond that is 1 meter or more...)
Click to expand...


Using cm instead of m is fine as long as you're consistent with units throughout your calculations and results. But the length of the part really doesn't matter, just because of that- using consistent units will always result in the same answer.
 
Originally posted by zer0signal667
but there is no cubed length unit in a thermal conductivity number, it's W/m*K, or using centimeters, W/cm*K
Click to expand...
We aren't talking about a one dimentional slab of metal. Such a thing can't exist in reality. (In all technicality, even with two dimentional it can't REALLY exist since it must be at least one layer of the smallest possible object thick to even exist.) REGARDLESS, a cubic centimeter STILL WORKS THE SAME. Basically, just instead of looking at one dimentional matter, we look at three dimentional matter, which works a little better here in reality. However, even if we simply ignore this fact, it still works the same. Again, just looking at a different scale.

Using cm instead of m is fine as long as you're consistent with units throughout your calculations and results. But the length of the part really doesn't matter, just because of that- using consistent units will always result in the same answer. [/B]
Click to expand...
At least you finally understood the part about units not mattering provided it's consistant. I had to say that about three times... However, the length does matter due to reistances/etc provided you don't have a magical material that has perfect conductivity (here in reality even diamond can't come close to being literally perfect.)

EDIT: To clarify, one cubic centimeter (cm^3) = 1 cm x 1 cm x 1 cm. In other words, if you must reduce it to a one dimentional object, it is 3 cm for 1 cc. I guess they should have used cL to prevent this confusion.
 
Originally posted by Nazo
We aren't talking about a one dimentional slab of metal. Such a thing can't exist in reality. (In all technicality, even with two dimentional it can't REALLY exist since it must be at least one layer of the smallest possible object thick to even exist.) REGARDLESS, a cubic centimeter STILL WORKS THE SAME. Basically, just instead of looking at one dimentional matter, we look at three dimentional matter, which works a little better here in reality. However, even if we simply ignore this fact, it still works the same. Again, just looking at a different scale.
...
Click to expand...

You're making Baby Fourier cry.

The units for thermal conductivity are usually W/(m*K). When you start getting into multidimensional heat transfer, the method of calculating heat transfer becomes a bit more complex. But the units of thermal conductivity are still quite firmly based off of power/(distance*temperature).

Reference for further information.
 
Originally posted by Nazo
We aren't talking about a one dimentional slab of metal. Such a thing can't exist in reality. (In all technicality, even with two dimentional it can't REALLY exist since it must be at least one layer of the smallest possible object thick to even exist.) REGARDLESS, a cubic centimeter STILL WORKS THE SAME. Basically, just instead of looking at one dimentional matter, we look at three dimentional matter, which works a little better here in reality. However, even if we simply ignore this fact, it still works the same. Again, just looking at a different scale.
Click to expand...

OK, well 3 dimensional heat transfer modeling still does not require the use of your magical unit of measurement. There are two pertinent quantities- thermal conductivity and thermal diffusivity. Thermal conductivity has units of power/length*temperature change. Thermal diffusivity has units of length/time^2 (m/s^2 in SI units). You can use both quantities to model a transient response, if you want to get really technical.


At least you finally understood the part about units not mattering provided it's consistant. I had to say that about three times... However, the length does matter due to reistances/etc provided you don't have a magical material that has perfect conductivity (here in reality even diamond can't come close to being literally perfect.)

EDIT: To clarify, one cubic centimeter (cm^3) = 1 cm x 1 cm x 1 cm. In other words, if you must reduce it to a one dimentional object, it is 3 cm for 1 cc. I guess they should have used cL to prevent this confusion.
Click to expand...

I always understood that units are relative, I'm an engineer, and I deal with units constantly. However a meter is a unit of length and a cubic centimeter is a unit of volume, they're not comparable.
1m /= 1cm^3.
And length does not matter, are you saying that you can't use certain units of measurement for objects of certain sizes? You can express your height in units of centimeters, meters, or kilometers and it always means the same thing. Try using a meter stick to measure a pen, and realize that you can relate the 10cm it measures to meters- it's .1 meters. Still the same length. Multiply by four to get the length of four pens. 40cm is still the same as .4m.
 
Originally posted by zer0signal667
OK, well 3 dimensional heat transfer modeling still does not require the use of your magical unit of measurement.
Click to expand...
Erm, precicely what part of the standardly world-wide accepted SI units is "magical"? Please, I'd be terribly interested to know the answer to that...

1m /= 1cm^3.
Click to expand...
Firstly, I'm not the one who "made up" this unit that you insist on assuming is made up. Secondly, volume is another word to refer to the three dimentional version of, well, distance. And no, 1m isn't 1cm^3, but I _NEVER_ said it was. I said it's the same thing on a different scale. Even if that weren't just completly off the whole point, it would be more accurate to say 1m != 1m^3, which is true because 1m^3 = 1m x 1m x 1m. I guess engineers are too used to flat paper.

And length does not matter, are you saying that you can't use certain units of measurement for objects of certain sizes?[/B]
Click to expand...
Length doesn't matter? Then why is it used in every one of them including the one you used? The amount of heat that can be transfered through an amount of material depends on how much of that material you HAVE! You assume you transfer the exact same amount whether you have a tiny pea sized heatsink as you would if you have one that pulls your PC over. So why don't they make that since it transfers the same amount? Ok, admittedly you have to make minor adjustments for the fact that it has to be actively cooled, so we are talking more like a quarter thick but wide heatsink so it's big enough for the fan to cover. There's no reason that physically can't be made. The problem is, it wouldn't transfer nearly enough heat away from the CPU. And if you are thinking right now to argue that more metal can CONTAIN more heat, but that doesn't work because eventually the "storage room" runs out as the amount of heat would continously build up over time until eventually it would simply be too high to handle (more storage only buys more time until it runs out, not decreases speed of lowering heat, the speed would remain the same.)
 
This isn't meant to be a flame, Nazo, so keep that in mind before firing something back.

Both Zer0signal667 and I are engineers, we know our units (unlike some poor NASA chaps), and we've both had coursework in heat transfer theory. Please, please read this before you say anything else.
 
What did you intend for me to see there exactly? It still says that you use a measurement of the distance of the "path" involved.

As for flaming, you might be surprised, but even when people seem to assume that I'm just seething mad, I rarely am.
 
Originally posted by Nazo
Erm, precicely what part of the standardly world-wide accepted SI units is "magical"? Please, I'd be terribly interested to know the answer to that...


Firstly, I'm not the one who "made up" this unit that you insist on assuming is made up. Secondly, volume is another word to refer to the three dimentional version of, well, distance. And no, 1m isn't 1cm^3, but I _NEVER_ said it was. I said it's the same thing on a different scale. Even if that weren't just completly off the whole point, it would be more accurate to say 1m != 1m^3, which is true because 1m^3 = 1m x 1m x 1m. I guess engineers are too used to flat paper.
Click to expand...

Just because it contains units of length and power does not mean that it MEANS anything or that it's USEFUL. Sure, you can say that there is some property with units of W/cm^3, but nobody cares because heat flow is characterized by units of W/m*K. You DID make up that unit, I have never seen anything like it before and I don't know where you would have.

1m/=1cm^3 <-- different units, bad
1m/=1m^3 <-- different units, bad
1m^3=1x10^6cm^3 <--same units, good!

I don't know what your definition of a "scale" is, but scaling units never requires a change of dimensions. I'm sorry, but I'm not going to argue about this anymore, I suggest you take Aggie's advice and do some reading.
 
Look, I was just posting something I had found and a simple idea I had out of curiosity, ok? I'm sorry I ever even TRIED to think up new ideas, god. I forgot that to do so you have to be 100% perfect and never mistype anything or whatever.

That said, I think it's a good time to end this. Remember the point of the whole thread.

No one has yet confirmed whether or not LiquidMetal (or similar products) would actually be bad or good at thermal conductivity unless I passed over it (quite possible mind.) However, we have seen arguments that it's quite doubtful. Anyone know more specifically? So far it looks bad though as those are very good arguments.

I guess the silver heatsink is still the best, ne? Anyone have the money to try that? Er, actually, if you do, instead of wasting on a novelty item like that, buy me a decent system... ^_^
 
You must log in or register to reply here.
Back
Top