mercury cooling system

alg7_munif

Supreme [H]ardness
Joined
Oct 9, 2006
Messages
5,862
I somehow managed to miss this thread... upon reading it I'm really suprised that people were discussing about magnetic motor damaging the harddrives but nobody mentioned anything about SSD?
 
Joined
Oct 1, 2008
Messages
687
SSDs are not magnetic, a magnetic motor wouldn't have the same effect on it as a hard drive (if it was powerful enough to damage a hard drive in the first place)
 

flogge

[H]ard|Gawd
Joined
May 20, 2008
Messages
1,888
anyone have any ideas about letting thermal convection to move the coolant in the loop move the mercury instead of using a pump....?

I think water has a higher thermal expansion rate than mercury and so would be a better all-around candidate for a pumpless design. http://physics.info/expansion/

I have thought about building a pumpless water-cooled rig in the past for absolute silence using thermal circulation. Perhaps ethyl alcohol straight or mixed with water?
 

AVT

Supreme [H]ardness
Joined
Aug 8, 2008
Messages
4,824
as some of us are aware diamond is one of the most heat conductive materials known to man...... and i am flattered my nerdyness made your day...... :D

My question is how you'd manufacture a block using diamond as a material. :eek:

OK apparently no one here knows how magnetic fields and Faraday cages work. Stop it.

The answer is a strong focused magnetic field will screw up a HD in a second.

The answer is motor magnetic fields are focused and the field strength falls off rapidly even if any of it escapes the motor housing the HD case will shield sufficiently from any leakage. Even an inch of space and or any other grounded ferrous metal will knock the field strength down significantly.

Take a 5 lb horseshoe magnet and set it directly on top of an HD and its turned into a raw drive instantly and you will be lucky if it low level formats properly again.


Dont make me have to stop this car. Carry on.
About the Mercury. You will shoot your eye out.

So use an SSD instead. :D

edit: damn, beaten to it on pg 5. :(
 

thaltek

Limp Gawd
Joined
Sep 15, 2009
Messages
206
My question is how you'd manufacture a block using diamond as a material. :eek:

to answer your question: it depends.... as there is more than one way to do this....
1. most direct solution get your hands on a solid high purity wafer at least 1" in diameter. this would be very expensive so probably not within reason for most people....
2. use smaller diamond inserts and cast high purity copper, silver, or gold to encase the diamond. this would be less expensive especially since you wouldn't need a large stone...
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
to answer your question: it depends.... as there is more than one way to do this....
1. most direct solution get your hands on a solid high purity wafer at least 1" in diameter. this would be very expensive so probably not within reason for most people....
2. use smaller diamond inserts and cast high purity copper, silver, or gold to encase the diamond. this would be less expensive especially since you wouldn't need a large stone...

As much as people talk about using diamonds for heatsinks, I'm pretty sure a continious liquid He stream poured directly over the outside of the CPU with no heatsink would be much more economical and more effective. Although I have no idea if liquid He is conductive, but I would guess that it isn't and I don't feel like googling it. :p
 

thaltek

Limp Gawd
Joined
Sep 15, 2009
Messages
206
As much as people talk about using diamonds for heatsinks, I'm pretty sure a continious liquid He stream poured directly over the outside of the CPU with no heatsink would be much more economical and more effective. Although I have no idea if liquid He is conductive, but I would guess that it isn't and I don't feel like googling it. :p
lol.... you do realize that liquid helium flows up, right?
 

thaltek

Limp Gawd
Joined
Sep 15, 2009
Messages
206
Umm, liquid helium is 100 times denser than air.... gas helium is less dense than air, but liquid helium is not.

when in transition from a liquid to a gas it has the rising fluid behavior...... which would start happening by the time you get the bloody stuff more than a few inches from its, well insulated storage container....... not to mention more so when brought near something that is 30C......................
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
when in transition from a liquid to a gas it has the rising fluid behavior...... which would start happening by the time you get the bloody stuff more than a few inches from its, well insulated storage container....... not to mention more so when brought near something that is 30C......................

/sigh
 

thaltek

Limp Gawd
Joined
Sep 15, 2009
Messages
206
Well, I could tell you what I think of your opinions but it wouldn't get us anywhere so I figured I'd just sigh to let you know that you aren't getting a reply.

sorry, i can be an ass some days.....

i guess what i am trying to say is that i don't feel that liquid He is the solution besides this thread is to discuss mercury cooling... feel free to start a liquid He thread if you so wish.... i would be interested to see just what comes of it.....
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
sorry, i can be an ass some days.....

i guess what i am trying to say is that i don't feel that liquid He is the solution besides this thread is to discuss mercury cooling... feel free to start a liquid He thread if you so wish.... i would be interested to see just what comes of it.....

Liquid He is all ready used. Someone has posted that you should make a diamond heatsink. I responded by saying it would be cheaper and cooler to simply pour a steady stream of liquid He. He pours just like water btw.
 

thaltek

Limp Gawd
Joined
Sep 15, 2009
Messages
206
Liquid He is all ready used. Someone has posted that you should make a diamond heatsink. I responded by saying it would be cheaper and cooler to simply pour a steady stream of liquid He. He pours just like water btw.

:)

anyways back to the subject.... i am starting to wonder if the mercury would even need to flow considering how conductive it is......
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
:)

anyways back to the subject.... i am starting to wonder if the mercury would even need to flow considering how conductive it is......

It's really not that conductive at all. You are talking like... 1/50th of copper or 1/5000th of a good heat pipe.
 

Temo Vryce

Weaksauce
Joined
Mar 6, 2007
Messages
93
You all are dancing around the biggest flaw with using mercury to cool your system. It's used in thermometers because of it's thermal expansion properties. Now take that and add in a closed loop with no air and no room for expansion and what you have?


A TOXIC BOMB in your house.
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
You all are dancing around the biggest flaw with using mercury to cool your system. It's used in thermometers because of it's thermal expansion properties. Now take that and add in a closed loop with no air and no room for expansion and what you have?


A TOXIC BOMB in your house.

1) It doesn't expand anywhere near what you think it does.
2) There is air in the reservoir.
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
whoops my bad...... i meant heat capacity and yes it would need to flow.....

Ok? I think the point still stands as to why? Even if you developed "hyper-super-ultra-uber-liquid" you would still be limited by:
1) Your CPU-block and it's ability to get the heat off the CPU and into the liquid
2) Even if you created an "insane-uber-super-awesome-cpu-block-of-doom" you are still limited by ambient temperature. (Assuming you created an "insane-uber-super-awesome-radiator-of-doom" to go with the cpu-block)

Seeing how the alternative, a mundane phase-change cooler, can easily go sub-ambient, Mercury just doesn't make any sense.
 

Matthew Kane

Supreme [H]ardness
Joined
Dec 1, 2007
Messages
4,233
This thread is still going on.

Well I'm not Science nerd or anything, but I asked my Science/Chem/Physics/Engineering teacher, Mercury Cooling is possible but it will cost tons to get the appropriate parts to push heavy mercury compared to water around a liquid cooling system due to its heavier density. And do you know Mercury eats metal for breakfast? Why do think they are stored in glass?
 

Halsey

Limp Gawd
Joined
Dec 29, 2008
Messages
131
This is the most ridiculously awesome thread ever! :)

I like the idea of letting thermal convection move the mercury, it would make the system (relatively) lower maintenance than using a pump (no moving parts to replace). The question is, would the design just be similar then to a heatpipe, if not exactly that? Or is there some other method besides a heatpipe design, that might possibly be more efficient?
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
This is the most ridiculously awesome thread ever! :)

I like the idea of letting thermal convection move the mercury, it would make the system (relatively) lower maintenance than using a pump (no moving parts to replace). The question is, would the design just be similar then to a heatpipe, if not exactly that? Or is there some other method besides a heatpipe design, that might possibly be more efficient?

I'll try to be nice forgive me if I'm not.

You would never be able to design a system in which natural convection moves the mercury with a lower temp delta than an air-cooled system, much less a watercooled system.

Heat pipes don't do that at all. Heat pipes phase change a liquid to a gas which then drives the movement to the other end of the pipe where it condenses back to a liquid. It is the phase change that captures and moves the heat.
 

Halsey

Limp Gawd
Joined
Dec 29, 2008
Messages
131
I'll try to be nice forgive me if I'm not.

You would never be able to design a system in which natural convection moves the mercury with a lower temp delta than an air-cooled system, much less a watercooled system.

Heat pipes don't do that at all. Heat pipes phase change a liquid to a gas which then drives the movement to the other end of the pipe where it condenses back to a liquid. It is the phase change that captures and moves the heat.

Ah, ok. Yeah, actually that does make sense when I stop and think about it. :p Guess I didn't quite grasp the concept of how a heatpipe works. My bad.

On to the next best thing: what was the final verdict on an EM motor? I followed those wikipedia links about Fleming's Right Hand Rule, and it seems like a solid option for a pump, although I'm far from an expert (obviously) in this subject.
 
Last edited:

thaltek

Limp Gawd
Joined
Sep 15, 2009
Messages
206
This thread is still going on.

Well I'm not Science nerd or anything, but I asked my Science/Chem/Physics/Engineering teacher, Mercury Cooling is possible but it will cost tons to get the appropriate parts to push heavy mercury compared to water around a liquid cooling system due to its heavier density. And do you know Mercury eats metal for breakfast? Why do think they are stored in glass?

*facepalm* i believe everything in your statement was covered in the first two pages of this thread..... and again mercury has different levels of reactivity with different metals.... for instance mercury is nonreactive with iron which is why iron containers are commonly used for transportation. aluminum on the other hand will form an amalgam which will rapidly erode the aluminum....
 

Matthew Kane

Supreme [H]ardness
Joined
Dec 1, 2007
Messages
4,233
*facepalm* i believe everything in your statement was covered in the first two pages of this thread..... and again mercury has different levels of reactivity with different metals.... for instance mercury is nonreactive with iron which is why iron containers are commonly used for transportation. aluminum on the other hand will form an amalgam which will rapidly erode the aluminum....

hence thats why I'm posting, to make more of a point how useless this ongoing thread is, but ridiculous amusing at the same time to read.
 

thaltek

Limp Gawd
Joined
Sep 15, 2009
Messages
206
hence thats why I'm posting, to make more of a point how useless this ongoing thread is, but ridiculous amusing at the same time to read.

or is it? :D

this is the last time i will explain myself for starting this thread.......
i wanted to test the waters and see how people would react to this type of subject matter, also i wanted to get the creative juices of this community thinking of other ways (creatively) to approach cooling problems... ya know like using something like mercury as the transfer medium.... and thinking of ways to make it work.......
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
or is it? :D

this is the last time i will explain myself for starting this thread.......
i wanted to test the waters and see how people would react to this type of subject matter, also i wanted to get the creative juices of this community thinking of other ways (creatively) to approach cooling problems... ya know like using something like mercury as the transfer medium.... and thinking of ways to make it work.......

My opinion is the VIABLE cooling options are as follows:

Ambient
Stock air cooling
Improved air cooling (~20$ range)
Ultra air cooling (Thermaltake Ultra 120 type)
----- limit of air cooling ----
Entry level water cooling (Corsair's HT-50 kit for example)
High end water cooling (triple radiator, high end block, etc)
---- limit of water cooling ---
---- limit of ambient cooling ---

Sub Ambient

Non-viable cooling options.
Peltair blocks - Only viable for lower power consumption systems. Going about 20C sub ambient on an i7 will run you 1000Watts of power consumption AND disipating 1000Watts+CPU on the other side. Phase change makes much more sense.

DICE - Non-continous
LN2 - Non-continous - Good for record runs
LHe - Non-continous - Good for record runs

Viable
Phase change, - compressor loops, i.e. hacked air-conditioners, "chillers", etc. Eventually I think we'll see someone splice a water block in the loop and make a commercial product capable of wattages needed for OC'ed i7s.
 

BrainEater

[H]ard|Gawd
Joined
Jul 21, 2004
Messages
1,087
You've missed a few.

I can assure you , submersion cooling works. :p

Related to this , spray cooling is also employed. (cray)

Also , special mention should be given to 'passive' cooling under the 'air cooling' section.
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
You've missed a few.

I can assure you , submersion cooling works. :p

Related to this , spray cooling is also employed. (cray)

Also , special mention should be given to 'passive' cooling under the 'air cooling' section.

Yeah, the list is not complete.

As far as submersion cooling goes, I see no real benefit vs standard water cooling, assuming of course that "coolness", "awesomeness", and "kickassness" are not among the metrics we are using.

I'm familiar with the technique you are talking about. I'm guessing some kind of methonal evaporative cooling technique? If so it seems it would fall under phase change. Even so, I'd think a simple carnot cycle would make more sense. But again, I'm not familure with the technique you've named.

Passive cooling is great, but I don't think it's going to help us overclock better. XD

That being said, I'm sure you could design (a very expensive) passively cooled system to take a 4Ghz i7.
 

thaltek

Limp Gawd
Joined
Sep 15, 2009
Messages
206
My opinion is the VIABLE cooling options are as follows:

Ambient

passive
Stock air cooling
Improved air cooling (~20$ range)
Ultra air cooling (Thermaltake Ultra 120 type)
----- limit of air cooling ----
Entry level water cooling (Corsair's HT-50 kit for example)
High end water cooling (triple radiator, high end block, etc)
---- limit of water cooling ---
---- limit of ambient cooling ---

Sub Ambient

Non-viable cooling options.
Peltair blocks - Only viable for lower power consumption systems. Going about 20C sub ambient on an i7 will run you 1000Watts of power consumption AND disipating 1000Watts+CPU on the other side. Phase change makes much more sense.

DICE - Non-continuous
LN2 - Non-continuous - Good for record runs
LHe - Non-continuous - Good for record runs

Viable
Phase change, - compressor loops, i.e. hacked air-conditioners, "chillers", etc. Eventually I think we'll see someone splice a water block in the loop and make a commercial product capable of wattages needed for OC'ed i7s.

a good list of cooling options... however this thread was not intended for the discussion of what people have already done and or methods currently being used so much as an academic discussion of possibilities....

@ vengence: this thread is not about viability, instead it is assumed that the person building this has infinite resources..... and they want mercury as the coolant.

@matthew kane: yes i realize this thread is something of a white elephant... just play along and assume its not ridiculous or useless..... besides even your science teacher will tell you that humans don't know everything just yet....

@ temo vryce: i think you are the second person to openly say the system would blow up... to that i lol'd.... in the future i would suggest making an effort to double check such a statement......
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
a good list of cooling options... however this thread was not intended for the discussion of what people have already done and or methods currently being used so much as an academic discussion of possibilities....

@ vengence: this thread is not about viability, instead it is assumed that the person building this has infinite resources..... and they want mercury as the coolant.

You could design a system to make a mercury cooling system. It would be incredibly expensive to design, construct, and operate. It would be out classed by even the simplest phase change coolers. It fails even if you use coolness as a metric as you could have a passive diamond heatsink instead.
 

BrainEater

[H]ard|Gawd
Joined
Jul 21, 2004
Messages
1,087
I'm familiar with the technique you are talking about. I'm guessing some kind of methonal evaporative cooling technique? If so it seems it would fall under phase change. Even so, I'd think a simple carnot cycle would make more sense. But again, I'm not familure with the technique you've named.

It does infact fall under 'phase change' , albiet the most extreme form.

As far as I know it's Fluorinert they use.It's used in the Cray X-1 system.I'll try to dig up the video I saw.

--------

Yea mercury cooling can be done . Drop the cash and get to work. Please be careful.

:D
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
It does infact fall under 'phase change' , albiet the most extreme form.

As far as I know it's Fluorinert they use.It's used in the Cray X-1 system.I'll try to dig up the video I saw.
I did some hunting and found it. It's pretty neat stuff, although I suprised they found it viable.
 

Matthew Kane

Supreme [H]ardness
Joined
Dec 1, 2007
Messages
4,233
@thaltek just out of curiosity, if you had the resources and money would you build such a system?
 
Top