Project Gallium (Liquid Metal Cooled PC)

[Tmt97]

Just now, Dash Lambda said:

What I mean by "too-perfect-to-be-real" is the essentially featureless reflective surface liquid metals have. You can get normal metal to look like that, sure, but it's something you have to maintain and it wouldn't apply to the channels inside the water blocks.

 

Of course, I'm assuming they're going to use flexible tubing. This is one of those situations where I think hardline would actually look worse than flexible, because then it actually would be hard to tell that it's liquid metal and not just metal piping...

Oh, I see what you mean. Just because it would be shiny without any oxidation, or fingerprints, ect... Yea, I get ya, although you get that same effect if you look at a heatsink in a mineral oil PC. I srsly doubt they will use flexible tubing though. I mean if they did it would certainly help their case as they try to convince people that its gallium and not just metal pipes, but you gotta remember how heavy gallium is. I question wether or not flexible metal tubing would hold up to the forces of gravity in this case.

[Dash Lambda]

1 minute ago, Tmt97 said:

Oh, I see what you mean. Just because it would be shiny without any oxidation, or fingerprints, ect... Yea, I get ya, although you get that same effect if you look at a heatsink in a mineral oil PC. I srsly doubt they will use flexible tubing though. I mean if they did it would certainly help their case as they try to convince people that its gallium and not just metal pipes, but you gotta remember how heavy gallium is. I question wether or not flexible metal tubing would hold up to the forces of gravity in this case.

Well, gallium isn't nearly as heavy as something like Mercury... It's definitely heavier than water, but I don't think it's heavy enough to be a problem.

That is, assuming they keep the tubing runs short. I can see it being an issue with a stretch of tubing that runs the length of the case or something.

[Tmt97]

Just now, Dash Lambda said:

Well, gallium isn't nearly as heavy as something like Mercury... It's definitely heavier than water, but I don't think it's heavy enough to be a problem.

That is, assuming they keep the tubing runs short. I can see it being an issue with a stretch of tubing that runs the length of the case or something.

Well its not the tubing itself that would worry me as much as it would the fittings. Liquid cooling loops just use compression fittings. You can see videos where even after you set one up properly you can still rip the tube right out of its fitting with a firm tug. These kinds of fittings just aren't meant for a lot of force to be put on them. I wouldn't trust them to hold up the full weight of a solid metal rod, which is essentially what a tube filled with gallium will weigh. Maybe if they used a different type of fitting they'd have a chance, as I don't imagine its heavy enough to rip a tube apart...

[iamdarkyoshi]

1 minute ago, Tmt97 said:

Well its not the tubing itself that would worry me as much as it would the fittings. Liquid cooling loops just use compression fittings. You can see videos where even after you set one up properly you can still rip the tube right out of its fitting with a firm tug. These kinds of fittings just aren't meant for a lot of force to be put on them. I wouldn't trust them to hold up the full weight of a solid metal rod, which is essentially what a tube filled with gallium will weigh. Maybe if they used a different type of fitting they'd have a chance, as I don't imagine its heavy enough to rip a tube apart...

My barbed fittings are strong enough to where the tube would break before the fitting seperates from the tube

[Tmt97]

Just now, iamdarkyoshi said:

My barbed fittings are strong enough to where the tube would break before the fitting seperates from the tube

Maybe they should use those then. I'd just hate to see it leak, as I'd imagine liquid metal is pretty conductive.

[Dash Lambda]

1 minute ago, Tmt97 said:

Well its not the tubing itself that would worry me as much as it would the fittings. Liquid cooling loops just use compression fittings. You can see videos where even after you set one up properly you can still rip the tube right out of its fitting with a firm tug. These kinds of fittings just aren't meant for a lot of force to be put on them. I wouldn't trust them to hold up the full weight of a solid metal rod, which is essentially what a tube filled with gallium will weigh. Maybe if they used a different type of fitting they'd have a chance, as I don't imagine its heavy enough to rip a tube apart...

Thus why a long stretch of tubing would be worrisome. Something else that could come into play is the momentum of the dense liquid metal as it travels through the tubing... Depending on what kind of pump they use and the bends of the tubing, it could possibly carry enough force to knock something loose.

 

Though, with weight as the only concern, they could easily just have it supported. Anything from hanging wire to a sort of scaffolding should work...

[Tmt97]

1 minute ago, Dash Lambda said:

Thus why a long stretch of tubing would be worrisome. Something else that could come into play is the momentum of the dense liquid metal as it travels through the tubing... Depending on what kind of pump they use and the bends of the tubing, it could possibly carry enough force to knock something loose.

 

Though, with weight as the only concern, they could easily just have it supported. Anything from hanging wire to a sort of scaffolding should work...

Yea, I suggested scaffolding if they used mercury, and they would definitely need it for soft tubing. Of course the scaffolding would be easier if it was hardline. Maybe a combination of the two where the parts reinforced with scaffolding are hardline and the rest is short runs of soft tubing would work. Your right about the momentum though. That could also break stuff internally potentially. Especially thinner stuff like those micro-channels you pointed out.

[Dash Lambda]

4 minutes ago, Tmt97 said:

Yea, I suggested scaffolding if they used mercury, and they would definitely need it for soft tubing. Of course the scaffolding would be easier if it was hardline. Maybe a combination of the two where the parts reinforced with scaffolding are hardline and the rest is short runs of soft tubing would work. Your right about the momentum though. That could also break stuff internally potentially. Especially thinner stuff like those micro-channels you pointed out.

I could only really see the momentum being a problem for free objects. Since it's not actually impacting the channels, it shouldn't cause that sort of damage...

But then again, maybe the pressure would be greater or something.

[Tmt97]

5 minutes ago, Dash Lambda said:

I could only really see the momentum being a problem for free objects. Since it's not actually impacting the channels, it shouldn't cause that sort of damage...

But then again, maybe the pressure would be greater or something.

I dunno... There could be free objects though in some of the components. I know nothing about the engineering of a pump for example, no clue what pumping solid metal could dislodge in one of those.

[Dash Lambda]

Just now, Tmt97 said:

I dunno... There could be free objects though in some of the components. I know nothing about the engineering of a pump for example, no clue what pumping solid metal could dislodge in one of those.

Well, that depends on the pump they use... There are a few different kinds.

 

@ElementalComputers, you mentioned using an electromagnetic pump- You're talking about one that directly uses induction to force motion in a conductive fluid, correct? I ask because I'm aware of at least three different kinds of pumps that I'd apply the term "electromagnetic pump" to that would work in this situation, and I have no idea which one is actually called that.

[pagani123]

21 minutes ago, Dash Lambda said:

Well, that depends on the pump they use... There are a few different kinds.

 

@ElementalComputers, you mentioned using an electromagnetic pump- You're talking about one that directly uses induction to force motion in a conductive fluid, correct? I ask because I'm aware of at least three different kinds of pumps that I'd apply the term "electromagnetic pump" to that would work in this situation, and I have no idea which one is actually called that.

i can answer this yes we are talking about one that uses induction to force motion. thats the only electromagnetic pumps i know of what are the other ones your thinking of? 

[Dash Lambda]

Just now, Maybach123 said:

i can answer this yes we are talking about one that uses induction to force motion. thats the only electromagnetic pumps i know of what are the other ones your thinking of? 

The other two I was thinking of were the type that uses a sort of magnetic rotor inside a loop of piping and the type that uses a pulsing electromagnet to oscillate a chamber with a series of spring valves.

 

Now that I think about it, that probably shouldn't have even been a question I had to ask.

[Dash Lambda]

35 minutes ago, Maybach123 said:

-snip-

And I just realized a question I should ask: Is this going to be a hardline or soft tube loop?

-Or, of course, is it still undecided?

[pagani123]

9 minutes ago, Dash Lambda said:

And I just realized a question I should ask: Is this going to be a hardline or soft tube loop?

-Or, of course, is it still undecided?

still undecided but most likely hardline. aesthetics are really the last thing i am thinking about regarding this project atm to be perfectly honest

[Dash Lambda]

9 minutes ago, Maybach123 said:

still undecided but most likely hardline. aesthetics are really the last thing i am thinking about regarding this project atm to be perfectly honest

Understandable, it's going to be an impressive feat to get this thing working. I can't wait to see what comes of it~

[pagani123]

.

[vanished]

(going back to before that whole discussion) I've seen a lot of talk about surface tension but nothing about contact angle.  That is also going to play a role in how the fluid "sticks" to things.  Ever wondered why water spilled on a table will stay together as a pool and spread out, while mercury (for example) will break up into little beads and seem to float as a ball above the surface? Contact angle.

[ElementalComputers]

so just an update we have had a discussion and looks like for now were going with a small cpu only loop to keep gallium cost down. Eventually we will add a gpu to the loop but for now just a cpu loop. Also on a side note we are building a website so look forward to that soon.

thanks  

[Stefan1024]

Any news on this very interesting build log?

[pagani123]

39 minutes ago, Stefan1024 said:

Any news on this very interesting build log?

I can answer this. Right now i am still in the process of doing somewhat long term tests with copper submerged in gallium  so far after a month i have only seen the slightest bit of degradation from the copper. We are also trying to work out a deal with the parylene deposition company and trying to find a supplier for the amount of gallium and indium we need. A lot of behind the scenes stuff. 

[thekeemo]

@Trik'Stari its not mercury but still metal

[ryanmmax]

Saw some of the questions about pumps, and I thought I would suggest using a Peristaltic Pump.  If your not familiar with them they are typically used for chemicals and have a few very beneficial properties for this use case. 

 

-Aesthetically could look freaking cool, with clear tubing you could see the metal being pumped!

-As "positive displacement" pumps so they handle very viscus fluid and high DP.

-Due to the design the "fluid" would not contact anything other then the tubing so reactions and other issues would be minimal to non-existent.  

[pagani123]

31 minutes ago, ryanmmax said:

Saw some of the questions about pumps, and I thought I would suggest using a Peristaltic Pump.  If your not familiar with them they are typically used for chemicals and have a few very beneficial properties for this use case. 

 

-Aesthetically could look freaking cool, with clear tubing you could see the metal being pumped!

-As "positive displacement" pumps so they handle very viscus fluid and high DP.

-Due to the design the "fluid" would not contact anything other then the tubing so reactions and other issues would be minimal to non-existent.  

interesting i might have that on the prototype cooler but i would like to go full electromagnetic eventually.

[ElementalComputers]

Update: The long term copper and gallium test is finished and so far we only see minor structural degradation to the copper which means we will be able to use copper pipe for the radiator but for now we are putting this build on hold until term ends. After that we will have more time to work on it but right now we are mostly focused on studying for finals. sorry for the delay.

 

[SImoHayha]

Doesn't Gallium destroy aluminum ?