Unique potential WC project - opinions please!

N

nonlnear

2[H]4U
Joined
Sep 22, 2004
Messages
3,273
I'm about to take the watercooling plunge with my ancient Athlon XP mobile. I'm doing it for fun. Not for performance value. Just putting that out there from the start so people won't tell me that upgrading will get me more bang/$.

Anyways, I'm going to be doing a fairly standard setup, but have been thinking about a rather unique project: using a vortex tube to separate the flow. Here's a schematic of what I'm thinking right now:

vortex_tube_wc.png


Advantages:
- extra cooling of CPU water
- extra heating of radiator inflow means greater deltaT. Thus more heat dissipation (everything else being equal - there's the kicker)

Disadvantages:
- possibly severe loss of pressure to waterblock

Here's my take on some of the issues involved:
A vortex tube is not known for its efficiency. They are usually used as spot coolers where a high deltaT is needed, and compressed air is plentiful. They need a lot of energy to achieve any paritcular level of cooling. This is possibly a big issue. However, the way I see it isthis: if (this is a BIG if!) a standard "large" watercooling pump can achieve sufficient pressure on the cold end of the tube, then there's no reason not to use the vortex tube.

The normal thermodynamic objections don't really apply here because this is a closed system. The energy input driving the vortex tube is the pressure head from the pump. If a large conventional watercooling rig can already dissipate this ""extra" energy from a given radiator size, it won't be driving it much harder to put a vortex tube in the system and turn the pump up a bit. After all, AFAIK the heat input of the pump is fairly small compared to the CPUs rate of dissipation. So even if you've got to double the pump's work, it's not going to put a much bigger cooling load on the system.

In most vortex tube applications, the heat output is looked on as waste. However, in a closed WC rig, the hot side is almost as valuable as the cold side. (In fact, possibly more valuable.) The reason is that the amount of heat dissipated by the radiator is largely determined by the temperature difference between the radiator and the ambient air. Thus for a given flow rate, a radiator will dissipate more heat with hotter inflow than with cooler inflow.


That's about it for my musings about this setup. There are some pretty big problems to be addressed in design and implementation. The biggie is just finding out about vortex tube design and performance. There are a handful of decent engineering papers I've found that pertain to some aspects of vortex tube design. Unfortunately, they are all specific to air tubes. I would need calculations that are for water or - better yet - are in terms of dimensionless constants only.

Then again, I could jsut make something and see how it works. That's probably what it's going to come down to, as I don't believe that there is a sufficient model of how a vortex tube really works. The papers I've found pertaining to tube design parameters have all been empirical tests. These are of limited use to me, as I'm using a different fluid. With a decent theoretical model, I could drop it into matlab and see what works best. Not really an option though.

The other big problem is fabrication. But that's a problem that is potentiall a fun one to solve. I've got a couple ideas of how to make the tangential nozzle(s) in a fairly straightforward way. I think I can take care of most of it with a trip to Lowes.

I'll add some of my more detailed design/implementation ideas as opinions roll in. I hope this is a fun project for batting ideas around.
 
Sorry for the ghetto artwork.

The flow direction arrows are black on transparent, so it doesn't work too well on the [H]'s color scheme. I'll fix it in a bit. On the other hand, if the flow direction isn't obvious to you, I probably am not interested in your opinions... :p

*edit: transparency problem "fixed"
 
This could work, but as you said getting the pressure on the cold side of the VT would be a problem. I've never seen a vortex tube use liquid before, as the air ones are generally the most common. They're LOUD too, have you ever seen one work?
 
Arcygenical said:
This could work, but as you said getting the pressure on the cold side of the VT would be a problem. I've never seen a vortex tube use liquid before, as the air ones are generally the most common. They're LOUD too, have you ever seen one work?
Click to expand...

Remember that the vortex tubes you've heard are all air. Water has very different turbulent flow characteristics. I suspect that a water vortex tube will be a lot quieter than an air one - as long as the system is completely bled... although a few bubbles in the works could make it whine, or possibly sound like a sink drain.

The pressure on the cold side is definitely going to be the big problem. The potential saving grace in that regard could be that I really don't need a dramatic temperature drop on the cold side. What I'm going for is maximum total entropy drop.

When going for the coldest cold-side temperature, the throttle valve on the hot side is left almost fully open, leaving very little pressure on the cold side, but a very cold outlet temp. This can give a large drop, but isn't very efficient. By throttling the hot side down, you get a lot higher flow on the cold side, but less temperature drop. Also, you get higher hot side outlet temps. These are all good things. In "typical" vortex tube applications - like machine tool cooling - these are not desirable, so they opt for the less efficient "coldest cold" mode. However, in a watercooling setup, there is no reason not to run the system at the most efficient vortex tube throttle setting.

It might not be too hard to get decent cold pressure after all... The blurb about vortex tube refrigeration here suggests that maximum efficiency happens at a cold fraction (proportion of flow out the cold side) close to 80%. Of course that doesn't say anything about outlet pressure. Of course that's all dealing with air. Things could be VERY different with water.
 
nonlnear said:
I
Advantages:
- extra cooling of CPU water
- extra heating of radiator inflow means greater deltaT. Thus more heat dissipation (everything else being equal - there's the kicker)
Click to expand...

This being a closed system, the vortex part neither adds nor removes heat so the only real benefit is if you can somehow get more cooling from the radiator because the water entering it is hotter than normal. I assume this is what you mean by the 2nd advantage mentioned above.

It's an interesting idea, but what if it does cause a drop in water pressure? How much will this drop be? Could adding a second, or larger, radiator to the system not cause the same drop in pressure while providing greater cooling?

Exactly how much pressure do you need going into the vortex tube for something like this to work well? Because I know with air, it is very high pressure air going into the tube typically.

Overall, I am just concerned that the closed system nature of this will ensure you don't gain much cooling at all from the setup, inspite of the deltaT thing.
 
arentol said:
This being a closed system, the vortex part neither adds nor removes heat so the only real benefit is if you can somehow get more cooling from the radiator because the water entering it is hotter than normal. I assume this is what you mean by the 2nd advantage mentioned above.
Click to expand...

That's exactly what I meant. Higher deltaT between the radiator fluid and ambient.

It's an interesting idea, but what if it does cause a drop in water pressure? How much will this drop be? Could adding a second, or larger, radiator to the system not cause the same drop in pressure while providing greater cooling?
Click to expand...

We'll never know until somebody tries it, right?

Exactly how much pressure do you need going into the vortex tube for something like this to work well? Because I know with air, it is very high pressure air going into the tube typically.

Overall, I am just concerned that the closed system nature of this will ensure you don't gain much cooling at all from the setup, inspite of the deltaT thing.
Click to expand...

Actually, some back of the envelope calculations are a little promising. The Reynolds number at the nozzle part of the tube is roughly given by

Re =U * D / nu,
where U is the tangential velocity at the outside of the tube near the nozzle, D is the diameter of the tube, and nu is the kinematic viscosity. I've been using the results from this paper (abstract here) as a rough guide. They used air in a D=12mm tube. I'm thinking of using 1" or 2" ID tubing for a water tube. That means to keep the same reynolds number in the tube, my nozzle velocity has to be roughly 1/32 (for 1" tube) or 1/64 (for 2") of the nozzle velocity they got. Of course, if you're scaling up the nozzle diameter with the VT diameter, those numbers become 1/8 and 1/4, respectively - still <1, but not as helpful.

There's a problem with the nozzle diameter, though. In the 6 nozzle setup in the paper referenced above, the nozzle diameter appears to be about 1.5mm. Even if you scale them proportionally to the 2" tube setup, there's no practical way to pump ~1800 GPH through six 6mm ID holes. This obviously requires a more inventive solution... (continued in next post)

* there are a couple other papers I'm drawing from. I'm not just winging it with the one.
 
The Swiftech MCP350 and DangerDen DDC-12V appear strikingly similar...

The blowup view on swiftnets.com makes it look like with a little tinkering, one could buld the pump into the intake end of a vortex tube. Thus, the vortex wouldn't be fed by the intake nozzle velocity, but would actually have an impeller right in the vortex tube itself.

The rotor of the pump appears to be hollow. I don't have a great view of it, so I'm not sure if this is the case. Anybosy who's tinkered with one, please let me know. Basically what I'm thinking is taking the top off of the pump, and standing the tube on top of it, drilling a hole in the bottom of it for the cold outlet (Is this even possible?), and hooking up an intake nozzle in the reverse direction of the factory exit nozzle. (or better yet, can this pump be made to rotate in reverse? If so, this would save a little bit of fabrication.) This system would of course have to be driven by another pump that's a good deal more powerful than the one spinning in the tube. After all, it's basically got to forcefeed the MCP350 (almost) in reverse. Maybe a huge ass eheim or something.

Lots of work if it's going to happen, that's for sure.
 
I have an engineering background but am not familiar with a Vortex tube. I suspect I'm not alone here... Can you give us a brief overview on what a vortext tube does and how it works?
 
nonlnear said:
The Swiftech MCP350 and DangerDen DDC-12V appear strikingly similar...

The blowup view on swiftnets.com makes it look like with a little tinkering, one could buld the pump into the intake end of a vortex tube. Thus, the vortex wouldn't be fed by the intake nozzle velocity, but would actually have an impeller right in the vortex tube itself.

The rotor of the pump appears to be hollow. I don't have a great view of it, so I'm not sure if this is the case. Anybosy who's tinkered with one, please let me know. Basically what I'm thinking is taking the top off of the pump, and standing the tube on top of it, drilling a hole in the bottom of it for the cold outlet (Is this even possible?), and hooking up an intake nozzle in the reverse direction of the factory exit nozzle. (or better yet, can this pump be made to rotate in reverse? If so, this would save a little bit of fabrication.) This system would of course have to be driven by another pump that's a good deal more powerful than the one spinning in the tube. After all, it's basically got to forcefeed the MCP350 (almost) in reverse. Maybe a huge ass eheim or something.

Lots of work if it's going to happen, that's for sure.
Click to expand...

There are companies that have made custom plexi tops for the DDC pump (search for DDC Ultra) which is the same as the MCP350. The Iwaki pumps may do the job as the primary driver... They are very powerful. Good info on pumps here... http://www.xtremesystems.org/forums/showthread.php?t=41495
 
virtualrain said:
I have an engineering background but am not familiar with a Vortex tube. I suspect I'm not alone here... Can you give us a brief overview on what a vortext tube does and how it works?
Click to expand...

There's a decent brief overviews of vortex tubes on wikipedia. Other than that, just start googling "vortex tube" and see what you find.

*edit: Sorry about the terseness - I'm not trying to be dismissive. It's just that I could go on for pages and still not touch on all of the subtleties that go into the design decisions involved. If you have any particular questions, I'd love to answer (if I can), but it's hard to know where to stop an answer to such an open ended question.
 
Do you happen to have any links to any companies that actually sell liquid based vortex tubes, or a link to a website that describes one functioning with liquid in it instead of air?

I am just wondering because I can't find a mention of Vortex tubes that doesn't include the fact that compressed air is the expected input to the tube. Even the wikipedia link you posted is specific about that. It would be nice to have a link to proof this worked with water at all before continuing discussion of the actual functionality of it.
 
nonlnear said:
There's a decent brief overviews of vortex tubes on wikipedia. Other than that, just start googling "vortex tube" and see what you find.

*edit: Sorry about the terseness - I'm not trying to be dismissive. It's just that I could go on for pages and still not touch on all of the subtleties that go into the design decisions involved. If you have any particular questions, I'd love to answer (if I can), but it's hard to know where to stop an answer to such an open ended question.
Click to expand...

The Wikipedia overview was great... Thanks.
 
Finally; an interesting idea!

Though I don't know much at all about vortex tubes, it seems like the method you'd be using will require at least one substancially powerful pump, and these tend to not come cheaply. Iwaki MD series are good performers, but they're also very large. Ideally, you want a Japanese RZ series MD pump or a no-holds-barred RD series (RD-20 or RD-30). These RD pumps are absolutely fantastic: tremendous head, reasonably good flow, fairly compact (comparatively) and not terribly loud. The snag is the pricetag - you're looking at anything between $220 and $270 for the pump itself and roughly $70 for a quality 24V power supply. Another potential snag is that these can only be ordered in quantity, so if you consider trying to get one, you'll have to wait for a group buy or organize one yourself. I don't know of any other mag-drive or centrifugal pumps with similar characteristics that don't use large, 1/4hp+ AC motors, but it's possible a couple other companies make somewhat similarly performing pumps at the same pricepoint.
 
Very Interesting Concept!
Sir, I APPLAUDE you LOUDLY!!!!!

Now on to the technical issues.
I'm a Manufacturing Engineer and I have used Vortex Tubes for years for spot cooling various machining operations. Especially on open machines like Bridgeport Milling Machines when you need cooling but don't want the mess associated with water or oil based coolants.

Based on the theory of how a vortex tube works and some of my half remembered fluids classes - We are interested in Flow Velocity more than pressure. Plus, in designing an air powered system, rule of thumb states that if you open it to external air - then concider all pressures to be zero - which means that we are converting the air pressure's potential energy to velocity
Go to Exair's Vortex Tube Page they are the ones that I generally use. They have a beautiful animation and a link to a theory and info page.
 
stormshadow said:
interesting concept...
but i looked thru wiki and saw this:

Vortex tubes have lower efficiency than traditional air conditioning equipment.

ummm....
Click to expand...

That's when you are considering their efficiency only as a cooler. The output of the hot end is counted as waste heat in that calculation. In this case, the hot output is just as valuable as (possibly more valuable than) the cold output. IT increases the effective heat transfer rate of the radiator.

There just isn't that much info out there that would let one run the numbers on total system cooling efficiency for the setup I'm suggesting. There's the temperature and pressure drops on the cold end, coupled with the effective cooling of the radiator due to the temperature increase from the hot end. To do the analysis properly, you need the temperature, pressure, and flow velocity from both ends as the cold (or hot) fraction, and the input flow rate change.

The analyses I've seen that look at a vortex tube as a cooler typically focus jsut on the cold end. (Almost nobody is interested in a vortex tube as a heater. After all, it's not that hard to make a heater that's virtually 100% efficent. Hence the lack of interest in analysing the hot end thoroughly.)
 
rodsfree said:
Very Interesting Concept!
Sir, I APPLAUDE you LOUDLY!!!!!
Click to expand...

Thank you.

Now on to the technical issues.
...
Click to expand...

The more informed opinions I can get about this project, the better. Thanks for the info. I'm just looking for as much pertinent info as possible, and knowing which pages are regarded as useful by engineers in the field is very valuable to me.

If you've got any clue where to find info about vortex tubes using water, I'd LOVE to hear about it.

Good to konw about the output preesures being low. I'm guessing the thinking there is that the tube uses up practically all the kinetic energy in the process of decreasing entropy (with the requisite waste of course). Practically speaking, I guess the plan would best be implemented with a (small) pump to repressurize the flow to the block. With a sufficiently wide rad, the radiator might not require repressurizing.

I've actually been thinking that the height of the tube presents an opportunity for an integrated radiator. In this idea, the vortex tube rises phallically from the case, with the hot output on top. I figured I'd save the feminist technology commentators the trouble! :p The hot output is fed into a plenum chamber in the form of a halo on the top of the vortex tube. A curtain of vertical (probably copper) tubes surround the vortex tube feeding flow down into an identical chamber on the bottom. I don't think fabrication would be that hard. The effect would be that the assembly looks like a fluted copper column. It'd be a high volume low velocity radiator that I think would be appropriate for the setup. Just a thought though. It'll probably be easier to just stand up a stock triple 120 rad (or two - in parallel of course) beside the vortex tube.
 
Im no engineer but my concern would be that since water is so viscous is it even feasible? Has anyone ever even made a water based VT before?

Also looking at the diagram to me it appears that you would have to have HIGHLY efficient separation of the cold and hot liquid for it to actually improve performance considering the added restriction and splitting of flow between the two loops.
 
kemist1117 said:
Im no engineer but my concern would be that since water is so viscous is it even feasible? Has anyone ever even made a water based VT before?[/i]

I don't know if it's been done before. The viscosity deal isn't as bad as you might think. AFAIK, the Reynolds number depends on the kinematic viscosity - that is the "normal" viscosity divided by the density. The difference isn't as big when reduced by density.

Also looking at the diagram to me it appears that you would have to have HIGHLY efficient separation of the cold and hot liquid for it to actually improve performance considering the added restriction and splitting of flow between the two loops.
Click to expand...


As far as I can tell, the less separation that happens in the vortex tube, the less head loss as well. That gives a wonderful tradeoff: If the VT pressure loss is higher, the temperature separation is higher too. I'm guessing that this will give a fairly generous regime of operation - assuming that it's feasible to make a water based VT in the first place!
 
I say again...

Can you provide a link to anyone in the world who has used a vortex tube to cool water?

This whole discussion is a waste of time without evidence that liquid vortex tubes actually exist and work effectively the same as normal vortex tubes, and I just don't see that kind of evidence anywhere I look.
 
arentol said:
I say again...

Can you provide a link to anyone in the world who has used a vortex tube to cool water?

This whole discussion is a waste of time without evidence that liquid vortex tubes actually exist and work effectively the same as normal vortex tubes, and I just don't see that kind of evidence anywhere I look.
Click to expand...

Who pissed in your corn flakes? Nobody wasted your time but you. I never made any claims. I'm just brainstorming on the forum hoping to get more information and feedback. If you have heard of anybody who actually tried to make a water vortex tube and documented the reasons for failure, that would be helpful.
If not, plese stop shitting on my thread.
 
If I am shitting on "your" thread it is no different than if you took a dump in an outhouse and then I went and took a dump there too....

It is YOUR inital post that assumes, and by extension implies your factual knowledge, that a water based vortex tube already exists and works just like an air based vortex tube. Everyone responding, including me at first, made the assumption that you actually have seen or heard of water based VT's and therefore have a valid reason for bringing this topic up at all. In fact your second paragraph could be interpreted to mean that you actually are about to build a system using a water VT. Most of us assumed you wouldn't do that without actually having evidence they exist.

So, since the whole freaking thread is really a discussion of water based vortex tubes that you implied the existance of, I simply ask you to provide evidence of their existance so this discussion can be validated.
 
I have in fact found one website where someone created a water based vortex tube. However, the site doesn't say anything about the resulting temperature of the output streams. It simply discusses separating oil from water (oil following the "cold" path), or using magnets to separate ORMUS water from regular water. Also the examples given on the page all show very low water pressure coming from the cold side, I know that low pressure was partly a choice of the designer, but I don't know how much greater pressure they could get really. Doesn't sound like they could get much more.

Another thing I have realized from researching this is that the amount of air output by a vortex tube is not balanced, and therefore the pressure level on each side is different. So even if you did set all this up, the CPU side would have lower pressure than the radiator side. You would probably need some way of merging the water that kept the higher pressure side from pushing back into the low pressure side, resulting in even slower water flow on the CPU side.
 
arentol said:
If I am shitting on "your" thread it is no different than if you took a dump in an outhouse and then I went and took a dump there too....

It is YOUR inital post that assumes, and by extension implies your factual knowledge, that a water based vortex tube already exists and works just like an air based vortex tube. Everyone responding, including me at first, made the assumption that you actually have seen or heard of water based VT's and therefore have a valid reason for bringing this topic up at all. In fact your second paragraph could be interpreted to mean that you actually are about to build a system using a water VT. Most of us assumed you wouldn't do that without actually having evidence they exist.
Click to expand...

First things first: detente. You are clearly more reasonable than I gave you credit for. I just get worked up when I see taunting in large fonts.

Secondly, I have no evidence that water VTs exist. However, as I have a smidge more than a pedestrian understanding of fluid mechanics, it is pretty clear to me that there is no fundamental difference between an air VT and a water VT. All you have to do is design the aparatus appropriately. The question is whether it would be feasible with accessible parts.

I really couldn't care less if it hasn't been done or even tried before. Actually, that would be encouraging. If it has been tried or doen before, that would be informative. I'm not afraid to be the first to try and fail. If this project actually results in any prototypes, it'll be a hell of a lot of fun whether it works or not. That's the point to me.

So, since the whole freaking thread is really a discussion of water based vortex tubes that you implied the existance of, I simply ask you to provide evidence of their existance so this discussion can be validated.
Click to expand...

The thread title has the word potential. To me that means something that might be. Not something that necessarilly exists already.
 
arentol said:
I have in fact found one website where someone created a water based vortex tube. However, the site doesn't say anything about the resulting temperature of the output streams. It simply discusses separating oil from water (oil following the "cold" path), or using magnets to separate ORMUS water from regular water. Also the examples given on the page all show very low water pressure coming from the cold side, I know that low pressure was partly a choice of the designer, but I don't know how much greater pressure they could get really. Doesn't sound like they could get much more.
Click to expand...

after a little googling (*thanks for the term ORMUS). googling: ' ORMUS "vortex tube" ' got me this. That answers your previous request for a link! :)

So the separation effect is clearly possible with pressures in the ballpark of watercooling. The encouraging thing to me from that acrylic water VT is that the whole immersed impeller is definitely not necessary. One big pump really ought to be able to drive the vortex tube, if not the whole system.

Another thing I have realized from researching this is that the amount of air output by a vortex tube is not balanced, and therefore the pressure level on each side is different.
Click to expand...

The "cold fraction" is typically not very big, that's true. But that's because most vortex tubes are built for spot coolers, where the goal is maximum temperature drop on the cold side. With water, you can't go for much temperature drop. Condensation will become an issue long before freezing. And freezing isn't really that far below room temperature. I know phase change and TEC people prep their rigs for condensation, and it's not a big deal. I just don't think it's a good first step to go for a big drop on the cold side.

So even if you did set all this up, the CPU side would have lower pressure than the radiator side. You would probably need some way of merging the water that kept the higher pressure side from pushing back into the low pressure side, resulting in even slower water flow on the CPU side.
Click to expand...

That's a very good point. That's one I toyed with for a while before I drew the schematic in the first post. The solution is to just dump them both into a reservoir. That way, you just "throw away" the pressure in the hot side, and it doesn't put backpressure on the cold side. Of course, that'll only work up to a point. If the pressure difference is too big, more drastic measures will be needed.

*edit: It's important that the reservoir be open (pressure wise). Either open to the atmosphere, or have a diaphragm that maintains equal pressure. Such a diaphragm could be as simple as a piece of saran wrap over a good sized hole, or a bag if necessary.
 
You know that the technical term for a Vortex Tube is a Ranque-Hilsch tube right?

Most VT tubes are capable of producing almost an 80'c Temperature differential at 600kPa. That's a HELL of alot of pressure... That's honestly like 6 bar... right?
 
Arcygenical said:
You know that the technical term for a Vortex Tube is a Ranque-Hilsch tube right?
Click to expand...

Yup. Also known as Maxwell's demon (although this term has a much more general meaning in most settings), and I've seen a third name very occasionally tacked on after Hilsch - I foget who, though. I just find VT quicker to type.

Most VT tubes are capable of producing almost an 80'c Temperature differential at 600kPa. That's a HELL of alot of pressure... That's honestly like 6 bar... right?
Click to expand...

That's a whole lot of pressure, for sure. It's hard to relate numbers from air VTs to water though. And I really don't need anywhere near that kind of temperature differential either.
 
InTheFlow said:
When are you planning on making this? It sounds very interesting.
Click to expand...

I picked up the fittings for the inlet end yesterday. I've got to do some scrounging for tools, as I don't have a dremel. and I haven't settled on a final design for the throttle on the hot end.

I'm going to go door knocking at the Mechanical/Aerospace Dept. this week to see if I can get some flow meters and inline temperature sensors. I'm going to need three of each to get any meaningful data. (A temp. sensor and flow meter on each of the inlet, hot, and cold ends.)
 
Well, I just blew the majority of my project budget. I'm beyond the point of no return now. I just secured the wining bid on an Iwaki WMD-40RLT (same as the 40RT). :D I think the "WMD" is appropriate.

I think that ought to do the trick. If it doesn't, I'll gladly declare the whole project a lost cause. :p It's got the regular impeller, but with a larger everything, it puts out virtually the same head pressure as the infamous Iwaki MD-20RZ. (I'm under the impression that's a favorite for large watercooling systems.) Best thing is the pressure stays flat near the max head up to a lot higher flow rates. i.e. the 40RLT pumps 3GPM at over 20feet of head pressure!!!!

The impeller could be replaced with a Z-type (it's not cheap), but I think it's better this way. I can operate at pressures attainable by a 20Z, and still have wiggle room on flow rate.

If this project doesn't work out, does anybody ahve a server rack they need cooled? I think this thing could do a whole lot more than one box. Then again, who knows what pumps people will be getting when the 4x4 rigs roll out. :rolleyes:
 
I'm going with a slightly more ghetto setup. Just getting a multimeter, putting some short copper pipe sections, and some thermistors. The prototype won't be hooked up to any watercooling loop, so corrosion isn't really a big deal.

I'm not really going to worry about pressure and flow rate measurements until after I've got a working prtotype. It looks like pressure and flow rate instrumentation will be significantly more expensive than temperature. Starting off with KISS seems like a good idea.

Right now, I've got most of the fittings needed for the nozzle end, a pump in the mail (hopefully), and no tools. I'll probably start a worklog thread when the pump arrives. I should have the tool situation worked out by then too.
 
I don't think that you really need to worry about head pressure.

Flow is going to be much more important to what you are trying to do.

Those Exair Vortex tubes operate from 1 to 150 cubic feet of air per minute.
One cubic foot of volume is equal to 7.48 Gallons of volume. So the pump that you are getting is going to flow just under 1CFM of water at almost zero head.

Another factor is that air is compressible and water is incompressible. Therefore with the air powered tubes we can trade the potential energy of the pressure for kinetic energy - i.e. velocity.
Remember that these things approach 1,000,000 rpm rotational velocity. In a 1/2" ID tube that is almost 1,500 MPH (if I've done my math correctly).

Water is never going to be able to do that. Just too much mass to move that quickly.

So,
What can we realisticly achieve?
Sub-ambient temperture - this should be possible. But not in the same realm as an air powered tube. If you get a 5C-10C drop then you should concider the experiment a success.
And a one of a kind water cooling rig - definately possible - do the body of the final version out of acrylic and light up the vortex with LEDs.


A suggestion - make the body of your tube as long as possible. This will give you a greater distance for the vortex to create a thermal transfer over. Most of the air powered tubes are less than 12" long X 1" in diameter. Personally, I would start with a 4 ft X 2" piece of PVC and vary up and down from there, to see what changing my pipe length and diameter would do for my temperture drops.
PVC is cheap compared to buying bigger and bigger pumps.

Good Luck!
And I'll definitly be watching for more!
 
rodsfree said:
I don't think that you really need to worry about head pressure.

Flow is going to be much more important to what you are trying to do.

Those Exair Vortex tubes operate from 1 to 150 cubic feet of air per minute. One cubic foot of volume is equal to 7.48 Gallons of volume. So the pump that you are getting is going to flow just under 1CFM of water at almost zero head.
Click to expand...

The 40RLT pumps a max flow rate of 13.7 GPM. That's almost 2CFM. Not an earth-shattering difference, but possibly significant.

Another factor is that air is compressible and water is incompressible. Therefore with the air powered tubes we can trade the potential energy of the pressure for kinetic energy - i.e. velocity.
Remember that these things approach 1,000,000 rpm rotational velocity. In a 1/2" ID tube that is almost 1,500 MPH (if I've done my math correctly).

Water is never going to be able to do that. Just too much mass to move that quickly.

So,
What can we realisticly achieve?
Sub-ambient temperture - this should be possible. But not in the same realm as an air powered tube. If you get a 5C-10C drop then you should concider the experiment a success.
Click to expand...

First, thank you for your thoughts. It's really helpful to get a lot of input about things to consider.


I've got to beg to differ with you about the importance of head pressure. I think that pressure is also going to be absolutely vital. The turbulence in the tube is going to extract energy from the flow. (Also, the decrease in entropy! :rolleyes: ) Possibly a LOT of it. That energy has to come from a pressure drop, because - as you said - water is incompressible, so the total flow in has to equal the total flow out.

As for your comments about what to expect, I fully agree. I'm definitely not hoping for anywhere near the temperature drops of air-tube spot coolers. The good thing is, my goals for the project aren't anywhere near the realm of those air tube coolers. The way I see it, if I wanted a large temperature drop on the cold side, I'd be better off just getting a Vapochill. If I see any statistically significant temperature drop on the cold side that's bigger than about 5C, I'll be thrilled. But if you couple that with the commensurate temperature increase on the hot side, and factor in what that does to the radiator's thermal dissipation, you've potentially got a nice "middle road" solution between the conventional WC setup, and the TEC/phase change solutions.

And a one of a kind water cooling rig - definately possible - do the body of the final version out of acrylic and light up the vortex with LEDs.
Click to expand...

If this works at all, I definitely want to do the final version in acryllic. But that's a big if... Until I know for sure if this works, the vortex tube isn't going anywhere near a computer.

A suggestion - make the body of your tube as long as possible. This will give you a greater distance for the vortex to create a thermal transfer over. Most of the air powered tubes are less than 12" long X 1" in diameter. Personally, I would start with a 4 ft X 2" piece of PVC and vary up and down from there, to see what changing my pipe length and diameter would do for my temperture drops.
PVC is cheap compared to buying bigger and bigger pumps.
Click to expand...

Right now, I've got a 5' piece of 1.5" PVC. But after fabrication, it might be closer to 4'. (I don't know how many tries it's going to take to cut/drill good nozzle holes.) So that's the ballpark I'm working in. I'll probably have quite a few prototypes by the end of the process, though. Good thing PVC is so cheap.

Good Luck!
And I'll definitly be watching for more!
Click to expand...

Thanks!
 
At least you are being very realistic in your expectations. ;)

And it's a great project just for the experiment itself.

And another thought....
Point the hot end up and the cold end down.
With the small tempertaure differentials you are probably going to see every little bit will help. Wouldn't want rising heat to ruin your data.
 
rodsfree brings up some very good points, but I have some doubts. For one, I do not think you'll have head pressure. For two, I do not believe that your backflow will be enough to cool a CPU. Although, I must confess that I hope you prove me wrong. ;)
 
rodsfree said:
At least you are being very realistic in your expectations. ;)

And it's a great project just for the experiment itself.

And another thought....
Point the hot end up and the cold end down.
With the small tempertaure differentials you are probably going to see every little bit will help. Wouldn't want rising heat to ruin your data.
Click to expand...

Heat only raises when convection is taking place. Convection only happens with compressable gasses that expand with heat (making them less dense and more boyant)

also, I am not an expert on vortex toobs (I have never even seen one), but from what little mysterius information I have seen on them, it appears that their cooling abilities come from using the venturi effect (fast moving fluids create a vaccume) to lower the pressure of one part of air (the resulting expansion makes cooler temps) and I bilieve this is only possible with compressible fluids.

someone was commenting on pressure... 6 bar is nothing (with air). thats about 84PSI... the tires on a bicycle have 85PSI in them...

also the origional refrence link mentioned equations for both compressable fluids and non compressable fluids. So I guess it implys that it would work with water. I am however doubtfull.
 
nhusby said:
Heat only raises when convection is taking place. Convection only happens with compressable gasses that expand with heat (making them less dense and more boyant)
Click to expand...


Natural convection occurs by the same mechanism as ice floating in water - difference in densities of materials. It doesn't have to do with compressibility, and does affect fluids like water.
 
zer0signal667 said:
Natural convection occurs by the same mechanism as ice floating in water - difference in densities of materials. It doesn't have to do with compressibility, and does affect fluids like water.
Click to expand...

right... convection occurs due to the boyancy of an expanding heated fluid.

If I remember correctly the density of water doesnt significantly change until within a few degrees of freezing. As water nears freezing point it begins to condense to about 99.8% of its previus volume, but then when it freezes it expands to about 110% of its normal volume because of the way the molocules align. its been a long time since chemistry, so my numbers might be off... but thats the way I remember it.

also convection takes a considerable amount of space, in order for the heated fluid to be able to flow past the non-heated fluid.

and as far as I know, if a fluid cannot be compressed it cannot expand and condense because its properties dont allow conversion between density and heat. (thats more of an assumption on my part)

please, if I am wrong, prove it and explain it to me. ;)
 
You must log in or register to reply here.
Back
Top