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:
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.
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:
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.