View Full Version : Liquid verses direct phase change effciency?
saratoga
03-17-2004, 08:49 PM
Initially I'd think a direct phase change system would be ideal since it offers less chance for heat to leak and put the least material between the die and the phase change.
However it does limit the size of the evap, so in so much as the evap is the limiting factor, the chilled liquid approach has an advantage.
I'm really not sure how different the two would really be. Please advise.
TheDogFather
03-18-2004, 12:37 AM
At minus 60 youre still going to be limited to the same evap size, whatever fits on your board once its insulated.
TDF.
water_cooler 20
03-18-2004, 04:17 AM
pretty much the main pro of the direct die is it's colder
and the main pro of a chiller is it can cool multi places at once
I hit 700 posts:D :spam:
Originally posted by saratoga
I'm really not sure how different the two would really be. Please advise.
The differences R like night and day. It would take a dual direct-die set up to outperform a single high performance chiller. With chillers you have the added flexibility of being able to cool multiple water blocks with a single unit. It is also easier to change out a water block then it would be to braze on a new evaporator new processor technology evolves.
Maybe I should clarify... Both have distinct advantages and disadvantages. Is a direct die evaporator more efficient, yes. Is it more flexible than a high performance chiller, no, not in my opinion.
afireinside
03-20-2004, 09:13 AM
OCme, you dont have to change the evap with everysocket. You might have to reinsulate it but all you need to do is make a new plexi hold down or drill more holes in the current one.
Originally posted by afireinside
OCme, you dont have to change the evap with everysocket. You might have to reinsulate it but all you need to do is make a new plexi hold down or drill more holes in the current one.
Your right, you don't have to put a new one on every time because sometimes you can make the old one work, but all the work involved in making the old one work and getting it to seat just right is in my opinion much harder than using a water block that is manufactured for a particular chip & socket. Besides if I wanted to cool ten cpu's with one chiller theoretically you could do that. Now, let me see a direct die setup do that! :banana:
afireinside
03-20-2004, 10:50 AM
Good luck building a chiller than can handel 2000 watts of heat :p
Originally posted by afireinside
Good luck building a chiller than can handel 2000 watts of heat :p
You are right, the type of cpu you are using and how hard you are pushing it would be determining factors. Building a chiller that would dissipate 500 to 1000 watts would not be hard to do though from what I have been told.
Gary Lloyd
03-20-2004, 11:50 AM
Hmmmm... I've worked on 12 megawatt chillers.
saratoga
03-20-2004, 10:59 PM
Originally posted by Gary Lloyd
Hmmmm... I've worked on 12 megawatt chillers.
Are such systems electrically driven or chemical? Just wondering . . .
Gary Lloyd
03-20-2004, 11:57 PM
These are driven by electric motors. When we start them, we have to notify the power company, so they can bring extra equipment online.
Originally posted by Gary Lloyd
Hmmmm... I've worked on 12 megawatt chillers.
Damn, how many computers could you cool with one of thoes. LOL Chillers obviously have a lot of potential.
Originally posted by Gary Lloyd
These are driven by electric motors. When we start them, we have to notify the power company, so they can bring extra equipment online.
what were they used to cool :O
Gary Lloyd
03-21-2004, 02:47 PM
The Renaissance Center in Detroit. Four towers. They have three of these 12 megawatt chillers, and one little 4 megawatt chiller.