|
|
Heatsink/ Heat Pipe / ThermoSiphon Cooling The cat will only make the mistake of putting its paw by your HSF once. :) Also the place to discuss the new high end heat pipe goodness. |
Thread Tools |
03-17-2004, 08:44 AM | #76 |
Cooling Savant
Join Date: Nov 2002
Location: portugal
Posts: 635
|
Just a quick note:
I think with peltiers and a cascade block, the peltiers cooling the water would proove to be a good sistem in terms of overclockability. But search here for 'stirling cycle engines'. They are the future.
__________________
back. |
03-17-2004, 10:33 AM | #77 |
Cooling Neophyte
Join Date: Mar 2004
Location: UK
Posts: 81
|
Thanks for the info jaydee.. This makes me wonder if it's Zog's text size which causes the side-scroll bar to appear? Are you using a large character size in IE, Zog? I will have to check the site out with larger text.
BigBen: Thanks for the link to the article.. I will check it out before long. As well as the text rewrite for the web-site: More emphasis will be placed on performance data in future. EDIT: I searched for stirling cycle engines.. All I found were large outdoor coal/wood power generators. I read the piece on stirling engines for CPU cooling though. Interesting. Re: Cascade.. I wonder if it would benefit peltier plates as well? From what I've read - they focus the jets over the area of the CPU core whereas the average cpu pelt is 50mm square. Perhaps the jets would still serve to cool the larger surface area of the peltier quite well? What does the developer have to say about this? Last edited by feathers; 03-18-2004 at 12:11 PM. |
03-19-2004, 03:15 PM | #78 |
Cooling Savant
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
|
Swiftech claims seem reasonable.
Using Kryotherm ,Waterloo, Since87,and Delta/img Get theoretical/empirical results:: EDIT: Notes on last graph Used Since87 to cslculate 220w Peltier performance in perfectly "Insulated condition"(Ri= infinity). Then iteratively corrected to allow for Insulation Losses/Gains using Kryotherm Rc Using my understanding that the Spreading Resistance is dependant on the Contact Conductance(h) of the Peltier-Coldside/Coldplate. I equate "h" to Contact Conductance of the Hotside/HS( = Area x 1/Rh). Giving h=1786 W/m*m*K (0.0025x 1/0.224) and h=2516 W/m*m*K Waterloo gives a "Total Conductive Resistance"(subtracting Contact Resistance from Total Resistance) of 0.103c/w for both cases. TIM resistance of 0.1c/w (for 10x10mm die) is added to give an Rc=0.203c/w Ri Insulation (Ri) =1 This is a real beast. It is used as a tool to estimate the secondary cooling/heating of the die.Nigh impossible in the thermal environment of a motherboard to use anything other than empirical values.However with very liitle or no data available this difficult.My present preferred value for CPU in motherboard is Ri=1( maybe 0.5c/w is better in view of this . Dunno ) Last edited by Les; 03-20-2004 at 05:01 AM. |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|