Since87
10-08-02, 09:41 PM
I created some spreadsheets to calculate the effectiveness of pelts under a variety of operating conditions. The calculations are based on data extracted from the paper, "Universal Thermoelectric Design Curves" by Richard J. Buist. (http://www.tetech.com/publications/pubs/IECEC1980RJB.pdf)
I came up with equations which match the graphs in the document with reasonable accuracy. (over operating regions of interest to overclockers) Using this spreadsheet you can get some gauge of the cooling effectiveness that can be obtained with a wide range of pelts under a wide range of circumstances.
The numbers calculated are somewhat idealized. They don't take into account heat leaking from the waterblock to the coldplate through the clamping screws, spreading resistance in the coldplate, and other similar factors. Actual temperatures will be somewhat higher than the value calculated.
The first spreadsheet is written for calculating the effect of a single pelt. However by selecting the inputs correctly, the behavior of a dual pelt system like Silver's can be simulated.
The second spreadsheet is written for calculating the effect of stacked pelts. It shows that it can be done effectively when the input heat is substantially less than the Qmax of the pelts in the stack.
I'd appreciate any comments, corrections, or suggestions people have on these. I'm sure these could be much better, but I'm not an Excel expert by any means and don't know how to use a lot of the nifty features. If someone is interested in putting up a website to do these calculations, I'd be happy to help.
I came up with equations which match the graphs in the document with reasonable accuracy. (over operating regions of interest to overclockers) Using this spreadsheet you can get some gauge of the cooling effectiveness that can be obtained with a wide range of pelts under a wide range of circumstances.
The numbers calculated are somewhat idealized. They don't take into account heat leaking from the waterblock to the coldplate through the clamping screws, spreading resistance in the coldplate, and other similar factors. Actual temperatures will be somewhat higher than the value calculated.
The first spreadsheet is written for calculating the effect of a single pelt. However by selecting the inputs correctly, the behavior of a dual pelt system like Silver's can be simulated.
The second spreadsheet is written for calculating the effect of stacked pelts. It shows that it can be done effectively when the input heat is substantially less than the Qmax of the pelts in the stack.
I'd appreciate any comments, corrections, or suggestions people have on these. I'm sure these could be much better, but I'm not an Excel expert by any means and don't know how to use a lot of the nifty features. If someone is interested in putting up a website to do these calculations, I'd be happy to help.