Alright guys, this is a big one. I’ve got almost all the parts I need for my current watercooling project. It’s going to be using TEC units instead of radiators, however, and rather than cooling the TECs with waterblocks on the hot side they will be cooled with CPU coolers. This can be done, there are at least 2 builds on OCN that show this works as a concept. Obviously not the best solution, but the best I can afford and I’ll still be looking at around ambient temperatures under load. That said, I have a number of questions regarding the selection of the CPU coolers for the TECs.
Starting off, to compare coolers I am looking at, I made this spread sheet.
The process here is simple, i found as many reviews as I could for each cooler, listed the appropriate information, and then estimated the heat load of the CPU to get a estimated thermal resistance. Thermal resistance in this case is a even comparison tool, as the different CPU, clock speed, voltage settings, whatever across reviewers can be accounted for. That said, there is still a great deal of guesswork in this, and I understand that this is only a “best guess” sort of average. I’d like to ask for Your suggestions for the power draw of the CPUs tested. Most of these are my best guess after a 10-15 minute google for each and looking at reviews and forum posts and trying to understand what they’re getting at (surprisingly good power consumption results for overclocked CPU are few and far between). Plenty of guy on here know a lot more about this than me, and advice on “a few more watts here, a few less there” would help the accuracy of this comparison greatly.
The next question I have regarding this is the difference in coolers. When you look at something like the Thermalright Macho, it looks like it should outperform a single tower cooler like the Thermalright True Spirit, but in many reviews, it seemed this wasn’t the case. The coolers both use the same fan, so it doesn’t appear to be an issue of airflow, so why is it that a cooler with similar coldplate design, heatpipe design, and much greater thermal mass does not perform markedly better? Based on the averages from my tests, the thermal resistance comparing the two is 0.3 for the TS and 0.28 for the Macho. Better, but not incredibly so. Is this the standard difference I should be looking for in comparing large thermal mass coolers and high performance single tower coolers? Which design better suits high thermal loads? Would either design style present any specific advantage for my particular use? For this type of heat load, would compact, high density fins (justice/TS) outperform larger, more spread out fins (fuma/macho)?
The last question involves the type of cooler that I can use. Unlike CPU, where the heat in concentrated in the center over the cores, with a TEC the heat is evenly spread across the whole surface. Will it be a problem then to use a direct heatpipe cooler, such as the Thermalright TS 140 direct instead of the power? Assuming complete coverage of the TEC’s surface, that is.
I know this is a lot to throw out in one post, just looking for some answers and trying to learn a bit. 
If anyone has advice on a cooler that I have looked over that can be bought in the US for around/under $50, please share. And please please please feel free to make suggestions on how to more properly adjust the power draw of the CPUs so these results can be as accurate as possible.
tl;dr
Click on the link, comment and correct my estimations on power draw
do these numbers seem accurate? based on the estimations and data gathered?
why do big fat coolers with far more mass only outperform single tower coolers by a bit, even with similar design?
would big fat coolers be better for higher heat loads? (beyond desktop cpu load)
will direct heatpipe design be an issue for TECs, assuming proper coverage?