I just read your article on the Peltier Radiator and found it interesting. I have some input on the results of your experiment. I have a background in mechanical engineering, with specialization in HVAC (Heating, Ventilation, and Air-Conditioning).
For the past 12 years I have worked with, and on the design of, large commercial Air Conditioning systems. Essentially what you have built is a what is called a chiller. Water is chilled at a remote location and pumped to another area to remove heat from the load. There are a lot of variants that can play a part in the results you got. Most likely it is flow and the load calculation on the TEC. Here is a simple formula to calculate load when water is the medium:
Q = 500*GPM*(Delta T)
Q = load in BTUh*
500 = constant
GPM = flow in gallons per minute
Delta T = temperature difference in Fahrenheit between the supply water and the return water
Play with the formula to find the desired results. If you know the load (e.g. 100 Watts) and the Delta T, you can figure out the flow you need to achieve the desired results. Assuming 100 watts to cool and a Delta T of 5 F, then simplifying and substituting:
GPM = ((3.412 * watts)/Delta T)/500
GPM = (.006824 * watts)/Delta T
GPM = 0.6824 / 5
GPM = 0.136
This means you need a pump that will move at least 8.2 gallons per hour (60 * 0.136) through the system. Note this is measured flow, NOT pump rating.
If you have a 70 Watt TEC, you may need a BIG heatsink to remove all the heat that the TEC generates. Water is the best refrigerant; air on the other hand blows. 🙂 Usually it requires a large surface area to remove large amounts of heat.