I'm planning on making a small water cooler for a project. I need to cool approximately 1.5L of water and have two TEC1-12706 (12v 6amp) peltier modules, and hoping to cool to as low as 6 degrees Celsius. Would it be more efficient/effective to instead use a single Pelt such as the 12715 (12v 15 amp) or run the two separate lower rated ones? Or even two 12715s at half power for example? Also any advice on calculating the ideal heat sink size/fan speed would also be very useful
Using two peltiers is going to be better than one:
1) You will get twice the heat flow (equivalent to current) and twice the cooling rate.
2) There will be less self heating assuming that you run them at a lower voltage (ie series).
It is impossible to say anything else about your system (like how low the temperature will go, or fan speed) because you need to know what your heat in and out of the system is. The insulation around the tank needs to be known as well as the temperature on the outside of the insulation (or ambient temperature).
The insulation (or material if it is just some plastic) will leak heat in from the environment, the pelters need to be able to remove this heat. They can provide 80C of cooling with no heat load at 15 Amp (black line). If you wanted to cool to 5C and ambient is 30C (for a high mark), then you would need 25C across the peltiers, and one would give you 130W of cooling.
If the water is stagnant, it's better to use two modules, one on each side of the container. If you wish to use a heat sink for faster cooling, place it on the hot side as this will remove heat from the module and thus allow it to operate more efficiently. From experience, I find that working the module at about 80% of full rated value is best for longevity. If speed of cooling is not a major concern, connect the two modules in series across 12 Volts and each would receive the same current (but lower) at 6 Volts each and thus you would use only 1/2 the power as in the original design.
6 degrees C might be ambitious for a single-stage Peltier cell. You may want to consider cascading cells: have a large cell between ambient and a smaller cell that actually cools the target.
For example, I would seriously consider using the large 15A cell as the outside cooler with both of your smaller cells mounted to that large cell. The smaller cells would be wired in series (as per one of the comments) to better equalize the cooling ratio. Obviously, the cool side of the smaller cells contacts the target.
The aforementioned "heat sink in the water" idea would be good. I have worked for a place that builds giant radiators for mining trucks and have a few engineers in the family.
1.) The more surface area and fins the better you will do. for air or water
2.) Water has a very high thermal mass (apparently one of the highest for everyday things). This means it takes a lot of energy to heat or cool compared to most things but will retain a lot of that warmth/cold.
Rocks also have fairly high thermal mass. other things like air can be cooled/heated quicker but will also loose their cool/heat quicker.
If you expand to a two stage cooling vessel how regularly would you need to extract/replenish the volume? Would you cool the heat produced by the peltiers with a recirculation cooling system. Generally you can put pelts with a heatsink wedged between them instead of directly stacking them. They might only activate to a change in state such as the fluid needs sudden rapid heating and your modest lowpowered series of pelts get kicked into their higher mode of operation. Adjusting the pulsewidth for them will ensure their greater longevity. Some control them quite diversify with digital algorithm software If you need to transport of cut the power first you can give it a power blast to bring temp of the fluid down perforehand. Again your insertion losses Both stages could be the same but with different volumes. By pump or gravity dispersion of the chilled fluid the output/secondary vessel could have a long shallow shape. A switch for the second stage could be time activated so that the loss of temperature on pour would be compensated then switches off or delayed from the time when deactivated. With the series pelts on the main vessel you can still achieve the correct temperature and perhaps the volume that runs into the second stage is fairly negligible for any power losses while the fluid remains still, circulated or diffused with oxygenated or carbonated gasses. What ever you transfer the the energy with like a heatsink or submersed radiatior might be as reasonable in design to hygine/cleaning as it would be for design performance. All in all do seal the peltier from direct contact. A plate of raw copper with heat transfer compound means you'd never need to open the vessel just to do some service or repairs. The copper will give the greater surface area for distribution more effectively especially if factored to variable changes in load vs ambient to desired temperatures and it will also make multiple peltsxattached to it more synchronised to function. 10deg C ambient temperature swing or more in the 24hr day means the system needs variable and automatic control usually taken care of by a programmable thermostat with a variety of adjustable parameters. One of these could be set for the secondary stage whereby the liquid will not outward flow until the nominated reset values have been made. When the max is reached to system will switch the chilling until a parameter is reached to the higher range set temperature. A simple programmable thermo controller will keep you as close to 6deg without going ballistic on customized softwarebased control systems. You can also incorporate a boost function to get everything quicker working when from anambient off state.