This article requires that you've read and understood the previous topics, in particular "Multiple TEC's"

The above topics can be summed up into 3 short statements.

1a. Lower the input voltage relative to Umax, the more heat can be moved relative to input wattage.

2a. Lower the input voltage relative to Umax, lower the maximum temperature difference between the two sides of the TEC.

3a Lower C/Ws of the hotside will result in lower hot and cold side temps.

This article will help you select the right Qmax of a single or multiple TEC's and the correct input voltage to run them at to achieve the desired results. The Sweet spot is a point where all the factors come together to achieve the desired results while using the lowest electricity possible.

Before we start

Before we can start we NEED to know two things

1b The heat load to be cooled in watts.

2b The temperature you'd like to cool the load (1b) too.

For the purpose of this document we'll be using a heat load of 200 watts and a target temp to 0c.

For the moment we'll pretend our hot side C/W is an impossible 0 and a hot side temp of 35c.

Umax ralative to Qmax

There is two obvious routes we can take when selecting input voltage and Qmax for our TEC's we could ether have one low powered TEC running at a high input voltage or one high powered TEC running at a low input voltage.

1c 474watts of Qmax can cool our load to 0 while running at 70% of Umax

Or

2c 949watts of Qmax can cool our load to 0 while running at 44% of Umax

While both systems are cooling our load to the same temp they aren't using the same amount of electricity to achieve this. Our first system with 474watts of cooling is using 372watts of electricity to achieve a temperature of 0 while the second system uses only 280watts of electricity to achieve the same temperature.

One could conclude that more Qmax is better. However this isn't the case.

9490watts of Qmax would cool our load to 0 running at 28% of Umax while using 969watts of electricity.

As you can see there is a point where increasing Qmax and decreasing the input Voltage no longer pays off. We need to adjust these two factors to fined the Sweet spot. In our case that's 1000watts of Qmax and 40% of Umax. If we double our load the sweet spot becomes twice the Qmax 2000watts while running at 40%.

At this stage it's very hard to decide which configuration is better when comparing 1c to 2c. 1c uses more electricity to run and there for will cost the most in the long run however 2c will cost more to implement initially as it requires twice the Qmax.

Hot Side C/W's

C/W's (centre grade per watt) is a measure of how efficient a cooling system is at removing heat. Lower the number the better. For the purpose of this article we'll be using a hot side C/W of 0.1 and an ambient temp of 35c. If we didn't TEC cool our 200watt load and simply used our 0.1 C/W cooling system to cool our load our load temp would be 20c above ambient or 55c.

The numbers below shouldn't be considered accurate. They are just used to illustrate my point.

Once we apply a C/W of 0.1 to the hot side of 1c our load temp now becomes 45c with a hot side of 90c !

If we apply a C/W of 0.1 to the hot side of 2c our load temp becomes 39c with a hot side of 78c

2c becomes the clear winner now because the hot side is cooler than 1c. The hot side is cooler because less electricity is being used and therefore there is less heat to be cooled by the cooling system. This difference in hot side temp more than makes up for the larger difference in hot and cold side temps of 1c.

The C/W obviously affects the "Sweet spot" by moving it down the voltage.

The sweet spot of 1c now becomes 52% of Umax instead of 70%. The reduction to 52% also reduced the load temps slightly. Any increase in voltage beyond 52% results in wasted electricity and /or increased load temps. The affect is the same on 2C but to a lesser degree the sweet spot now moves to 40% from 44% of Umax.

Interestingly enough 1c now uses less electricity than 2c. However 2c still produces lower temps. This is a clear indication that 1c requires more Qmax (more TEC's)

That happened to 0c ?

Our original target temp was 0c. However after adding a hot side C/W of 0.1 the best we managed was 39c.

Due to this hot side C/W it's impossible to cool our load to 0c no matter the Qmax or input voltage. The only way to achieve 0c would be to decrease the hot side C/W.

It would seem we'd require a truly awesome cooling system to be cooling the hot side of the TEC's to achieve a 0c temp.