Michael Ethridge | Mechanical Engineer

For this research opportunity, I studied the efficacy of thermoelectric materials to produce effective cooling. At peak testing, I was able to produce a temperature of -55 degrees Celsius in the cooler. 

I began this project by teaching myself C# and C++, the two languages that those who have worked on the project in the past have used to communicate with the power supply, thermocouple, and data collection interfaces. I trained myself very quickly looking through the 600+ line long codes that has worked in the past to learn the syntax and was able to experiment very quickly. The device worked by sticking the hot end of the thermoelectric cooler to an anti-freeze based chiller capable of reaching -35 degrees Celsius, insulating the cold end with insulating foam, and running a current through the material, producing a temperature gradient over the device and cooling off the cold end to temperatures as low as -55 degrees Celsius when current and power are optimized. The cooler short circuited one day, and I discovered that moisture had gotten into the cooler and froze the circuit. While waiting on a replacement to come in, I worked on improving the testing rate by improving the method of how data was transferred between the computer and interfaces. Originally, the method of communication between the power supply/multimeter and the computer was inefficient because it would send the data points back one at a time, leading to the testing rate only being 6.67 measurements per second, which created inaccuracies when measuring transient effects on the device. With my new method I coded, the system gathered all of the data points on local storage within the multimeter, then sent over the full trial to the computer in one transmission. This increased the testing rate to 50 measurements per second on average, matching the conditions present in the outlets of the building and therefore completing measurements in a much more efficient and accurate way.