Some background
Peltier devices have been around for a long time, and there are many articles, websites and YouTube videos on their application and limitations.
This website is different, and this article gives some background why and how it came about.
Generally, the common wisdom is that Peltier devices are interesting but not very efficient, and suited only for very narrow applications. In actual case, they are unbounded in their efficiency, with both theory and experiment confirming efficiencies over 2000% (COP > 20) although this is . In real world conditions, it’s not difficult to maintain a COP for heating in the range of 2.0 to 3.0 (200-300% efficiency), making them competitive with compressor technology taking into account the infrastructure required to operate.
So why is the common wisdom misguided?
In a nutshell, most set ups have one or more unintuitive mistakes that significantly reduce the heat pump effectiveness. Datasheets can also give unrealistic expectations for a single Peltier, often suggesting cooling potential of 85W or more. The lukewarm results in practice encourage users to push Peltiers harder, often well into the region where the waste heat far exceeds any cooling ability. This then reinforces the false image that while Peltiers do seem to work, they are terribly inefficient.
In more detail, there are five key points as to why this website is different:
The thermal circuit: for any heat pump, understanding the role of the thermal circuit is critical to successful design. The model used in this website is not complicated and uses basic electrical components (resistors, voltage sources, current sources) to represent the thermal elements, and can be easily understood with high school level circuit analysis.
The full circuit: many websites use the information from Peltier datasheets to predict performance, and ignore the external thermal circuit such as hot and cold side heat exchanges, cold side insulation, the Peltier contact thermal resistance and so on. For efficient application of any heat pump (including compressor based systems) knowing the full thermal circuit is critical.
The values: while a model is fine, it only works if typical values are known, such as the thermal resistance of a heatsink with a cooling fan, the thermal characteristics of the Peltier itself, or the insulation of a cool box. This website introduces typical values for Peltiers, common heat exchanges and insulation systems, and is planned to include more comprehensive data in the future.
The twists: many aspects of thermal systems are counterintuitive, for example water has excellent ability to store heat but a terrible conductor of heat; adding Peltiers in parallel can often make set ups worse, the common use of large heatsinks with cooling fans are actually not that efficient, and the reason why stacking Peltiers is usually ineffective is more subtle than expected. The model used in this website (thermal circuit) helps to show what’s going on.
Measurement: a model is only useful if it can be confirmed by measurement, including not only temperatures but actual heat flow. While typical users won’t be able to do this, the information on this website is derived from experiment which provides confidence in the model proposed and the conclusions from the analysis.
Why is this possible when others have failed?
To be clear, this didn’t appear overnight. It’s culmination of about 3 years of experimenting with Peltiers. Initially many experiments had success rate of about 5%, a lot of stupid mistakes were made, with the early work was more like fumbling around in the dark. But there were occasional successes, and this sparked a desire to figure out why some things worked while others set ups failed miserable. Pieces started falling into place, eventually leading not only to model but the experimental data as shown on this website. These days around 90% of experiments work, and those that don’t are much easier to diagnose.
In the background, this website is really a spin off of MEDTEQ: a small business that develops custom test equipment for evaluating medical devices according to international standards. Medical devices of course should be accurate, and the test equipment used to check them should be about 5 times more accurate again. And if you sell that equipment you need reference equipment which is about 5 times more accurate again. The result: a facility that can measure temperature with 0.001°C resolution, DMMs and power supplies with 0.05% precision, scales with 0.05% resolution and more. Combine that with nearly 40 years of experience in test and measurement, the ability to make custom equipment for measuring temperature and heat flow, 24bit ADC multichannel datalogging, custom software and a keen interest to understand Peltiers and it all adds up to something different.
Enjoy!