Tuesday, August 23, 2016

A refrigerator that works by stretching rubber bands

Handheld CNC router: https://shapertools.com/

Rubber band thermodynamics: https://en.wikipedia.org/wiki/Rubber_band#Thermodynamics
Rubber band bicycle wheel:  https://www.youtube.com/watch?v=dBXL93984cQ

Thermal camera in this video:  Opgal Therm-App 25Hz: http://www.therm-app.com/ Let me know if you're interested in seeing more detail about this device in a future video.

Support the creation of Applied Science videos: https://www.patreon.com/AppliedScience


  1. Hey Ben

    Thanks for this marvel - very inspiring. Yes: show us more details!



  2. I initially thought your timing belt was the rubber band. On first thought, this might make the mechanism much simpler (and a continous process). Also it just needs two small holes instead of a cutout for the whole wheel (plus drive wheel axles).

  3. This polymer-based heat pump literally made me laugh out loud. Bravo!

    I have a challenge you may enjoy. Create a circuit producing very high frequency noise that can be sampled digitally to produce streams of non-algorithmic random data (based on some physical phenomenon.) The noise cannot originate electrically (or via inductance/radio) and should not be predictable after having been sampled. The higher the frequency and sampling rate, the better. The simpler the design, the better - Ideally a mediocre EE would be able to look at the circuit and agree that the device cannot generate predictable data.

    The application is to reduce the data into physically random data that is not based on an algorithm or (reducable) algorithm. (Note: Ring oscillators don't count - they're essentially digital and too easy to hide in the design. Also, extra points if you stay away from all but the simplest of ICs.)

    If you can do this simply and cheaply, I will buy you a cookie. Or...1000.

  4. I've been thinking about how to actually make a heat pump out of this phenomenon. This might be really interesting to build and test the performance (but, alas, I have no access to a shop...): http://i1325.photobucket.com/albums/u628/amguilmet/Rubber%20Band%20Heat%20Pump_zpspogz51bm.jpg

    It's pretty simple and self-explanatory, but even though I think it would work well enough to warrant an actual build, it'd be a PITA to _truly_ optimize the design. You'd have to place the source and sink blocks in the most ideal location, use the most ideal contact surface profiles to maximize the amount of time the band spends in contact with the contact surfaces, while minimizing the amount of heat generated due to the friction from the band sliding while in contact with the surfaces. Urg, too complex for me.