Imagine it's hot and humid outside. You see a pedal-powered air conditioning booth. You get in the booth, close the door, and start pedaling.
The pedals are connected to a shaft-driven (automotive-style) air conditioning compressor via chain and sprockets. The system is sized and geared to work at a sustainable power and cadence for a moderately fit person. (Say, around 100 watts at 80 crank RPM.) The pedals also drive two fans, one each for the evaporator (inside the booth) and the condenser (outside the booth, with a drive shaft poking through the exterior wall). There are no electrical generators or motors -- the system is entirely mechanical to minimize energy loss. Inside, the cooled breeze is aimed at the cyclist. The booth is well-insulated from the environment (like a refrigerator). The booth is stationary (it does not move). To avoid suffocating, the cyclist breathes through a snorkel that vents to the outside.
Could this work at all, given the state of air conditioning technology? Is it feasible to build a system that would keep our cyclist at a comfortable temperature, or would they be cooked inside from the trapped heat of exertion?
I want to say it's feasible, here is why. This figure (from this book) suggests that when a human pedals to make 147 watts of mechanical energy, they also produce about 400-500 watts of heat. This allows us to answer a related question: Can an air conditioner reject 500 watts of heat with 147 watts of input energy? Yes, I think so! That requires a Coefficient of Performance (CoP) of 3.40 (500 / 147), which is easily within the capability of a modern air conditioning system. On the 'web site where you buy things', you'll find mini split AC systems with a SEER of 20, which appears to convert to a CoP of 5.87. Is it possible to build a smaller pedal-driven system with similar efficiency? Am I failing to consider anything important?