Passive Thermosiphon Solar Still
A solar water still built from under $60 of hardware-store parts, driving its own circulation by convection alone — no pump, no moving parts.
The problem
Solar stills are only useful where they’re affordable. Anything with a pump adds cost, a power requirement, and a part that will eventually fail — which defeats the point in exactly the places a still is most needed.
The constraint was a still that runs on sunlight alone, from parts someone could actually source, for under $60.
How it works
The design drives its own circulation by thermosiphon. Water in the collector warms, becomes less dense, and rises; cooler water sinks to replace it. That density difference is the entire pump.
Which means the geometry is the pump. Get the height difference between reservoir and collector wrong, or the tube runs wrong, and the loop never establishes — there’s no motor to cover for a bad layout.
That sketch is where the layout got settled: a serpentine collector run in ½-inch PVC, driven by the pressure difference the sun’s temperature gradient creates, feeding a tank with a one-way vapour path to a condenser.
Simulating before building
Because the whole design depends on convection actually establishing itself, we modelled the collector in OpenFOAM before committing to a build. A thermosiphon that doesn’t siphon is just a pile of tubing.
The simulation showed the convective flow developing through the serpentine run, and put the expected output at roughly 3 L/hour — comfortably past the 2 L per day the requirement asked for.
Building it
The collector is a serpentine run of PVC on a timber frame; the reservoir is a Home Depot bucket, elevated to give the loop the head it needs.
Result
- ~3 L/hour predicted output — from the OpenFOAM model, not a measured yield
- Under $60 in parts, all from a hardware store
- No pump, no moving parts — passive from end to end
- Convective flow and pressure containment both confirmed on the physical prototype
Worth being precise about which of those came from where: the flow behaviour and the sealing were verified on the build, but the 3 L/hour figure is the simulation’s prediction. A full sunrise-to-sunset yield test on the physical still is the piece still missing.
We presented at an end-of-semester expo to judges, classmates, and members of the community. Our professor consistently ranked the project at the top of the class.
