A few years ago I got an itch to grow a garden.  The only problem was that we lived in a small apartment in Wallingford with no access to a garden space, and the length of the waiting list for P patches in the city was measured in years.

I'd always had an interest in hydroponic systems as a means of cultivation, and so I started thinking about how I could scratch my itch by building a small hydroponic garden in our apartment.  It became pretty clear that yeah, I certainly could, but with one big constraint: it couldn't leak.  Ever.  

All of the hydroponic systems I was familiar with had a few components in common:

  • An external liquid nutrient reservoir of some kind, usually a glorified bucket full of water with some dissolved fertilizer.
  • A trough or other container for the plants to grow in
  • Plumbing - namely, a pump to move the liquid to the plants through some tubes, and a return system to bring the liquid that wasn't taken up by the roots back to the reservoir through some other tubes.  Sometimes the same tube.  I digress.

The issue is the third point.  All of that plumbing means opportunities for leaks.  Joints need glue, threads need tape, tubes get caught in vacuum cleaners, and at the end of the day you're almost guaranteed a leak somewhere.  Resolved that neither our downstairs neighbor nor our landlord would appreciate 5 gallons of nutrient-rich water under the floorboards, I designed a simple system that couldn't leak.  

Basically, the solution is simple: make the reservoir and the plant trough the same container.  If the job of the pump is only to move water from beneath the roots to above the roots, it should be possible to make sure there are no external tubes whatsoever.  Furthermore, if the total volume of the nutrient solution isn't enough to even overflow the container where the roots are, the worst that can happen is that your pump moves all of the liquid to the roots, pumps on nothing until it craps out, and you're out a pump... but your neighbor stays dry.

A crudely drawn self-contained hydroponic garden.

The entire system was run off of a small single board computer that ran an artificial light and the pump on a schedule, and it worked!  We grew 6 kale plants in a system like this for about 6 months, which yielded give or take about a bunch of kale every week.  By the end of the experiment, those kale plants were some tough old plants, but they were still bearing leaves.  

There will be much to come about the successor of this system as my friend Amelia and I resurrect it.  Next up will be more about modeling this garden in software and how the COMPUTER actually runs it.