Autoponics: Hands-free Hydro
- 1 Overview
- 2 Physical Design
- 3 Irrigation System
- 4 Lighting
- 5 Related Links
- 6 Project Log
Create an autonomous, controlled-environment hydroponic system that can support the entire growth cycle of a plant from seed to harvest without intervention by a human operator.
Smart cities of the future are cybernated, self-sustaining systems. The ability to efficiently and reliably produce food for inhabitants is a critical characteristic of these city systems. Urban agriculture and vertical farming have the potential to provide an abundance of food to city dwellers more economically and sustainably, with an output magnitudes greater than traditional farming methods. By implementing automated food production in the urban center and scaling these systems vertically inside of skyscrapers, it may be that a footprint one city block in size could produce enough food to feed 50,000 people. (see related links below)
In the near term, these urban agricultural centers will provide jobs for the local community and increase food security. Eventually, through total cybernation of seed-to-harvest, packaging and distribution processes, autoponics will replace human labor throughout agriculture.
As we stand on the threshold of a global post-scarcity society, we are approaching a milestone in the long arc of agricultural progress. Amidst unprecedented abundance, we will see the commoditization of food become irrelevant, hunger and starvation eliminated, and the notion of working for food laid to rest, freeing more of our collective intelligence to lift us towards the next phase of civilization.
A basic do-it-yourself hydropopnic system consists of a reservoir, some plumbing, emitters and a water pump, and lighting to provide primary or supplemental light. In a higher-end or commercial system, sensors monitor temperature, nutrient density, pH, system pressure, light intensity, etc. and provide varying degrees of automated self-regulation.
The Autoponics project aims to create a system that incorporates the best parts of high-end commercial systems in terms of functionality and automation, but to do it using inexpensive do-it-yourself solutions like Arduino and open-source software to make it operable from a distance over the web. We'll identify all the components of a commercial hydroponic farm and develop each subsystem. For example, high-output, narrow-spectrum LED lighting is becoming more common in grow operations with small plants, offering an alternative to high-output fluorecent, sodium vapor and metal halide lamps. However, prebuilt systems are still very expensive. We could build our own LED lighting system.
- Main reservoir
- Flushing reservoir
- Nutrient reservoir
- pH+ reservoir
- pH- reservoir
- Plumbing (PVC)
- Valves for main water line & reservoirs
- Water pumps
- Lights: Full-spectrum flourescents & ballasts, or LEDs
- Reservoir levels
- EC/TDS (elec conductivity / total dissolved solids)
- Structured light for 3D mapping
- RGB cameras
- Visual interface for remote monitoring & operation
- Logging of sensor data
4" PVC (10-100+ ft) and endcaps
- 1" PVC (10-20+ ft), angles, tees and endcaps
- 1/2" and 1/4" polytube (50-100+ ft)
- Polytube fittings (joints, tees, plugs)
PVC glue Water pressure gauges (0-100 psi)
- Water pressure sensors
- 1/2" compression fittings
- 5 micron water filter
- Regular mesh water filter
Large plastic storage bins (for reservoirs)
- UV water filter
- 1/2" ball valves
- 1/2" 2-way electric water valves (solenoid, ball)
- 1" 3-way electric valve
- Power hand drill
Hacksaw (or other device for cutting PVC pipe)
- Circle saws (2",3",4")
- pH electrode
- EC/TDS electrode
- Arduinos, ethernet shields
- High-output LEDs
- 4' T8 / T5 fixtures, ballasts, lamps
- 1" Water Pump (40-80 psi)
- Misting nozzles (40-80 psi)
- CCD/CMOS video cameras
- Peltier elements
Wires Soldering kits
Physical DesignTo make the most of the wall space in buildings, a wall-mounted aeroponics system could serve both a functional and aesthetic purpose.
One of the designs I'm playing with uses 6" schedule 40 PVC pipe for the main chambers with 3" holes for the pots. The nutrient transport manifold is 1" PVC and the sprayer lines are 1/2" PVC, fitted with misters arrayed between the pots. The lights are standard 6' shop lights, but Evan mentioned a project he's starting to develop LED grow lights. We can combine projects and use the LED grow lights in this system just as easily.
I have a 1/2 HP 1" water pump from Northern Tool we can try out.
Things to consider
We want the system to be modular, composed of interchangeable parts so it can scale to fit a living room or a warehouse by simply adding/removing units. Each module therefore needs to provide feedback on its condition--are the lights for that module working? Are the sprayers operating? Temperature, humidity, plant maturity...we should be collecting data at each module so that, imagining a system with hundreds or thousands of modules, someone doesn't have to go around checking each one visually or by hand. The system should report to us when a component fails.
- How will the positioning of the modules affect water pressure, and what size pump will be needed?
- How do we best handle drainage back into the reservoir?
- How can we build the system to be modular, so sections can be super easily added/removed at any time?
- Periodic cleaning inside the grow tubes. Access ports?
- How much energy will the system consume?
- Lighting. Flourescents? LEDs? Metal halide / HPS? Pump in sunlight from outside with fiber optics, solar tubes?
- Sterilization? We could cycle the nutrient through a UV chamber.
- Filtration? Basic mesh to catch roots and particulates so they don't clog the pump or sprayers.
- Generating our own C02 by growing yeast. The yeast is a good nutrient source for people, too.
Internal Spray Lines
Within each grow tube is a smaller spray line (1/4-1/2") fitted with misting nozzles in between each grow site. Misting nozzles can either be threaded or barbed. Barbed heads are or use with flexible piping called polytubing or micro tubing. Threaded heads can be placed into rigid PVC pipe by drilling small holes and using a tap to create the threading.
In aeroponics, the size of the particles should be small enough to provide a good mix of air and liquid to the root structure. If the particle size is too large, the roots won't receive enough air and will drown. If too small, not enough liquid will make contact with the roots for optimum growth. Misting heads like the kind you'd find for cooling off guests at a sunny outdoor venue provide a good balance of flow rate, coverage area and particle size. These generally operate between 40-80 psi, so maintaining that pressure throughout the system will be important.
DIG Irrigation products supplies a 10/32" threaded misting nozzle as well as a barbed one. If we were to use the barbed type, the internal spray line would be of the flexible polytube type. If we go with the threaded, we'll use something like 1/2" or 3/4" rigid PVC. We should test each type of tubing with various misting nozzles.
Tests at SSD on Dec 18, 2010 produced 65 psi using the 1" water pump at the space. The misters worked well at this pressure.
Another option is not to use sprayers at all, but to simply drill holes in the spray line. The high pressure and small opening might potentially produce a fine spray, or if a jet of water is produced, the reflection off the inside of the tube at high pressure may break up the stream into fine particles.
Because the misting nozzles have a tiny aperture, as small as 5 microns, they are prone to clogging. A filter must be used to prevent solid particles from clogging the system. A calcium inhibitor filter might be considered.
Sometimes algae and other invaders can get into the irrigation. A UV disinfection filter would be a good way to eliminate cysts.
A basic mesh filter should also be used to catch roots and pieces of the grow medium.
For the SSD wall garden, something like a 160 watt 72" T12 bulb could be fine for growing small herbs and lettuces. These lights are $20-40 each, but the price comes down to around $10 each when you buy a case. Larger or more complex plants would need more candlepower than the T12's can emit, so then we're getting into T8 and T5 bulbs, the VHO (very high output) flourescents that start reaching the price of metal halide lamps.
- Correll Lab is run by a CU professor whose main research focuses on Autonomous Robotic Gardening.
- Autoponics: The Future of Food Dec. 9, 2010.
- DIY Aeroponics system
- Article about PlantLab in the UK - growing food indoors in controlled environments.
- The Vertical Farm Project
- Design concepts by The Vertical Farm Project
- Despommier Interview Podcast in which Dickson Despommier, a microbiologist and ecologist at Columbia University’s School of Public Health and author of The Vertical Farm, explains how much food can be produced in a small footprint with vertical farming.
- Critique of vertical farming in The Economist
- PlantLab May 23, 2011 posted on Inhabitat
- Alpha Farm is working with Dickson Despommier on a pilot vertical farm experiment in England.
- Solar tubes and other fiber optics are a potential alternative to powered lighting.
- Helpful article that explains lumens, illuminance, foot-candles, etc.
- Urine + wood ash as fertilizer.
- Thermolectric heating/cooling might be considered for its low-maintenance characteristics, although it is less efficient than traditional coil-based refrigeration.
- Live demo of an Arduino LED controlled via a web interface
- Arduino pH meter circuit
- DHT22 humidity and temperature sensor
- DHT22 sensor library developed by Ben Adams
- Light Intensity to Frequency IC from SparkFun
- Liquid Level Sensor from SparkFun
- Photo Interrupter
- Japanese strawberry-harvesting robot
- Vision Robotics is working on an orange-harvesting robot and other agricultural robots.
- Founder of Harvest Automation talks about robots working in greenhouses
- Cityscape Farms Urban agriculture company
- Gotham Greens Urban farming in NY
- Verticrop system
- Terrashpere Systems vertical farming in Canada
- Vision Robotics builds agricultural robots
- Nuvege is a Japanese vertical farming compay
Dec. 18, 2010
Posted by Daniel Zukowski
We met up at Solid State Depot around 1PM and started drawing up some plans for a setup to test different sprayers with the 1/2 horsepower 1" water pump. Pretty quickly, we realized that we needed bins, more pipe and fittings, and a ball valve to make it all work. So, we went on a run to McGuckins to pick up supplies.
Liz, Willy and I worked on cutting, priming and glueing the PVC. We drilled and tapped 3 holes into a length of 1/2" PVC for the sprayers, and put a pressure gauge at the end of the length to get a reading on psi.
- Video of the finished test setup.
The glue set while Dan and I cruised over to Bococo. The guys there were kind enough to hook us up with 2 IDE hard drives to get some new workstations up and running at our space.
When we returned, I filled up the reservoir and turned on the pump. Nothing happened. Well, the pump was running but no water was flowing. It turns out you have to get water into the system first, at least with the approx. 1' rise we had in our intake line. Filled the pipe with some water by hand, then turned it on and it started flowing.
Before the test we had no idea how much psi we'd be able to achieve with the pump. To operate misting nozzles you need 40-80 psi. When the system started up, the regulator valve was fully open, so the sprayer line was receiving hardly any pressure. At about halfway closed, the 1" regulator valve raised the pressure in the spray line to about 25 psi which was enough to get the sprinkler-type spray nozzles working pretty well. Turning the regulator valve to about 80% closed sent the pressure up to 65 psi and the misting nozzles started producing a nice fine mist.
The brass nozzle creates the finest mist with the lowest flow rate. The white plastic nozzle creates a mist that's slightly less fine, and it looks like more water flows out.
We wrapped the spray line with a section of 4" sewer pipe to simulate the tube-within-a-tube setup for an aeroponics rig. In this video you can see the brass mister in the foreground, a sprinkler in the middle that is off (serving as a plug), and the other mister at the far end. The black thing at the near end of the sprayer line is the pressure gauge.
- Video - inside the tube
In the above video you can see how the mist from the far misting nozzle is bouncing off the tube and creating a nice chaotic blizzard of fine particles. This is good because it means the roots should get fairly even coverage as the mist blows around the interior of the pipe.
Thanks to Henry for bringing the waterproof video camera.
Dec. 24, 2010
Posted by Daniel Zukowski
- Video - Experimental idea for pipeless aeroponics
Dec. 26, 2010
Posted by Daniel Zukowski
To make a waterproof tube using this concept, one might start with two pvc endcaps and fasten something like bamboo rods along the inner surface of each, then wrap the sheet around and use a waterproof adhesive to fasten the sheet to the endcaps. The point of this approach is to limit cost, weight, and materials.
We might try printing up some of our own ABS endcaps with built-in connections for cord or bamboo.
Jan. 1, 2011
Posted by Daniel Zukowski
I tested out the spray line using nothing but 1/16" holes drilled into the 1/2" PVC. At 50 PSI, the flow rate was way to high. The jets of water coming out were just too powerful and although they could probably create a decent mist on the inside of the tube, the spray might also damage the roots.
I'd like to try another test with smaller holes. The smallest drill bit at the space I could find is 1/16", and anything smaller than that will be challenging to drill by hand without snapping the bit. Maybe I can try hammering a small nail into the PVC. A drill press would be handy right about now.
When I first started the pump, nothing happened, despite being primed with water. I took off the PVC fittings and the pump looked like it had developed a good amount of rust in the two weeks since our last test. I took a screwdriver and manually turned the piece inside that pushes the water through in order to loosen it from the rust. That got it spinning, and it worked fine after that. So, if we're going to leave the pump unused for a few days, we'll need to find some way of keeping it from rusting.
Jan. 18, 2011
Posted by Daniel Zukowski
Tested the < 1/16" holes drilled by Sebastian. The smallest of the holes (the exact size of which I will post when I can get back to the space and reference Sebastian's note) looks to be the most promising. All of the holes produce a more or less jet-like spray at ~50 psi except for one which had a burr remaining. That one produced a more divergent spray, but not quite a mist. I think the tiny holes spraying a jet of water onto the inside of the external tube will make for a usable setup that will provide a good balance between flow rate and spray characteristics while keeping cost and maintenance minimized.
Apr. 9, 2011
Posted by Daniel Zukowski
Worked with Aaron and Sarah, driving around town and picking up supplies for construction of a few different designs for boxes. I'm going to build the below design created in SketchUp:
Aaron is building a smaller unit designed to accommodate a single plant.
Apr. 16, 2011
Began construction of the box frame, using 2x4 lumber which will be sanded down, primed, and painted with a weather-resistant white paint.
The 2x4 box frame turned out to be not so much a box, but more of a torqued, skewed object that needed to be disassembled as soon as it was created. We took it apart and used the warped 2x4s for a workbench. I've changed the design to use 2x2 redwood, and selected very straight lengths at the lumber yard. A 5' box has been built, but I'm contemplating disassembling this one as well and scaling it down to a tiny little 18" unit. All I really need at this stage is a small box to start testing sensors and such. The 5' box was designed to accommodate 4' T8 light fixtures, but I think for now we can go with a standard compact fluorescent bulb just to get the basic circuits and software built and tested.
Took a break from spending $$$ on the Autoponics experiments to tackle a smaller, more tangible goal that could be used to raise some funds for the bigger Autoponics project. I've been working on the models for a couple of months now and finally have something I can prototype. If you've been to the Hackerspace recently, you know the basic concept. I'll post some images and descriptions once I have a prototype. I would post everything now, but seeing as how I'm hoping I can sell a few of these things to make a little something to fund Autoponics, I'll have to keep things on the DL until I'm a bit closer to having the marketable product ready. Stay tuned!
A research team has been assembled and the Autoponics project has received a small research grant to pursue the computer vision aspects of the system.
Follow our progress at autoponics.org