The OpenAg Food Server™ is the Yogi to the OpenAg Personal Food Computer’s™ Boo-Boo. Based at the MIT Bates Linear Particle Accelerator Research Center, my team and I are transforming four shipping containers into Mars-worthy indoor farms. The infrastructure we've designed allows for precise control over a plant’s environment, including variables like air temperature, humidity, light spectrum, photoperiod, nutrients, and others. Food Server also collects data on its incubated plants and environments. OpenAg is using this agriculture lab to correlate plant characteristics like flavor and rate of growth to the environments from which they were grown and expressed.
As part of our mechatronics team, I'm designing and building out Food Servers' mechanical components and infrastructure, including LED lighting, water distribution, and HVAC infrastructure.
I designed and built the Food Server’s hydroponic rack systems, now found in all of our active units in Boston and India. I iterated through the design enough times to have officially patched all of its leaks. Once requiring two dozen components and a near-Home Depot aisle’s worth of tools to assemble, the rack is now built from a simple bill of materials and no hand tooling at all. It has a bunch of features I like:
● Vertical space — Up to 150 plants can be grown in one rack. Each rack houses up to three vertical layers, allowing for 24 sq-ft of growing capacity in an 8 sq-ft footprint.
● Waterbot connectivity — The rack plugs into OpenAg’s automated fertigation system, WaterBot, which I also helped design. Waterbot delivers programmed nutrient, pH, and stress inducing solutions to each rack in Food Sever. The feature enables OpenAg to run 12 unique root water environments in a single shipping container.
● Multi-purpose — Racks can easily change modes from shallow water culture, deep water culture, nutrient film channel, or aeroponic propagation.
● Mobility and modularity — Precisely 12 racks fit inside of Food Server, leaving a middle corrider down its center for researchers and engineers to work and swap them out.
One of Food Server’s main points of environmental control is its lighting system. I desigend the Food Server’s LED lighting system under a set of electrical, mechanical, and biological design constraints. The panel’s form factor, power consumption, spectral output, and heat emission properties guided my final design.
I also led the panel's production from prototype to a manufacturing run of 60 units. Intern Alex Flamm from Cornell University, pictured above, assisted me with assembly. Aren’t interns the best? After prototyping in-house, I coordinated contracting out of house for the final production run.
I’m now onto creating a custom chip on board (COB) lighting system for the OpenAg Tree Computer project. Check out updates on my Tree Computer page as it progresses.
As engineering tends to go, these new lights created complexity in our HVAC system. Adding light into the box implictly added heat, which our biology team—and our plants—didn’t want. Extra heat? No sweat. The OpenAg mechatronics team and I hacked our Triton shipping container’s HVAC system to reroute its airflow, and exhaust its high temperatures. The end result is an homogenous atmospheric environment across all points inside of the chamber. Our design reduced the floor-to-ceiling temperature delta from 12C to less than 1C. We now have precise control of Food Server’s air temperature and humidity to match the control of its LED’s.
I’ve also assisted on other infrastructural scale elements of the Food Server, like this waterproof floor membrane installation. Spilling and water overflow was an issue while I was running OpenAg’s Food Server v1.0 at the Media Lab. The mechatronics team and I made the entire floor of Food Server 2.0 a drain itself, so there’s no such thing as a spill while you're inside of it.
OpenAg Food Server™ v2.0 is an evolution of version 1.0, designed by our Super PI, Caleb Harper, and formerly located at the Media Lab. After I oversaw the system in 2016, I learned from its original design—things I liked, things I didn't like. Our OpenAg Food Server™ v2.0 array now has all sorts of updates in its mechanical design, automation, and sensor integration. Pictured above, a shot I took of v1.0 with Tom in 2015.
After building out hardware and infrastructure components, I oversaw OpenAg’s biology experimentation and horticulture inside of the container labs from 2016-2017. For the work, I raised hundreds of plants, among a few dozen species, from seed to harvest, to sample vial.
With Food Server, using the tool has been (almost) as fun as building it. The controls and data collection systems I’ve helped build allow for new kinds of research questions that I couldn’t ask with my previous lab’s standard growth chambers:
● How do scheduled water table changes create better quality cotton?
● How do environmental stressors, like chitosan and other insect proteins, enhance plant stress response and flavor?
● Can cotton be grown aeroponically?
● Using our data, what environments can machine learning suggest we implement inside of Food Server for improved flavor, size, and other traits?