Method is building a Modular PFC in London

#1

Hello Ag-Hackers!

We are Adrian, Felix, and David from Method, a strategic design consultancy based in London and New York City. The three of us have started building a Personal Food Computer in London for our office as a passion project. We have some lofty goals for this one, so we wanted to share our ambitions and progress with the community.

We were inspired by Caleb’s mission of democratising and digitising food production in service of the hardships humankind will face in the future. We see huge implications for this across a variety of industries – agriculture, logistics, retail, pharma, to name a few. We think the PFC is an incredible project to get these industries to start thinking differently.

Our goal is to learn and explore first-hand by building a PFC, and to contribute our learnings and ideas back to the community. Method has a rich work history of delivering high-class user experiences in our client work, so part of our ambition is to make PFC engagement more accessible and end-user driven. At the same time, we love getting our hands dirty and flexing our maker skills (we have staff with expertise in hardware, software, physical construction, etc.)


(front facing shot coming soon :grin:)

Challenges we see so far

In deciding our approach, we knew we wanted to make sure our PFC was using and contributing to the OpenPhenome database, but found the physical build of v3.0 to be very rigid (by design, we know). We also found that this rigidity made it really expensive to fabricate the PFC according to the specs, both for the physical chassis and the mainboard. Incorporating flexibility was really necessary to make it more accessible.

When looking at v2.0, it seemed to be the opposite – the Rolls Royce of PFCs. It’s a great benchmark for how complex a PFC could get, but not necessarily feasible on a budget, especially not at once.

We want to create something in between – something modular that can start with relatively less effort and be added on to later.

We also found it hard to get up to speed on the requirements to grow from a baseline hydroponics point of view. As novice first time growers, this has required some heavy research to help us make decisions about what should be in or out of our build. We think think all of this learning can be better communicated for future nerdfarmers.

Our build

Based on v2.0 and v3.0, we’re putting together a new build that fits in between. We’re calling it an mPFC, or Modular PFC.

Key Features

  • Modular mainboard with I2C bus and 12v relays
  • Separated lighting from mainboard in order to maximise batch PCBs and add more lighting if necessary
  • Separated air sensors from mainboard so they can be suspended into the chamber
  • Chassis built from recycled & available materials
  • 3D printed connectors
  • Software compatible with v3.0 driven by OpenPhenome recipes

CPU

  • Beaglebone Black Wifi

Actuators

  • Water Heater (a heating mat under the tank)
  • 36 LEDs, 6 of each color (far red, photo red, green, blue, warm white, cool white)
  • Airflow control fans
  • Air intake fan
  • Aerator (pump + water stone)
  • Humidifier

Sensors

  • CO2 sensor
  • Humidity / air temperature sensor
  • 2 cameras (top down, side)
  • Atlas water temperature sensor
  • Atlas EC sensor
  • Atlas pH sensor
  • Water level sensor
  • Light intensity sensor

Future Add-ons (need additional research)

  • Air cooler
  • Air heater
  • Nutrient pump + tank
  • CO2 gas tank + pump
  • Dehumidifier
  • pH tanks & pumps

Progress to Date



Chassis

Our aim is to recycle parts that we already have in our office space, rather than buying new building material. So far we are using a recycled clear acrylic box (50cm x 50cm x 50cm) as our main component, and wooden beams as a supporting structure. We are 3D-printing our connectors and sensor housings. So far we’ve created 3D models in Rhino, but will have to wait for our circuit boards to arrive to see if everything fits together.

So far, we have divided the chassis into these modular parts:

  • Bottom module: water reservoir, water sensors, heating mat, base plate for plants
  • Middle module: atmosphere, air sensors, radial fan
  • Top module: LED lighting and camera
  • Side module: main boards, side camera, air sensors
  • Second side module: tanks, nutrient pumps, cannisters (we’re not currently developing this one – it will be part of a future add-on initiative)


Electronics

We’ve created a much smaller and cheaper set of PCBs:

Mainboard (1)

  • Based on the design of v3.0
  • Has 8 open slots for new I2C modules. (4) are Grove connectors (from Seeed Studio, they have a ton of great I2C-ready sensors). (4) are regular 2.54mm 4-pin connectors.
  • Has 16 slots with relay switches. (12) connect a device to 12V and GND, (4) are wired to accept any VCC in.
  • Connects to the Air Sensor Module through USB Mini (see below)
  • Connects to the Lighting Driver Module with (6) V_SET pins for controlling channels

Lighting Node Module (10)

Lighting Driver + Node Module (2)

  • We’ve split the lighting drivers across 2 identical modules, each with 3 drivers.
  • The 6 V_SET inputs from the Mainboard flow into the first module and out to the second for easy plugging
  • This module also has a lighting node with 3 LEDs of different channels

Air Sensor Module (1)

  • The CO2 and Humidity/Temp sensors are on a single board that sits inside the growth chamber
  • We used a 3D printed housing to mount it with a minimal hole through the acrylic growth chamber
  • Plugs into the Mainboard via mini USB

Currently these are designed & being manufactured. We ordered enough for 2 PFCs, and the total was $66 (not assembled) from www.allpcb.com

Potential Problems

Airflow & Temperature Management
In our current plan, we don’t have any means of changing the temperature without introducing outside air, other than heating up the water. We’ve thought about routing air from the LED arrays into the tank to heat up the air, but this will surely bring in fresh air from the outside, changing the CO2 composition. We’re still looking for ways to solve this problem.

Software for Modular Add-ons
We will have to write the code that integrates with the v3.0 software, but allows for new sensors or the simple relay array we’ve included. We think we can get this working, but it will be a challenge.

We’ve only just begun a few weeks ago, so we’ll be diligently updating on this thread. We’re open to any questions, feedback, or help! We can also elaborate on anything if you are interested.

TL;DR

  • We’re a design agency building a PFC in London.
  • We want to help the community make PFCs more accessible.
  • Our approach is to make it as modular as possible.
  • More updates to come!
6 Likes
#2

Congratulations Method team!
I just noticed that your device uploaded its first image. Great job!
Will you be open sourcing your designs for the community to use?

#3

Wow cool! We didn’t even realize that was uploading! Yes we are planning on publishing everything next week. Our PFC will be sitting next to a couple of MIT’s in the Barbican at the AI:More than Human exhibition opening next week in London!

We’re having some last minute issues getting the UI to display images and viewing the UI in remote view with serveo – do you have a minute @rbaynes to take a look?