The Green Observatory

Table of content

Sketches and first ideas

References and related projects

Principles and concept

Electronics Design

Sensing the electrical activity of a plant

Structure: design and fabrication

Respiratory movement: programmable air

Respiratory movement: Flexible part

Code: From input to output

Assembly: Putting everything together

Introduction

A device that transforms the electrical activity of plants into a physical and meditative movement that we, humans, can observe and contemplate.

Sensing the electrical activity of a plant

sensing the activity
sensing the activity

Sensing the electrical activity of a plant

External resources

Kanban board List of materials

Created 15/06/2020

Updated 15/06/2020

Sketches and first ideas

This page is where I collect some of my thoughts on my final project idea.

Listen to what the plants have to say

Plants communicate as we do. But it's hardly noticeable for us because we don't send and receive signals at the same speed as they do. This is why I would like to build a kind of machine that can help any plant to express its needs at a speed that we can actually perceive and understand. Or should I reformulate in a less human-centric way: a machine that can help humans understand what the plants have to say.

A machine that helps us understand plants

The machine I have in mind is able to get the data from a plant [temperature, light level, air quality, humidity level, nutrient level, vibrations, ?] and transform it into something visual or audible. The machine itself could be small, like a tool that helps humans understand plants in everyday life, or large, as an art installation that invites the public to join a piece of nature, to watch and listen to what it has to say.

Breathing

Breathing (or respiration) is a very common process between plants and animals. This is the kind of movement I can use to represent how the plant feels in relation to its environment.

final-project-drawings
final-project-drawings

Data visualization

Before trying to understand what's going on between complex structrures such as plants, a good first step for me is to explore and familiarize myself with the data that I can collect about them and their environement, with the tools we have in fablabs and with the process of digital fabrication. Understand and visualize the needs and activity of plants. Here is my fab academy masterpiece, my final project which represents what I learned.

Mechanical movement

I'm thinking about a machine that can represent the movement of the breath, by using mechanics instead of air, and therefore I can have the control on what's happening, ahd what kind of movement is made.

sketch-mechanism
sketch-mechanism

I imagine a grid of rods pushing into a soft material and therefore deforming it. Each stem is linked to a data type (temperature, moisture level, etc.) and adjust its level depending on the state of the data. The soft material takes a unique shape according to plant's needs.

sketch-mechanism
sketch-mechanism

Luminous choreography according to the plant's electrical signal

The electrical signal of the plant, collected using electrodes and a signal amplifier, adjust the level of the LED's to give a visual signal on the pulses of the plants.

Global reaction to a human touch

If a human touches the plant, the installation reacts as well, inviting the people to give a special attention to it. This can be made using the touche sensor, as seen here.

Sketches

sketch
sketch
sketch
sketch
sketch
sketch

A kind of "spaceship" for plants. A machine that help them to move according to the light or a animal presence?

sketch
sketch
sketch
sketch

Using air pressure instead of mechanics. Get something more organic, less predictable.

sketch
sketch

Lots of interative elements of differents size. Create a landscape.

sketch
sketch
sketch
sketch
sketch
sketch
sketch
sketch

An installation that could hang on a tree or be dropped off somewhere. A system, a network.

sketch
sketch
sketch
sketch

Use of 5 differents sensors: light level, temperature, moisture level sensor, touche, voltage of the plant.

Each module could represent either one of these activy or a sum of several. A plant could send its data to one module or to several.

One could compose a unique installation that invites us to get closer to plants, to observe them, to take care of them.

sketch
sketch

A module could be assembled with others on a common structure to build a massive installation. Or a module could be kept apart as a stand-alone piece.

Respiration

How and why we measure photosynthesis

Principles

This project convert the data of a plant and its environment into a an installation which aims to invite us to get closer to nature and to observe how plants live and re/act.

Represent the health and activity of the plant

A portable device collects and analyzes light, air temperature and soil moisture level in a plant's environment, as well as its electrical activity.

These values will allow the device to determine how the plant is feeling, according to its needs and whether it is satisfied or not.

The device mimics the mechanisms and movements of the respiration, because it is a very common process between plants and animals, to represent the health and activity of the plant. For example, a plant that is in a good environment "breathes" slowly and deeply, and a plant that is in danger "breathes" briefly and jerky.

cycles
cycles

Interact with the living

A touché sensor, or at least a calibrated capacitive sensor, will allow the device to sense a human-plant interaction and to reveal its nature.

  • Is it a caress, a strike or a caring touch?
  • What does a plant feels when we touch it?
  • What do we feel when we touch them?

Play with a modular system

A device is composed by two items. One is the plant, capturing data, linked to a plant. The other one is the clone, showing the health and activity of the plant.

Data type

  • Light level
  • Air temperature level
  • Soil moisture level
  • Plant's electrical activity
  • Plant-human interaction

Communication

On the plant, one can choose the communication channel (between 1 and 5) which will be used to send the data to the clone.

On the clone, one can choose its communication channel and the types of data that are used to represent the health and activity of the plant.

  • plant 01 send light, temperature and interaction to clone 01
  • plant 02 send activity and moisture to clone 02.
  • plant 03 send interaction to clone 03

interface
interface

This kind of configuration allows to build any kind of networks, with several plants and tailor-made visualizations. One plant can be linked to several clones, and vice-versa.

The panel

The panel, a wooden structure designed to assemble modules together, can hosts up to 5 clones, creating a more detailed installation that facilitates the understanding of the captured values.

Public

  • Massively used in a large room and re-creating a space forest, it invites an audience to interact with plants and clones
  • A device plugged to an houseplant, acting as a companion in the everyday life

Two possibilities

Mechanical system

My first idea was to have a mechanical system that pushes rods on a flexible material in order to deform it and give it a unique shape according to the sensations of the plant.

clone
clone
clone
clone

I like this solution because I can easily imagine having total control, mastering the speed / movement ratio of the rods and therefore the shape of the clone.

clone
clone
clone
clone

But this solution requires a lot of material to be produced, and it will be too expensive compared to the idea that I have of it. I want to build something as affordable and easy to assemble as possible, to make sure it reaches a large audience.

Inflatable system

The other system I have in mind uses air pressure to inflate or deflate a flexible bag. This replaces the need for ± 12 motors with a single solenoid and an air pump. The control of the final shape is reduced but also the technical barrier, which is a good point. In addition, the movement should be more organic.

inflat
inflat
inflat
inflat

Materials

My project is made up of three elements: the plant, the clone and the panel.

The plant

The plant is a device connected to a plant. It has sensors and a microcontroller equipped with a WiFi transceiver module. Its role is to sense data from the plant and its environment, process them and send them to the clone.

  • Microcontroller (ESP32)
  • Light sensor
  • Temperature sensor
  • Soil moisture sensor
  • Electrical activity sensor (ADC, Operational amplifier)
  • Capacitive sensor
  • Power supply
  • Enclosure (wood and acrylic)

References

The clone

The clone is a device that receives data from the plant and its environment and transforms them into physical movement using an air pump to inflate and deflate a flexible material, creating a movement inspired by our breathing cycle, revealing how the plant feels.

  • Air pump
  • Solenoid
  • Pipes
  • Balloon (made out of recycled plastic)
  • Cover (made out of flexible fabric)
  • Power supply
  • Enclosure (wood and acrylic)

References

The panel

The panel is a wooden structure that can host up to five clones to create a more detailed and unique installation with custom parameters.

  • Wood panels
  • 3D printed joinery

Created 02/04/2020

Updated 02/04/2020

References and related projects

Concept

Technique

  • FYI on DIYs in PHL: Data Garden: "The sensors are “psycho galvanometers” which graph changes in galvanic response (electrical “skin” conductance/resistance) by producing a square wave of variable frequency and pulse width. Built from a 555 timer IC and a handful of electronic components, each sensor is attached to a leaf on a particular plant using an electrode of silver wire and conductive gel."
  • The Lessons To Be Learned From Forcing Plants To Play Music: The consumer version of the invention includes sensors that issue small signals through the plant, measuring variations in electrical resistance between two points within it. "The variation in the connection is largely related to how much water is between those two points, which changes a lot as the plant is moving water around while it's photosynthesizing, then we graphed that change as a wave, and then we translate that wave into pitch, so then essentially we're getting a stream of all these pitch messages coming from the plant."
  • Singing plant: Make yout plant sing with Arduino, Touche and a Gameduino.
  • Biodata Sonification kit using breadboard, for Arduino Uno 'Sheild'
  • Biodata Sonification on Instructables
  • Programmable-Air (Github project) is an open source hardware kit for controlling inflatable soft robots.
  • Open Soft Machines
  • Touche for Arduino: Advanced Touch Sensing

Created 01/05/2020

Updated 01/05/2020

Principles and concept

Why am I making another electronic device

Created 15/05/2020

Updated 15/05/2020

Electronics Design

Components

  • 4 solenoids
  • 1 air pump
  • 1 microchip ATMega16U2

ATMega16U2

  • Same microshop used in the Arduino Uno in order to exchange information via USB
  • Has enough pins for my project

Steps

  • Understand the USB part of the Arduino Uno schematics
  • Test how the mosfet + solenoid part is acting in term of voltage drop
  • Check the footprints of the different elements
    • Are they available?
    • Is it ok to solder them?
    • Aren't the pads to close to each other, or the lines too thin (or thick)?
  • Are all the components available at the lab?

Crystal

Oscar recommended using a resonator instead of the crystal because it is easier to use and solder than the crystals, but I found another piece of advice in the documentation from a former student.

From Andrew Mao's documentation:

In order to use Full-speed (12 Mbit/s) USB, the microcontroller needs to be able to generate a precize 48 MHz clock with a deviation of no more than 0.25%. Since resonators have 0.5% tolerance, This means only quartz crystals can be used, and moreover they need to evenly divide into this frequency in order for a phase-locked loop to generate this clock. For the ATMega16U2, this will require a 8MHz or 16MHz crystal - no substitutes.

But using a crystal induces having more components and thus harder to assemble. According to Oscar, at our level, in the Fab ecosystem, we don't especially need that precision. A resonator is totally fine for what I need.

Design tips

  • Check the correct clearance before tracing the routes
  • Apply a common ground at the end of the design
  • Global deletation > tracks, as soon as I'm blocked
  • Clean tracks to be sure there are no undesired portions of track

Export Kicad

Oscar's project to export SVG to PNG from Kicad using the Inkscape API. No holes for now, so I'm using Via's. I've to document that. The outer zone of the traces file must be filled in black and then inverted when imported into fabmodules.org. - link to gitlab project - Command line SVG -> PNG - What Oscar has been changed, for Python3 compatibility

Soldering

  • I have changed some components when I was doing my shopping list because not all of them were available at the lab. Here is the updated list and what I had planned to use:
  • Mini USB is much easier to solder than the micro usb, that why I used it. The only thing with that USB version is that is a bit more complicated to find the right cable.

Debugging

  • When plugged to my computer, I first tried to read my USB ports by typing $ lsusb but the new device didn't appear there. Then I checked more deeply with $ dmesg -w and saw that I had an over-current condition on the new device's port.

ISP programming

First of all, I have to program the chip to tell it what will be its behaviour. For that I am going to use a AVRISP mkII programmer.

Questions

QUESTION: How many pumps and solenoid are really needed?

  • In the Programmable Air project, they use 2 pumps and 3 valves in order to be able to manipulate entirely an air flow. I would like to reach that precision level but I'm not sure how much is needed

QUESTION: One single board or few modular ones?

  • One single board is easier to do because everything is physically connected together and thus we don't have to think about communication between the different parts. On the other hand, a modular approach allows to prototype faster, to make errors and improvements without having to fabricate everything again. I will begin to design and fabricate a single board and then iterate to different modules according to my time management.

References

Created 01/06/2020

Updated 01/06/2020

Sensing the electrical activity of a plant

Introduction

I selected two projects, related to mine, that listen to the electrical activity of plant in order to produce music (or at least sounds). They both are very useful for the understanding, design and fabrication of the sensing part of my project.

References

The first is Pulsum Plantae (Github project). This project amplifies the low voltage of a plant to make it readable, then converts it into an interactive installation that makes sounds. Leslie Garcia, Thiago Hersan, and Paloma López (the artists) have opened the sources of their project to make it accessible to people like me, allowing me to learn how things work and what could be a technical solution for this specific project.

The other comes from Data Garden and their Midi Sprout. Midi Sprout is a device that converts the electrical activity of plants into midi signals that you can send to any device such as synthetizers in order to produce musical melodies made by plants. The project isn't open-source but, fortunately for me, someone (Sam Cusumano) open-sourced the electronic part of it to make it accessible to all. It can be found here and there.

Sensors

I will prototype them both and compare the results to see which one is best for my use.

Sensor prototype n°1: 555 IC timer

The first sensor I want to prototype is build around a 555 IC timer. This component works as a square wave generator and the values it gives can be modified by changing the resistance of the circuit.

555 timer schematic
555 timer schematic

This sensor is used to graph changes in galvanic response (electrical “skin” conductance/resistance) by producing a square wave of variable frequency and pulse width. The values I get from this custom sensor can help me understand what happens inside the veins of the plant, or at least understand its electrical activity.

Arduino and components
Arduino and components

Here is the list of the components I used to assemble this sensor:

  • 555 Timer
  • Capacitor 0.0042uF
  • Capacitor 1uF
  • Capacitor 47uF
  • Resistor 100K
  • Jack Input 3.5mm
  • Electrode pads
  • Electrode lead
  • Potentiometer 10k
  • 16MHz Crystal Oscillator
  • ATmega328P (Arduino Uno, in case of this first prototype)

More information: List of materials

From input to output

Atom & Arduino screenshot
Atom & Arduino screenshot

In order to know if the data I collect from the plants is usable, I first have to do some data treatment. In this case, I have to convert an digital input into something that looks more alive than just 0 1 0 1 1 0.

The idea of using the 555 IC timer is to measure how often the length of the 1 compared to the 0 and calculate the duty cycle.

durationHigh = pulseIn(pinTimer, HIGH);
durationLow = pulseIn(pinTimer, LOW);
dutyCycle = float(durationHigh) / (durationHigh + durationLow) * 100;

It gives me a value in percentage that varies according to the conductivity of the plant.

I will elaborate more about this topic on its dedicated page:

From input to output

Alocasia Amazonica

The Alocasia Amazonica is one of my favorite plant, sitting on my desk. Its nerves are big and apparent which is supposed to make her a perfect candidate for my experiment.

And the results are way better than what I expected. The plant is super reactive to any touch and gives me beautiful data which seems quite easy to manipulate.

  • Duty cycle: from ~43% to ~47%
  • Reacts when I touch
  • Some interferences are perceptible (computer, keyboard) or reactions to stimuli?

Strelitzia reginae (Birds of Paradise)

The Strelitzia reginae and its big leaves is even more reactive than the Alocasia. Small variations in the data I collect are visible, any simple touch anywhere on the plant makes a tiny jump in the curve. I played with it during an hour without getting bored a single second.

  • Duty cycle: from ~40% to ~51%
  • highly reactive to touch
  • the curve oscillate more than with the Alocasia

Sensor prototype n°2: Op-amps

To do: Prototype with an Arduino Uno and a breadboard

Questions

QUESTION: How to program the ATmega328P?

QUESTION: What connections do I need for sending data to the pumps?

QUESTION: How to connect a power supply?

QUESTION: How to convert 12V to 3V?

QUESTION: Is 3V enough for the sensor?

Created 01/06/2020

Updated 01/06/2020

Structure: design and fabrication

structure overview
structure overview
structure detail
structure detail

Created 02/06/2020

Updated 02/06/2020

Respiratory movement: programmable air

Created 03/06/2020

Updated 03/06/2020

Respiratory movement: Flexible part

Making the mold

  • freecad design
  • laser cut test
  • laser cut test, clean the lens
  • assembling, holes too small, had to re-do them with the hand drill

lasercut acrylic transparent 5mm: 95% power, 0.45 time, 20000HZ

Preparing the casting

  • Spray release agent, mask and gloves, wait for 10 min
  • calcuate the volumes I need to cast (Pi * radius * radius * height)
  • 50% of A, 50% of B, mix super well for 5 min
  • put a timer of ~ 20 mins (according to the specs of the silicon I'm using)
  • Pour the solutions slowly, to avoid the creation of bubbles, in the mold
  • Wait for 5 hours

Other side

  • Unscrew the structure and flip the mold
  • Clean everything
  • Place the disk in the center, on the top of the first layer of silicon
  • tape the holes (with transparent tape)
  • Calculate the new volume to cast, and prepare the silicon mix
  • Cast and try to reach around 1 or 2 mm of thickness

Casting

Mistakes encountered

  • The two holes in the design were too small compared to the pipes, I had to enlarge them with an hand-drill
  • I calculate the volume without substracting the inner pieces, I ended with a waste of silicon (which is expensive)
  • I've used white tape, which is way more too visible and thus un-aesthetic
  • The first try of casting the second layer was too thin, the inflation worked but I tore the membrane

Created 04/06/2020

Updated 04/06/2020

Code: From input to output

Created 05/06/2020

Updated 05/06/2020

Assembly: Putting everything together

Created 06/06/2020

Updated 06/06/2020