

June 3, 2020
A U T O M AT E D H Y D R O P O N I C
SYSTEM BUILD

Last Updated: May 8, 2022

Hydroponic farming is a method of growing crops without soil, with the main benefits of
environmental and nutrient control, water conservation, and reduction of labor. This
technique relies on a number of technologies that the principles of automation can be applied
in order to improve yield and consistency. In this article and accompanying video, I’ll show
you how to built a hydroponic system with basic components and automate it using the open
source software Mycodo running on a Raspberry Pi single board computer.
Watch the video Automated Hydroponic System Build on YouTube

Many of the aspects that make hydroponic farming so efficient can also make it challenging to
implement. Water and nutrients must be provided in a balanced ratio while being constantly
monitored and adjusted as plants develop. Temperature regulation can also become a
problem since the circulating water system acts as a heat pump. Additionally, the microbial
communities that can benefit or harm your crops are now competing in an aqueous rather
than a terrestrial environment, sending us even further into unknown and potentially
dangerous territory. And to make matters worse, if you have a prolonged power outage, you
may need to manually water or have a generator to keep the water pump going, else you may
injure your crops.

Even with these challenges, there are several aspects that make hydroponics attractive
(Barbosa et al. 2015). An optimal amount of nutrients can be supplied directly to plant roots
in an immediately-usable form. Plants can be grown in areas with poor soil or limited space, or
grown vertically to maximize plant density. Water waste is reduced by recirculating it back
into the system. There’s also no need to worry about weeds.

Although there are a number of different hydroponic and cultivation techniques, this article
explores only one of the ways to build a hydroponic system and apply principles of
automation to monitor and regulate it. This is meant primarily to demonstrate the process of
applying automation techniques, which can be adapted to many other systems. And while the
hydroponic system I built is relatively small, it’s a modular design that can be scaled to
different growing spaces, be it a small grow tent or a large commercial greenhouse.
Table of Contents

1
Support
2
Introduction
2.1
Hydroponic System Automation Features
3
Disclosures
4
Materials & Parts List
4.1
Tools (Essential)
4.2
Tools (Non-Essential)
4.3
Base Hydroponic System
4.4
Planting
4.5
Environmental Monitoring and Regulation System
4.6
Power Control Box
4.7
Water Condition Sensing
4.8
Water Dosing
4.9
Water Flow and Level Sensing
4.10
Air Condition Sensing
4.11
Electrical Power Sensing
4.12
3D Printed Parts
5
Building the Hydroponic System
5.1
Frame
5.2
Water Channels
5.3
Water Input
5.4
Water Collection
5.5
Airflow
5.6
Lighting
6
Building the Automation Hardware
6.1
Control Panel
6.1.1
Connecting the Sensors, LCD, and Pump
6.1.2
Connecting the Float Switch
6.1.3
Connecting the RJ45 Surface Mount Jack
6.2
Liquid Solution Dispensing
6.3
Sensor Sample Reservoir
6.4
Water Flow Sensing
6.5
Electrical Current Sensing
 6.6
Power Control Box
7
Installing and Configuring the Automation Software
7.1
Install Raspberry Pi Operating System
7.2
Install Mycodo
7.3
Port Forwarding
7.4
Configuring Mycodo
7.4.1
Adding and Configuring Inputs
7.4.2
Adding and Configuring Outputs
7.4.3
Calibration
7.4.4
Adding and Configuring Functions
7.4.4.1
Grow Light Schedule
7.4.4.2
Electrical Conductivity and pH Regulation
7.4.4.3
Vapor Pressure Deficit Regulation
7.4.4.4
Configure Email Notification/Alert Settings
7.4.4.5
Email Notification for Air Temperature
7.4.4.6
Email Notification for Water Level
7.4.4.7
Exhaust Timer
7.4.4.8
DSLR Camera Timelapse
7.4.5
Adding and Configuring the LCD
7.4.6
Adding and Configuring the Camera
7.4.7
Adding and Configuring a Dashboard
7.5
Automation Epilogue
8
Plant Cultivation
8.1
Biocontrol
8.2
Planting Seeds
8.3
Maturation
8.4
Harvest
9
Future Directions
10
Log In to View My System
11
References

Support
If you find this article or video useful or you use Mycodo and want to support my future work
and software development, consider becoming a Patron at patreon.com/kylegabriel

Introduction
At the core of this automated hydroponic system is
the Raspberry Pi single board computer. This $35
computer is only the size of a deck of playing cards,
but it has all the components you would find in a full
computer, including CPU, RAM, video/audio output,
USB, WiFi, Bluetooth, Ethernet, and micro SD slot.
The most important feature of the Raspberry Pi for
our application is its general purpose input-output
(GPIO) pins, which will be used to connect inputs,
such as switches and sensors to measure the
environment, and outputs such as pumps and relays
to move liquids and control electrical devices that Raspberry Pi 4 single board computer
will manipulate the environment. (courtesy of WC)
Until recently (this month, in fact), the major
Achilles Heel of the Raspberry Pi has been an inability to boot from external storage devices,
forcing the use of a micro SD card. Micro SD cards were designed for cameras where there
are minimal read/write operations. Running a full Linux operating system along with a time-
series database recording measurements severely reduces the life expectancy of the SD card
and reduces the stability of the system due to file system corruption. In May 2020, firmware
that allows the Raspberry Pi 4 to natively boot from an external storage device entered beta.
This drastically improves its stability and suitability in critical automation applications.

The software which will be installed on the Raspberry Pi to automate the hydroponic system
is a free and open source software package I’ve developed, Mycodo, which is essentially a
universal tool for environmental measuring, monitoring, and manipulation, and has been used
for a number of interesting and diverse applications. Mycodo is a user-friendly yet powerful
tool for building automation into almost any system. Below is a brief overview of the
functionality that will be configured in Mycodo to operate as a hydroponic system controller:

Hydroponic System Automation Features

Measure water conditions: pH, electrical conductivity, temperature, level, and flow.
 Measure air conditions: temperature, humidity, vapor pressure deficit, and carbon
dioxide (CO2).
Control relays to modulate lights, air exhaust fan, and humidifier.
Control four peristaltic pumps to dispense specific volumes of solutions: acid, base,
nutrient A, and nutrient B.
Automatically adjust water to target pH range by dispensing acid/base solutions.
Automatically adjust water to target electrical conductivity range by dispensing
nutrient solutions.
Automatically adjust air to target humidity, temperature, and CO2 concentration
ranges by modulating an exhaust fan and humidifier.
Automatically regulate air vapor pressure deficit (VPD) with a humidifier and exhaust
fan, using a PID controller.
Use timers to schedule grow lights, air exhaust, and a DSLR camera shutter for high
quality time-lapse photography.
E-mail alert notifications if select measurements fall outside acceptable ranges (e.g.
temperature too high, water level too low, water flow has stopped, etc.).
Measure electrical energy usage with a transformer to automatically calculate
operating costs.
Use the Raspberry Pi Camera to monitor plants with a live video stream and conduct
time-lapse photography of plant growth.
Configure dashboards with gauges, graphs, camera feed, and other widgets to view all
relevant data on a single page.
Forward port 443 through your router to the Raspberry Pi to be able to access the
Mycodo web interface from anywhere with internet access (using a user login system
for security).

Disclosures
I reached out to Atlas Scientific to ask if they would be willing to contribute to this project,
and they were kind enough to donate 4 peristaltic pumps, an electrical conductivity sensor
kit, and a pH sensor kit. I’ve used Atlas Scientific instruments for years and have incorporated
support for several of their sensors and pumps into Mycodo, all prior to this project. I’d like to
thank them for their donations and for making quality products.

Additionally, I earn a commission from qualifying purchases. What this means is if you click a
link on my site to a product and purchase it, I may earn a small commission, without any
additional cost to you. Please consider using these links if you would like to support my work
and see more like it in the future.
Materials & Parts List
Each part of building and operating the hydroponic system will be covered in detail. As needs
vary, so too will the choice of materials one incorporates into his or her system. The materials
list below is only meant to provide a complete list of the components used in this particular
build.

Tools (Essential)
These tools are the minimum required to build this system, unless you want to adopt some
creative fabrication techniques.

Description Purchase

Soldering Iron Amazon

Solder Amazon

Saw or Dremel with Cutting Kit Amazon, Amazon

Heat Gun Amazon

Cordless Drill Amazon

Drill Bit Set (16-Piece) Amazon

Step Drill Bit (1/4 to 1 3/8 in.) Amazon

1 3/4 in. Hole Saw Bit Amazon

Screwdriver Set

Thread Seal Tape Home Depot

Tools (Non-Essential)
These tools aren’t required to build the system but are useful for verifying measurements and
tuning the system.

Description Purchase
 Description Purchase

Photosynthetically active radiation (PAR) Meter Amazon

Clamp Electrical Current Meter Amazon

Base Hydroponic System

These are the components necessary to build the base hydroponic system that can operate to
grow crops but will need to be manually operated. The materials further below will automate
various parts of the system.

Description Purchase Quantity

Grow Tent Amazon 1

Grow Light Amazon 1

Microtube Grommet AmHydro 6

Microtube AmHydro 6

2 in. x 3 in. Vinyl Downspout Home Depot 1

4 in. PVC Sewer and Drain Pipe Home Depot 1

4 in. PVC Sewer and Drain Cap Home Depot 2

3/4 in. Bulkhead Union Home Depot 1

1 1/2 in. ID Swimming Pool Hose (2 feet) Home Depot 1

1 in. PVC Pipe Home Depot 1

1 in. PVC End Cap Home Depot 2

1 in. x 1 in. x 1/2 in. PVC Tee (S x S x FPT) Home Depot 1

1/2 in. ID x 1/2 in. MIP Plastic Hose Barb Adapter Fitting Home Depot 1

10 Gallon Bin Home Depot 1

Submersible Water Pump Amazon 1

1/2 in. ID Clear Vinyl Tubing Home Depot 1

Stainless-Steel Hose Clamp (3/8 in. – 7/8 in.) Home Depot 1

 Description Purchase Quantity

Stainless-Steel Hose Clamp (3/4 in. – 1 3/4 in.) Home Depot 1

Nylon Filter Bags Amazon 1 Pack

Plant Growth Nutrient Solutions (Two Part, A and B) Amazon 1

pH Adjustment Solutions (Up/Base and Down/Acid) Amazon 1

Mycostop Biocontrol (Streptomyces griseoviridis bacteria) Amazon 2 Grams

Planting
These are the materials for germinating and growing plants.

Description Purchase Quantity

Seeds Johnny’s Choose

Rockwool Grow Cubes or Horticubes Amazon, Amazon 1 Pack

Seed Germination Tray and Dome Amazon 1

Waterproof Seedling Heat Mat Amazon 1

Environmental Monitoring and Regulation

System
Description Purchase Quantity

Raspberry Pi v4 Amazon 1

Micro SD Card (32 GB) Amazon 1

Raspberry Pi v2 Camera Module Amazon 1

Raspberry Pi Camera Ribbon Cable Amazon 1

HDPE Plastic Mounting Panel Amazon 1

6-Outlet Power Strip 1

 Description Purchase Quantity

20 character, 4 line I2C LCD Amazon, 1

#1 x 7/16 in. Stainless Steel Screws Amazon 1 Pack

Round Nylon Spacers (H 6 mm, OD 7 mm, ID 4 mm) Amazon, 1 Pack

Mini Breadboards Amazon, 1 Pack

Jumper Wire Kit Amazon, 1 Pack

Power Control Box

These materials are for building a power control box that has 4 individually-controllable
outlets, allowing any 120 VAC device to be controlled by the automation system.

Description Purchase Quantity

6 in. x 6 in. Junction Box Home Depot 1

RJ45 Panel Mount Connector Amazon 1

RJ45 Surface Mount Jack Amazon 1

RJ45 Crimper, Connectors, and Cable Test Tools Amazon 1

RJ45 CAT-6 Ethernet Cable (5-foot) 1

IEC 320 Panel Mount Power Socket Amazon 1 Pack

IEC 320 Power Cord Amazon 1

Wall Outlets Home Depot 2

2-Gang Wall Plate Home Depot 1

2-Channel Relay Module (Mechanical, 10-Amp) Amazon 1

2-Channel Relay Module (Solid State, 2-Amp) Amazon 1

14 Gauge Copper Wire Home Depot 1 Spool

Water Condition Sensing

These materials enable various water chemistry to be measured and provides feedback for
the water dosing system.

Description Purchase Quantity

Atlas Scientific PT-1000 Temperature Sensor Kit Amazon 1

Atlas Scientific EZO Carrier Board (for Temperature Sensor) Amazon 1

Atlas Scientific pH Sensor Kit Amazon 1

Atlas Scientific Electrical Conductivity Sensor Kit Amazon 1

Aquarium Specimen Container Amazon 1

Water Dosing
These materials enable the water chemistry to be adjusted by dispensing different liquid
solutions.

Description Purchase Quantity

Atlas Scientific Peristaltic Pump Amazon 4

Silicone Tubing (ID 3 mm x OD 5 mm) Amazon 1

1/8 in. Hose Barb Union Home Depot 1 Pack

12-Volt 5-Amp DC Power Supply and Female Barrel Connector Amazon 1

Water Flow and Level Sensing

These materials enable the measuring of water conditions that allow for the system to be
monitored for proper operation.

Description Purchase Quantity

Atlas Scientific Flow Meter Atlas Scientific 1

Atlas Scientific Flow Meter Totalizer Amazon 1

Atlas Scientific EZO Carrier Board Amazon 1

1/4 in. O.D. x 3/8 in. MIP NPTF Push-to-Connect Adapter Fitting Home Depot 2
 Description Purchase Quantity

3/8 in. FIP Brass Pipe Coupling Fitting Home Depot 2

Water Level Float Switch with Pipe Mount Amazon, 1

Assorted Resistor Kit Amazon, 1

Air Condition Sensing

These materials enable various air conditions to be measured.

Description Purchase Quantity

HTU21DF Temperature/Humidity Sensor (or AM2315, SHT31, etc. Amazon, Amazon, 1

See Supported Sensors) Amazon

Dupont Crimping Tool and Connectors Amazon 1

MH-Z19B CO2 Sensor Amazon, Tindie 1

UART to USB Converter (USB Interface) Amazon 1

Electrical Power Sensing

These materials enable the amount of power consumed by the system to be measured.

Description Purchase Quantity

Analog-to-Digital Converter Amazon 1

AC Current Sensor with Split Transformer or Greystone CS-650-R1 Solid Amazon 1

Core Current Sensor (only one, not both)

Extension Cord (1 ft.) Amazon 1

3D Printed Parts
These are various 3D models of simple parts I created and printed with my 3D printer that
helped with the build.
 Model 1: Sensor Probe Reservoir Lid

Model 2: Bottle Support Mount

3D Printer: Creality Ender 5 Pro

3D Modeling Software: Autodesk Fusion 360

3D Slicing Software: Ultimaker Cura

3D Printer Control Software: OctoPi (OctoPrint for the Raspberry Pi)

Building the Hydroponic System

This hydroponic system being built relies on a method known as the nutrient film technique
(NFT). These systems are popular for their ease of use and the high yields that can be
achieved. Plants that benefit the most from NFT are leafy greens, such as lettuces and herbs.

The design of an NFT hydroponic system is simple. Water is pumped up from a reservoir to
fluid channels, where the channels are placed at a slight angle to allow the water to flow from
the high end to the low end by gravity, where the water is then returned to the reservoir.
Plants are placed in holes cut in the top of the channels and are supported by a growth
substrate. The substrate provides a place to absorb water and/or support the plant. As the
plants grow, their roots extend into the fluid flowing across the bottom of the channels. This
particular design was inspired by do-it-yourself projects as well as systems built by the
commercial hydroponics company AmHydro, from which the grommets and microtubes
being used in this build were purchased.

Frame
A frame is needed to support the water channels as well as the water inlet and collection
pipes. One end should be higher than the other to allow water to flow through the channels
by gravity. I had 1-inch punched square steel tubing left over from a previous project, so this is
what I used to construct my frame (drawing below). However, there are many ways to support
channels, from raising one end of a table to building a custom frame. Steel is a relatively
expensive material, so I would suggest using angled aluminum bar or tubing if you’re going to
design your own frame. Another inexpensive and modular option is to use 3/4 inch electric
metallic tube, a metal tube cutter, and Maker Pipe connectors.
Detailed drawing if the 1-inch square steel tube frame.

Between 2° and 3° is a sufficient angle for water to flow through the channels. If the angle is
too small, water may not flow fast enough through the channels and cause overflow and
spillage on the inlet edge. Cut the square tubing into two 16.25-inch lengths and two 15-inch
lengths, to serve as two long and two short leg sets. Cut two 23-inch lengths to attach to the
top of each leg set (upon which the channels will rest), and bolt flat steel bars angled at 90° to
secure them to the legs. Secure all 90° flat bars using 4 bolts to improve rigidity, so no cross-
bracing is necessary. Cut a 24-inch length and bolt each end to each of the rear leg sets, to
serve as a spine. This spine will also later serve as a support for the pumps to mount to.
The frame with 90° steel bars securing the rear-left leg to the spine.

Water Channels
The water channels are made by cutting vinyl gutter downspouts into five 2.5-foot segments.
I used 2 in. x 3 in. downspouts, but 3 in. x 4 in. downspouts were also available, which could be
a good option if you have more space or are building a system with longer channels or for
larger plants. Using a hole saw, cut 1.75-inch diameter holes on the widest side of the
channels, at a spacing of approximately 6 inches, on center. I alternated the cutting so when
the channels were placed side-by-side, the holes alternated and formed a checkerboard
pattern, which maximizes space between plants to grow. Drill a 1/4-inch hole in the top of
each channel, on the inlet side, about 3 inches from the end, which will be used to supply
water to each channel. The bottom edge of the inlet side of each channel can also be heated
and bent upward to reduce the chance of water spilling out in the event water backs up in the
channel (e.g. due to excessive root growth or if the channel angle isn’t steep enough). Place
the channels on the frame, with the inlet sides with the 1/4 inch holes resting on the higher
end.

Different plants may require different channel spacing. Additionally, channels can be made
with holes closely-spaced (2 inches on center) to be used as nursery channels to grow many
seedlings prior to transplanting to finishing channels with wider spacing for mature plants.

Water Input
To supply water to each channel, 1-inch PVC is routed along the length of the inlet side of the
frame and connected to a water pump. Cut two 11-inch PVC lengths, then cement a PVC tee
in the center and PVC end caps to the ends. Screw a 1/2-inch hose barb adapter into the tee
after applying thread seal tape. Drill five 3/8-inch holes into the inlet pipe, one in front of each
channel. Remove any burs and install a microtube grommet into each hole. Secure the inlet
pipe to the frame edge with zip-ties or other fasteners, just under the channels. Connect one
end of a 1/2-inch inner-diameter hose to the inlet pipe’s hose barb with a hose clamp and the
other end to the water pump. Last, connect a microtube from each grommet to each channel’s
1/4-inch inlet hole.

Water Collection
To collect the water exiting the channels, we’ll construct an open pipe on the outlet side to
route the water back to the reservoir. Cut a 22-inch length of 4-inch PVC pipe (S&D PVC that
I am using for this project is thinner, cheaper, and easier to work with than schedule 40 PVC,
but it is not rated as food-safe. For safety, always use food-grade, schedule-40 PVC). To
increase the area to open the 4-inch pipe, cut two PVC end caps to remove excess plastic so
the caps are shortened to a width of about 3/4 inch, then cement them to the ends of the
pipe. Mark a straight line down the length of the pipe, from end cap to end cap, then cut the
pipe along the mark. Cut about 4 inches at a 90° angle at both ends, creating a flap that can
open down the length of the pipe. Using a heat gun, slowly and evenly heat the pipe along the
flap hinge while applying slight pressure to bend the flap open. Once the flap opens far
enough to be able to fit the channels in, stop heating the pipe and allow it to cool and harden
in the open position.

Fit the channels into the pipe opening along the length of the pipe and position the pipe at the
lower end of the frame to determine an optimum mounting position. In the center of the pipe,
mark the lowest point where water will pool. This is where a 1.25-inch hole should be drilled
to attach the bulkhead union that will drain the water back into the reservoir. Once the hole
has been drilled, use a heat gun to gently heat around this hole so when the bulkhead is
tightened, the curved PVC flattens so the rubber gasket of the bulkhead forms a water-tight
seal. Secure the 4-inch pipe to the edge of the stand with large zip-ties or other fasteners and
insert all the channels. Attach swimming pool hose to the bulkhead with a hose clamp and
secure the other end in the water reservoir. Attach a nylon filter bag to the end of the hose to
catch any large particles from entering the reservoir that may potentially clog the pump (e.g.
rockwool, plant material, etc.).

Fill the water reservoir with a few gallons of water until the pump is fully submerged. Turn the
pump on and adjust the flow knob on the pump so all microtubes are outputting enough
water to make a thin film of water flow down each channel. Inspect all parts of the system for
leaks and monitor everything for several hours or days to ensure it continuously runs as
expected. A cover should be placed over the water reservoir to reduce evaporative water loss
and algae growth.

Airflow
Airflow is important for modulating temperature, humidity, carbon dioxide, oxygen, and
vapor pressure deficit, as well as reducing the risk of mold growth. An airflow that gently
bends the plants also promotes thickening, resulting in a more crisp taste. For both
greenhouses and grow tents, adequate air circulation and exhaust is crucial to plant
development. Since this system is operating in a 2 ft. x 4 ft. grow tent, small 80 mm to 150 mm
PC fans are sufficient for circulating and exhausting air. A circulation fan was connected to an
always-on outlet and placed near the humidifier to distribute humid air being produced. The
larger fan was positioned in one of the wall ducts to blow outward and was connected to the
power control box so the control software can modulate exhausting.
To retain humidity when not exhausting, shutters were added to both the intake and exhaust
ducts of the grow tent. Two 6-inch lengths of 4-inch PVC were cut and inserted into the
ducting sleeves of the tent to provide a large opening for air to pass when the exhaust fan
turns on. These PVC pieces had a flat piece of plastic placed over the opening on the inside
the tent and a piece of tape used to secure the top of the plastic to the top of the pipe.
Another shutter was added to the exhaust pipe on the outside of the tent. These act as simple
one-way air valves that will fall flat over the openings when air isn’t being pulled into the tent
and will open when the exhaust fan turns on. To improve the seals of the shutter against the
PVC pipe, angle the pipes slightly upward so gravity forces the shutters to press the shutters
against the pipes.

Lighting
I won’t spend much time discussing specific types of lighting, as this is out of the scope of this
article and there are already numerous resources devoted to this. If growing outdoors with
good sun exposure, you usually won’t need any additional lighting, unless you desire a nursery
for seedlings or to extend growing hours beyond sunset. However, if growing indoors, you’ll
need at least one grow light. I’m using a Sun Blaze (T5HO 24) 96 Watt fluorescent fixture for
germinating seeds and a Galaxyhydro (HYG05) 300 Watt LED fixture for the mature plants in
the grow tent. I’ve also used both lights in combination in the grow tent.

The grow light puts off a lot of light and can result in poor photo quality. Since I want to
capture photos with both a DSLR and Raspberry Pi camera, I needed a less bright and more
neutral light source for illuminating the plants during photo acquisition. I also needed a way
to turn off the grow light(s) while this secondary light was active and the photos were
acquired. Although this will be covered in detail in the automation configuration, briefly, the
following is how this was achieved. First, two simple outputs were created, one that
controlled the switching of a light for taking photos, named Photo Light, and another that
controlled the switching of the grow light, named Grow Light. Next, a Python Command
Output was created, named Grow Light On Only, which executes user-defined Python
commands when turned On and Off, with the On action executing code to turn Photo Light off
and Grow Light on, and the Off action executing code that turns Grow Light off. Next, a Python
Command Output was created, named Photo Light On Only, with the On action executing code
that turns Photo Light on and Grow Light off, and the Off action executing code that turns
Photo Light off and Grow Light On Only off.

To understand how this works, I’ll describe how the light timer works and explain the order of
operations when photos are acquired. There are two timers that are used in order to keep
Grow Light on only during the chosen grow period (e.g. 6 AM to 8 PM). The first is a duration
timer, that at a regular period (e.g. every 15 seconds) between 6 AM and 8 PM, will set Grow
Light On Only on, if it is not already on. This will cause Photo Light to turn off and Grow Light to
turn on. A second timer (a single time point timer) will set Grow Light On Only off at 8:01 PM,
which will turn Grow Light off. This covers keeping the Grow Light on during the grow period.

Now, the photo captures can be configured. My DSLR camera has a remote shutter cable that
I wired to two outputs to activate the focus and to activate the shutter (auto focus is turned
off, but this is still required by the remote shutter device to allow the shutter to actuate). I
created a function that will switch the focus output, wait 1 second, then activate the shutter
output, effectively causing the camera to take a photo. This function was set to repeat every 2
hours, which for a 24-frame per second video would equate to 1 second of video for every 2
days of photos. To make only the Photo Light active during the photo, I simply added the code
to turn Photo Light On Only on before the code to capture the photo, and code to turn Photo
Light On Only off after the code to capture the photo. When Grow Light On Only is turned on
initially by the duration timer at 6 AM, only Grow Light is on. When Photo Light On Only is
turned on by the time-lapse function, Photo Light is turned on and Grow Light is turned off. This
provides a neutral white light for the photo. After the photo has been captured, turning Photo
Light On Only off turns both Photo Light off and Grow Light On Only off (which means now no
lights are on). Since the grow light timer checks every 15 seconds whether Grow Light On Only
is on, and only if it’s off will it turn it on, Grow Light On Only will now be turned on, causing
Grow Light to turn on.

This allows a single output (Photo Light On Only) to control the state of two lights, so that only
one of them is permitted to be on at any given time. The Raspberry Pi camera was set up the
same way to use this light configuration, by actuating the Photo Light On Only output on
before photo acquisition and off after acquisition.

Building the Automation Hardware

We’ve just finished building a basic hydroponic system that can be manually operated and
monitored. Next, we’ll build the capacity for automated growing by expanding on this base
system. The automation hardware is essentially everything added to the base hydroponic
system that will allow autonomous monitoring and control. This includes the sensors for air,
water, and power measurements, peristaltic pumps, and power control box that will modulate
fans, lights, humidifier, or other devices.

Control Panel
A central location is needed for all the hardware components to interface together to the
Raspberry Pi, which will orchestrate the automation processes. Weatherproof enclosures are
preferred, but for this build, I chose to use 0.375-inch thick high-density polyethylene (HDPE)
plastic panel to mount everything. This is a very rigid material that is easily cut with hand or
power tools. Cut the panel to 12 in. x 10 in. and lay all the components on the board to
experiment with different layout configurations until you find one that works well. Here are a
few things to consider when developing a layout:

Wire lengths should be minimized to reduce the likelihood of introducing excessive

noise or voltage-drop.
Cables should be secured to the panel to prevent wires from being pulled out or
excessive strain on connections (unless the connector is designed for that purpose,
such as BNC, SMA, etc.).
Avoid complex wire routing that makes following or understanding circuits difficult.

Start by connecting components with jumper wire. Below is a diagram of the Raspberry Pi
and all pin designations on the 40-pin header, including ground pins, 3.3 and 5 volt pins,
communication protocol pins, and the general purpose input output (GPIO) pins.
Diagram of a Raspberry Pi 3 B+ pin mapping (image courtesy of Raspberry Pi Spy).

Solderless breadboards
are used in this build to
create many of the
connections. A breadboard
essentially allows us to
rapidly prototype by being
able to easily make and
break connections. The
Mini solderless breadboard with a diagram of how the terminal
contact points of a
strips connect the contact points.
breadboard are connected
by terminal strips to allow
us to connect to a row of pins for easier access to connections if we desire multiple wires to
be connected together. However, in a production setting you should use a board with screw
terminals or a PCB to solder to for more secure connections.

The sections below describe how to connect the components, which should aid in developing
the layout of the board. Once a good layout has been determined, place and secure the
components to the board. Mark the mounting holes for each component on the board, then
use an appropriate sized drill bit and screw to mount the component. For most of the
components I used, a 1/16-inch drill bit with 7/16-inch #1 screws were sufficient. Spacers
may be needed to mount some of the boards in a raised position, in order to prevent damage
to components on the underside of the circuit boards if they come into contact with the
mounting surface. For components without mounting holes, such as the CO2 sensor, double-
sided tape or zip-ties can be used to secure the USB cable and the sensor to the board. Last,
drill two larger holes near each corner so the board can be secured to two lengths of rope
hung from the grow tent frame, which will allow easy adjustment in any direction on the wall.

Connecting the Sensors, LCD, and Pump

The sensors, LCD, and pumps will all connect to the same four pins of the Raspberry Pi to
enable communication. These are the ground pin, 5-volt pin, serial data (SDA) pin, and serial
clock (SCL) pin. These SDA and SCL pins represent the Inter-Integrated Circuit, or I2C, bus,
which was designed for short-distance communication with sensors and devices. Connect the
ground, 5-volt, SDA, and SCL pins from the Raspberry Pi to a breadboard so each connection
is broken out into its own terminal strip, allowing multiple devices to be connected to these
pins on the breadboard. For the number of devices we’ll be connecting, you will need to use
male-to-male jumper wires to further break out these 4 pins to as many terminal strips as
needed to connect all devices (I ended up needing to use 2 mini breadboards to connect all
devices). Then, connect the sensors, LCD, and peristaltic pump communication lines to these
connections of the breadboard(s).

The CO2 sensor components listed on the Material’s List need RX <-> TX, TX <-> RX, power,
and ground to be connected from the sensor to the UART-to-USB board. The board can then
be connected to the USB port of the Raspberry Pi and can be accessed at /dev/ttyUSB0 (use
this device later in the software configuration).

The figure below is not an exact replica of the control board I built, but merely serves as an
example of how to break out the I2C lines to connect all your devices. For a more accurate
schematic, click here.
Circuit connecting the sensors, LCD, pumps, and float switch to the Raspberry Pi. Not
pictured: analog-to-digital converter and current sensor.

Note 1: Make sure all your I2C components are 3.3-volt compatible. Never allow 5 volts to be
connected to any Raspberry Pi GPIO pin or the device or Raspberry Pi can be damaged. All
I2C devices mentioned here can operate on 3.3-volts, but if you need to use a 5-volt device on
the I2C bus, it’s best practice to use a logic converter to drop the voltage of SDA and SCL
coming off the device down to 3.3 volts before connecting to the Raspberry Pi.

Note 2: In the schematics, you may notice I have all my I2C devices connected together and
powered by 5 volts. There is only one circumstance when it is safe to do this, and that is if
none of your devices pull SDA or SCL high to 5 volts with a resistor. Since none of the circuits
of my I2C devices do this, it is safe to power them with 5 volts, since SDA and SCL are only
pulled low for communication (1, 2). This is done because, although most of my I2C devices
are both 3.5- and 5-volt compatible, my LCD only operates on 5 volts. Rather than connect a
level shifter for just the LCD, I chose to just use 5 volts for everything. If using this method, be
sure you do not inadvertently allow 5 volts to pass to SDA, SCL, or any other GPIO pins of the
Raspberry Pi.

Note 3: Atlas Scientific sensors and pumps are set by default to the UART communication
mode, not the I2C communication mode. You will need to refer to the datasheet provided by
Atlas Scientific for each device to manually switch each device to I2C mode before being able
to communicate with them on the I2C bus. This is a simple procedure that involves making a
single connection on the sensor board and then powering it for a few seconds. For example,
see page 41 of the Temperature sensor EZO circuit board datasheet for how to manually
switch from UART to I2C mode.

Note 4: If mounting the air temperature and humidity sensors to the control panel, make sure
to position the panel where the sensors are at the same height or slightly higher than your
plants. If there isn’t vigorous mixing of air by fans, this is where a temperature and humidity
gradient may form (with hotter air above this level). To yield the most accurate temperature
and humidity regulation at location of your plants, the sensors should be positioned among or
right above your growing plants.

Completed control panel with wires secured with cable fasteners.

Connecting the Float Switch

The float sensor is a magnetic switch encased in a rigid body, with two wires. A buoyant ring
containing a magnet slides up and down this body, causing the switch to open or close the
circuit depending on the water level. For the Raspberry Pi to use the float sensor to detect
when the water level is high or low, a simple circuit needs to be created that yields either a
high voltage (3.3 volts) or low voltage (0 volts) when the switch if open or closed. A GPIO pin
on the Raspberry Pi can then be set as an input and measure this signal.

Briefly, to construct this circuit, connect a GPIO pin, a ground pin, and a 3.3-volt pin from the
Raspberry Pi to their own terminal strips on a breadboard. Pull the GPIO connection on the
breadboard low by connecting the GPIO terminal strip to the ground terminal strip using a 10
kΩ resistor. Last, connect one end of the float switch to the GPIO terminal strip and the other
end to the 3.3-volt strip. Below is a minimal schematic of the circuit.

Float sensor wiring schematic.

When the water level is low and the buoyant ring isn’t in proximity to the magnetic switch,
the circuit is open (disconnected) and the GPIO pin is connected to ground and will measure 0
volts (low). When the water level is high and the magnet actuates the switch, the circuit is
closed (connected) and the GPIO will be connected to the 3.3-volt pin and will measure 3.3
volts (high). When the circuit is closed, ground and 3.3-volts is connected by the 10 kΩ
resistor that limits any current to a safe amount. When we monitor the GPIO pin, we can
determine the water level based on whether the signal being measured is low or high. Note
that either a pull-up or pull-down resistor can be used to change the voltage of the circuit
when the water level is high or low. By changing this, either water level state can yield either a
high or low signal. It does not matter which is used, as the software can be configured either
way.

Connecting the RJ45 Surface Mount Jack

The power control box will allow the Raspberry Pi to independently control the power of four
AC outlets with four relays. The control box will connect to the control panel by an RJ45
(Ethernet) surface mount jack, which will then connect to ground, 5-volt, and four GPIO pins
on the Raspberry Pi. These connections will power and trigger four relay modules in the
power control box, with each GPIO pin switching each relay. Make sure each connection of
the RJ45 jack corresponds to the proper connection at the other end in the power control
box. You can use the continuity setting of a multimeter to test each connection at both ends
to ensure they’re connected. The ground and 5-volt pins of the Raspberry Pi should connect
to GND and VCC of the relay modules, respectively, and each GPIO to each channel of each
relay module. Once the power control box is constructed, a short Ethernet cable can connect
the RJ45 jack on the control panel to the RJ45 connector on the power control box.

Liquid Solution Dispensing

Four peristaltic pumps will be used to adjust the water chemistry by dispensing small
amounts of acid (lowers pH), base (raises pH), and nutrient A and B solutions (raises electrical
conductivity). The pumps need to be located close to the water reservoir to reduce hose
length and have enough room to accommodate four solution tanks below the pumps. I found
the back horizontal bar (spine) of the frame to be an ideal mounting point.

Cut another piece of HDPE panel to 12 in. x 6 in. that the pumps and solution reservoirs will
be mounted to. I modeled and 3D-printed mounts to hang a 12-ounce plastic bottle under
each pump, but your bottles can rest on the ground or by some other mounting option.
Position the pumps and mounts, mark their mounting holes on the panel, then drill 1 1/4-inch
holes for the pumps to extend through the panel and smaller holes for securing the pumps to
the panel. To connect the 12-volt DC power supply to the pumps, solder 4 sets of negative
and positive wires together, respectively, and connect them to the appropriate negative and
positive screw terminals of the female barrel connector that the power supply male barrel
connector will plug into. Then, connect each set of 12-volt wires to each pump’s power screw
terminals. Last, the tubing needs to be installed to pump fluid from the bottles to the
reservoir. Drill a hole into each bottle lid, then connect 5 mm tubing to each pump input and
feed it to the bottom of each bottle. Last, cut four more lengths of tubing, connect them to the
output lines of the pumps, then a zip-tie them all together over the reservoir.
Sensor Sample Reservoir
A sample reservoir was constructed to divert a small volume of water from the main reservoir
for the water sensors to measure. This is done in order to sub-sample the main water
reservoir, which will buffer our sensors from the immediate effects of large pH and electrical
conductivity swings that can occur in the main reservoir when the pumps are dosing
solutions.

For the sample reservoir construction, I modeled and 3D printed a lid that fits on the top of
the sample reservoir to hold the temperature, pH, and electrical conductivity probes.
However, drilling holes in a flat piece of plastic or zip-tying the probes to the side of the
sample reservoir would also work. Rubber gaskets were slid onto the EC and pH probes to
prevent them from sliding all the way into the holes and allowing their height in the fluid to be
adjusted. These probes will need to always be submerged in water, so inflow and outflow
holes need to be added to the sample reservoir. Ensure the outflow hole is sufficiently large
so water can exit the reservoir faster than it enters. Drill an outflow hole or slot on the side of
the sample reservoir 2 to 4 inches above the bottom of the highest sensor probe. Drill a 1/4-
inch inflow hole near the top of the sample reservoir, on the side opposite side of the outflow
hole. Since the plastic material is fragile, I used a Dremel to cut the inflow and outflow holes
rather than using a drill bit. Secure the sample reservoir near the water inlet PVC pipe
connected to the water pump. Make sure the water pump is turned off, then drill a 3/8-inch
hole in the inlet PVC pipe near the sample reservoir. Place a microtube grommet in this hole,
then attach a microtube from the grommet to the sample reservoir. Turn the water pump
back on and ensure there are no leaks and the water fills the sample reservoir before spilling
back into the main reservoir.
Water Flow Sensing
A water flow meter can be easily installed to the water line feeding the sample reservoir, and
is useful for detecting if the pump is operating properly (and notifying you when it is not).
After applying thread seal tape to the flow meter, screw push-to-connect fittings to both ends
of the flow meter with couplings. Cut out a section of the microtube feeding the sample
reservoir that’s the length of the flow meter, then press the microtube ends into the push-to-
connect fittings. Last, connect the three wires of the flow meter to the Atlas Scientific Flow
Meter Totalizer on the control panel.

Electrical Current Sensing

In order to measure the amount of electrical current being consumed by the system, we’ll use
a transformer to measure one wire of the main alternating current (AC) power cord that
powers the system, upstream from any devices that are plugged in. The transformer will
convert the AC current proportionally to a DC voltage. This DC voltage will then be
converted to a digital signal with an ADS1115 analog-to-digital converter connected to the
Raspberry Pi. Using this voltage, the Mycodo software will then be able to calculate the
amount of current that’s consumed over time, in Amps. This can be further used by Mycodo
to automatically calculate the monetary cost to operate over periods of time (day, week,
month, year, or custom dates) and based on your local cost per kWh.

To do this, we’ll need to isolate a single wire in our power cord. I sacrificed a 1-foot extension
cord rather than cutting into my power strip cord. Cut away a few inches of the external
insulation of the extension cord, being careful not to cut into the insulation of the individual
wires. This exposes single wires to clamp the transformer around. Depending on which
current sensor used from the Material’s List, there are two ways to connect the sensor.

The spit-core transformer is easier to use because it can clamp around a wire, however it is
less accurate than a solid-core transformer. If using the split-core transformer, open it and
close it around the live wire. From there, connect the output connector from the transformer
to the current sensing board. Last, connect the negative and positive DC voltage outputs
from this board to the analog-to-digital converter’s Ground and Analog 0 Input (A0)
connections, respectively.

A solid core transformer is more accurate than a split-core transformer, but it requires
feeding a wire through the core or cutting and feeding a wire through the core before
soldering the wire back together. First, ensure the cable is not plugged in. If using the solid
core transformer, cut the live wire, strip their ends, and feed one end through the solid-core
transformer. Solder the wire ends back together with the transformer still surrounding the
wire and use heat-shrink tubing to cover the solder joint. last, connect the negative and
positive DC voltage outputs from the transformer to the analog-to-digital converter’s
Ground and Analog 0 Input (A0) connections, respectively.
Power Control Box
The power control box will allow the 3.3-volt DC signals from the Raspberry Pi GPIO pins to
safely switch AC devices, such as grow lights, exhaust fans, humidifiers, or other devices.

To understand how this works, here’s a brief

introduction to how the two main types of
relays operate. Mechanical relays use a small
input voltage to energize electromagnets that
cause the physical movement of metal contacts
in the output component of the relay that turns
on and off much larger voltages. Solid state
relays (SSRs) use a small input voltage to
produce a light that shines across an air gap to
energize a sensor that switches the output
connections. Mechanical relays make an
audible “click” sound when tuning on and off,
and compared to SSRs, are less expensive, can
typically handle higher electrical currents, but
have a lower life expectancy, with a mean time
before failure (MTBF) in the several hundred
hours range. Solid state relays are silent when
turning on or off, can handle lower currents,
but have a high life expectancy, with a MTBF in
2-channel 10-Amp mechanical relay
the several million hours range (yes, million,
module (top) and 2-channel 2-Amp solid
with some as high as 40 million or more, which
state relay module (bottom)
equates to 4566 years, under ideal conditions).
Although this is not always true, as relays of
course vary in their specifications, it does illustrate the general differences.

I typically use mechanical relays for operations that require high current loads and slow
switching, such as grow lights that only turn on or off a few times a day. I typically use SSRs
for applications that may need to switch much faster, such as heating and humidity control
(several switches per minute) or pulse-width modulation (PWM) signal generation (several
switches per second). Although mechanical relays can often handle many applications (please
don’t try PWM), increasing their rate of switching can increase their chance of failure (and be
annoying as heck to hear their constant clicking). When possible, I try to only use SSRs
because of their superior reliability.
For these reasons, I built my power control box with two mechanical relays to handle slow-
switching, high-current devices (grow lights), and 2 SSRs to handle fast-switching, low-
current devices (exhaust fan and humidifier). Always adhere to the current limitations of your
relays and don’t plug in any devices that could exceed 10 Amps for the mechanical relays or
any that could exceed 2 Amps for the SSRs (or whatever the specific current ratings are for
your particular relays). Since alternating current is dangerous, I’m protecting the circuitry and
relays in a 6 x 6 inch junction box.

Start by placing two sets of outlets and their cover on the junction box lid and mark where the
main opening for the outlets will be. Remember to leave space for the mounting holes to
secure the outlets to the lid and to use the outlet cover to ensure you’re adequately spacing
the outlets. Cut the main opening in the junction box lid, drill the mounting holes, and dry-fit
the outlets. On the junction box, mark the holes that needs to be cut for the panel mount
power socket and the panel mount RJ45 connector near the bottom of the box. Cut each hole
and dry fit each part.

Next we’ll wire everything. Inside the junction box, the panel mount power socket will have 3
blade terminals, for neutral, ground, and live (diagram below). Use 14 or 16 gauge wire for all
connections involving AC power, and either crimp female blade connectors or solder the
wires directly to the power socket connections. I recommend using heat shrink tubing to
cover all exposed AC connections. Follow the wiring instructions and diagram, below, for
wiring the remaining connections.
 Connect G (ground) from the power socket to the green ground connector of each
outlet.
Connect N (neutral) from the power socket to the neutral connector of each outlet.
Connect L (live) from the power socket to one of the two output connectors for every
relay (i.e. every relay output will control a live connection to an outlet).
Each remaining output connector from each relay will connect to the live connector of
each outlet.
One Ethernet wire to the ground (-) input of each relay module.
One Ethernet wire to the positive (+) input of each relay module.
One Ethernet wire to the channel 1 signal input of the mechanical relay.
One Ethernet wire to the channel 2 signal input of the mechanical relay.
One Ethernet wire to the channel 1 signal input of the solid state relay.
One Ethernet wire to the channel 2 signal input of the solid state relay.

Wiring diagram of four relays controlling mains electricity to four outlets.

Note: many outlets have a metal tab that connects both outlets to the same circuit (one tab
for neutral, one tab for live), which can be broken off on each side. When unbroken, this
allows one set of wires to power both outlets. Since we want to power each outlet
independently, the live tab needs to be broken off with pliers. The neutral tab can remain
intact, allowing you to connect only one neutral that will be distributed to both outlets. The
diagram, above, has been drawn with the assumption there are no tabs present.
Warning: If you leave the live side tab intact, both outlets will be powered when either live
wire is energized by either relay. The live tab must be broken to independently control each
outlet. If you don’t break the tab, you will be potentially powering two devices instead of
one when the relay actuates, potentially exceeding your relay’s maximum current.

Each relay will have output screw terminals that will be used to switch the AC power. Wire
the power socket’s live connection to one of these output screw terminals for each relay. if
there are 2 connections, it doesn’t matter which, so long as it’s connected to only one of the
two relay output terminals. If there are 3 connections, connect to the normally-open (NO)
connections so the power is off when the relay is not active. Next, connect the second output
screw terminal of each relay to the live connector of each outlet. This will allow the live
connection of each outlet to be switched by each relay. You should mark which outlets are
connected to mechanical relays and which are connected to SSRs.

Now connect the signal wires to the relays from the RJ45 connector inside the junction box.
There are two ways to do this: A) cut a small length of raw Ethernet wire and crimp a
connector on one end, or B) cut a consumer Ethernet cable leaving a short length on the end.
This will give you an Ethernet cable with a connector on one end that will plug into the RJ45
connector inside the box. The other end of the Ethernet cable should have 6 of the 8 wires
stripped for connecting to the relay modules. Be sure to note the color of the wires and what
they’re connected to. Since there are two relay modules, the wires you designate for ground
and positive will have to connect to the ground and positive of both relay modules. Much like
the outlets, you can run the ground wire to the mechanical relay’s ground screw terminal
(typically labeled “GND” or “DC-“), then run a short wire from that same screw terminal to the
other relay’s ground connector. Since the SSR has only male header pins, you can simply cut
and strip a jumper wire with a female header to make this connection. Do the same for the
positive connection (typically labeled “VCC” or “DC+”). For the four relay input signal
connections (typically labeled “CH#” or “IN#”, with # denoting the channel number), you will
need to use four wires from the Ethernet cable. Connecting to the screw terminals is
straightforward, but to connect to the SSR input signal male header pins, you will have to
either crimp or solder jumper wires with female headers to the Ethernet wires.
Once all the connections have been made, screw the relay modules to the bottom of the
junction box using spacers and screws, screw the lid to the junction box, screw both outlets to
the lid, then screw the face plate over the outlets. Connect an Ethernet cable from the
external RJ45 port of the control box to the RJ45 surface mount jack of the control panel. We
now have a power control box that can independently control the power to four outlets using
a common Ethernet cable.

WARNING: Always be ready to quickly disconnect power when testing electrical devices
you built yourself for the first time. No one likes to see smoke come from their hardware,
but excessive damage can sometimes be avoided by quickly removing power.

Of course, if you want to buy a commercial version of a power control box like this instead of
building it, you can do that as well.
This concludes the hardware section of the build process and getting the hydroponic system
ready to be automated with the Mycodo software. Read on to learn how to install the
Raspberry Pi operating system on your Raspberry Pi and install and configure Mycodo to
automate the system.
Installing and Configuring the Automation
Software
Install Raspberry Pi Operating System
Now that we have a functioning hydroponic system and all our electrical hardware has been
set up, we can install the operating system and Mycodo and begin configuring the system to
run autonomously.

You should always follow the latest instructions for installing Mycodo, on GitHub, as this will
be where the most up-to-date instructions can be found.

As of this article’s writing, briefly, you will need to install the Raspberry Pi operating system
to a micro SD card by downloading the latest Raspberry Pi OS Lite Image, then use one of the
recommended flashing tools (Etcher or the new Raspberry Pi Imaging Tool) to burn it to your
micro SD card. Once the image has been flashed, you should see one of the SD card partitions
recognized by your computer, labeled “boot”. Create a blank file in this boot drive named “ssh”
(with no file extension). This will allow the ability to connect to the Raspberry Pi via a secure
shell (SSH) connection after booting it for the first time. Also, if you would like your Raspberry
Pi to automatically connect to your WiFi router, create another file on the boot drive named
“wpa_supplicant.conf”. I recommend using a good text editor for this (such as Notepad++ or
Sublime Text), as Windows notepad or other simple text editors often cause errors due to
extra characters they create in the files. In “wpa_supplicant.conf”, add the following text,
changing “NETWORK-NAME” to your WiFi’s network name and “NETWORK-PASSWORD”
to your WiFi’s password, then save.

1. country=US
2. ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev
3. update_config=1
4.
5. network={
6. ssid="NETWORK-NAME"
7. psk="NETWORK-PASSWORD"
8. }

Insert the micro SD card into your Raspberry Pi, then plug in the power. After a minute or so,
it should be connected to your WiFi. Alternatively you can connect it to your router with an
Ethernet cable. Now, you will need to find the IP address of your Raspberry Pi that’s
connected to your network. This can often be done by logging into your router and viewing
the devices that have connected through DHCP, which should show a device named
“raspberry”. Note the IP address of your Raspberry Pi. Since multicast DNS is supported by
the Raspberry Pi OS, you might also be able to reach your Pi at the address raspberry.local.
Open a terminal or command prompt and type ping raspberry.local and if you get a
response, you’ll find your Raspberry Pi’s IP address. More information about other methods
to find your Raspberry Pi’s IP address can be found here.

Next, start up your favorite terminal. Linux and Mac already have terminals you can use for
this, but if on windows, I recommend Putty. If on Linux/Mac, open a terminal and use the
command ssh [email protected] replacing “192.168.0.50” with the IP address of your
Raspberry Pi that you noted earlier. Putty requires settings the host to [email protected]
and port to 22 to connect. You should be prompted for a password, which by default is
“raspberry”. Once logged in, you should be looking at a terminal with a flashing cursor at the
end of pi@raspberry:~ $

Before installing Mycodo, it is recommended to use the command sudo raspi-config and
change the default password to something more secure, set the time zone, set the wifi
country, then reboot for the changes to take effect. More information about raspi-config can
be found here.

Install Mycodo
Now we can install Mycodo. Again, always follow the latest instructions for installing Mycodo,
on GitHub. As of the writing of this article, it is a single command that is entered in the
terminal:

curl -L https://kizniche.github.io/Mycodo/install | bash

It will take a few minutes for Mycodo to install. Once complete, the text on the terminal will
display the web address to use to access the web user interface (UI). Open a web browser to
this address and you will be greeted with a screen to create a new administrator user, then
the login page.

Port Forwarding
If you would like to access Mycodo from outside your local network, you will need to log in to
the modem/router that your Raspberry Pi is connected to and forward port 443 to your
Raspberry Pi’s IP address. This can be as simple as forwarding external port 443 to your Pi’s
port 443, or you can forward a different port, such as 5000 to 443 if you want to access your
Pi from an non-standard port. Keep in mind the modem/router that your Raspberry Pi is
connected to will need to be connected to the internet for this to work. Once forwarded, you
should be able to access your Raspberry Pi from anywhere with internet by visiting your
modem’s IP address. Your modem’s IP can be found by visiting whatsmyip.com while
connected to your home network. If your modem’s IP address is 122.22.22.22, and port 443
was forwarded to port 443 of your Raspberry Pi, you can simply open a browser to
https://122.22.22.22 (notice the s in https), however, if port 5000 was forwarded to port 443,
then you would specify the port and navigate to https://122.22.22.22:5000.

If your IP address changes, you won’t be able to connect from outside your network until you
find what the new IP address is. This is where dynamic DNS can provide you a hostname that
will point to your IP address, and will update when your IP address changes. This allows you
to connect, for instance, to https://mypi.dyndns.org and be assured it will always forward to
your modem’s current IP address. Check out freedns.afraid.org for free dynamic DNS
services.

Configuring Mycodo
While following this guide, keep in mind that Mycodo is constantly being updated and what’s
documented here may deviate from what can be found in the latest version of Mycodo you
may have installed. Therefore, if you notice any discrepancies, it’s recommended to first refer
to the Mycodo Manual (also available as PDF, HTML, and TEXT) before seeking support
elsewhere.

Adding and Configuring Inputs

After logging in, ensure all inputs that are connected by I2C are detected by the operating
system. To do this, you can either execute the command i2cdetect -y 1 in a Raspberry Pi
terminal or navigate to the Configure -> System Information page on the Mycodo web UI
and find the section “I2C Bus”. You should see an address for each I2C device or sensor that’s
connected (figure, below).
Navigate to the Setup -> Data page and use the drop-down menu to search for each of your
sensors and add them. The default configuration options are usually okay to use, however
some inputs may need additional configuration prior to activation:

If using the ADS1115 analog-to-digital converter, the voltage that’s measured will
need to be converted to Amps. If using the clamp transformer and current sensor
board listed in the Materials list, you’ll need to set your ADS1x15 Input options as
follows: Unscaled Unit Max to 2 (volts), Rescaled Unit Max to 20 (amps), and Rescaled
Measurement to “Electrical Current: Amp (A)”. If using the Greystone current sensor
listed in the Materials List, switch the transformer’s switch to L and set your ADS1x15
Input options as follows: Unscaled Unit Max to 0 (volts), Rescaled Unit Max to 10 (amps),
and Rescaled Measurement to “Electrical Current: Amp (A)”.
For the Water Flow Input, you’ll need to set the Flow Meter Type to “Atlas Scientific
3/8″ Flow Meter”.
For both the Electrical Conductivity and pH Inputs, set the Temperature Compensation
Measurement to the Atlas Scientific PT-1000 temperature, which will improve accuracy
by calibrating each sensor to the temperature of the water prior to acquiring
measurements. Note: The PT-1000 Input will have to be created before it will be
available for selection in other configuration menus.
For the MH-Z19 sensor, if using the UART-to-USB board, set the Input’s UART Device
to “/dev/ttyUSB0”.
To set up the Input for the float sensor, add a “GPIO State” Input, then set Pin (GPIO) to
the GPIO pin you used when the float sensor circuit was made, previously.

Review the rest of the Input options, then activate your Inputs.
You can then verify measurements are being successfully recorded by navigating to the Data
-> Live page (figure, below).
Adding and Configuring Outputs
First, ensure all Atlas Scientific peristaltic pumps have been changed to I2C mode.
Additionally, all pumps come from the factory set to the same I2C address by default, but are
able to be changed with a simple command. If multiple pumps with the same address were to
be connected at the same time, we wouldn’t be able to communicate with them individually.
Therefore, we’ll first need to set each pump to a unique I2C address before they can be used
independently by Mycodo. First, disconnect all pumps from the Raspberry Pi’s I2C bus. Then,
start the Atlas Scientific I2C script in a terminal with the following command:
 ~/Mycodo/env/bin/python ~/Mycodo/mycodo/devices/atlas_scientific_i2c.py

Next, use the “List_addr” command to list the I2C addresses of the connected devices.
Connect one of the pumps and execute the command again to determine which address the
pump is using. The pump should be using 0x63 by default, which as an integer is 100 (and is
how the script should list it). Use the command “Address,100”, to start communicating with
address 100. Now, you can issue the command “I2C,50” to change the address of the pump to
50 (0x32). Repeat the procedure with all four pumps, using a unique and unused address for
each pump. Since the addresses 0x63, 0x64, 0x66, and 0x68 are in use by other devices, we
can’t use 4 consecutive address next to the default address, so I chose to use 0x32 (50), 0x33
(51), 0x34 (52), and 0x35 (53) for my four pumps. Once all pumps have had their address
changed to unique addresses, you can plug them all in at the same time, verify you’ve
successfully changed their addresses, and continue with their setup.

Navigate to the Setup -> Output page and add four “On/Off (GPIO)” and four “Atlas Scientific
Pump (I2C)” outputs. Set the Pin (GPIO) for each On/Off output (see the Raspberry Pi pin map
image, above) to the pins you connected your control box signal wires to earlier. Then, test
whether the outputs works by plugging in a device, such as a lamp, and clicking On for the
corresponding output. Set up each pump by setting the proper I2C Address you set for each
pump earlier, using the hex format (e.g. 0x32, 0x33, etc.). Then, test that each pump dispenses
when you instruct it to dispense a volume.

In line with the previously-mentioned light setup, I named my grow light output Grow Light.
This is all you need to do for the output configuration for your light, if you have a simple light
configuration that you only desire to control one outlet (which can have one or several lights
attached to). If a second light is desired to create more neutral light for automated
photography (time-lapse), add the additional light outputs, below.

Create “On/Off (GPIO)” output named Photo Light and set the GPIO pin connected to the
relay controlling the output with the light used to take photos plugged in. Create an On/Off
(Python Command) output named Grow Light On Only, and set the On Command to to
following, changing the IDs to correspond to the IDs of your Photo Light and Grow Light
outputs:

1. # photo light off

2. control.output_off('PHOTO_LIGHT_ID', output_channel=0)
3. # grow light on
4. control.output_on('GROW_LIGHT_ID', output_channel=0)

Set the Off Command to to following, changing the ID to correspond to the ID of your Grow
Light output:
 1. # grow light off
2. control.output_off('GROW_LIGHT_ID', output_channel=0)

Create an On/Off (Python Command) output named Photo Light On Only, and set the On
Command to to following, changing the IDs to correspond to the IDs of your Photo Light and
Grow Light outputs:

1. # photo light on
2. control.output_on('PHOTO_LIGHT_ID', output_channel=0)
3. # grow light off
4. control.output_off('GROW_LIGHT_ID', output_channel=0)

Set the Off Command to to following, changing the IDs to correspond to the IDs of your Photo
Light and Grow Light On Only outputs:

1. import time
2. # photo light off
3. control.output_off('PHOTO_LIGHT_ID', output_channel=0)
4. time.sleep(1)
5. # grow light only off
6. control.output_off('GROW_LIGHT_ON_ONLY_ID', output_channel=0)

This allows the Grow Light to be the only light on when Grow Light On Only is turned on, then
Photo Light On Only can be used to turn Grow Light off and Photo Light on during photo
acquisitions. For a more detailed description of the use of these functions, see the Lighting
section of this publication.
Calibration
The peristaltic pumps as well as the pH and electrical conductivity sensors may require
calibration before they operate accurately. Mycodo has built-in functions for calibrating all of
these devices, and provides an easy walk-through guide to instruct you at each step of the
way to calibrate them. Navigate to the More -> Calibration & Setup page and select which
device you want to calibrate, then follow the instructions. Note: Inputs being calibrated will
need to be deactivated prior to beginning calibration. Briefly, the steps for calibrating each
device are as follows.
The peristaltic pump calibration entails priming the tubing to remove all air, then instructing
the pump to dispense a calibration amount, such as 10 ml. Measure how much fluid was
actually dispensed (e.g. 9 ml), and input this value when prompted by the Mycodo calibration.
The pump will now accurately dispense volumes and can be verified by instructing it to
dispense a volume from the Setup -> Output page and verifying it actually dispenses that
amount. Make sure to calibrate all peristaltic pumps.

The pH sensor calibration will require either two or three calibration solutions. A three-point
calibration will yield higher accuracy than a two-point calibration, but two is still better than
none. These solutions should be at 4, 7, and 10 pH. The probe should be placed in each
solution and allowed to acclimate before the calibration to each standard pH solution is
carried out. Rinse the probe with distilled water before being placed into the next solution.
The calibration steps are repeated until all solutions have been used to calibrate.

The electrical conductivity sensor will require two calibration solutions. Enter the values of
the solutions when prompted, then insert the probe in the first solution and instruct Mycodo
to calibrate to that solution. Rinse the probe with distilled water and repeat for the second
solution.

Adding and Configuring Functions

Navigate to the Setup -> Function page, where several functions will be created that will
automate the system. They include:

Timer to turn on a grow light from 6 AM to 8 PM.

Function to regulate electrical conductivity and pH with the peristaltic pump outputs.
Function to regulate vapor pressure deficit with a humidifier and exhaust fan.
Function to monitor air and water temperature, water level, or other important
measurements and alert by email if they fall out of acceptable ranges.
Timer to periodically exhaust air.
Other functions for your application (e.g. I’m using a timer to trigger my DSLR camera
to capture photos for a high quality time-lapse).
Grow Light Schedule
Add “Trigger: Timer (Daily Span)” and “Trigger: Timer (Daily Point)” functions. For the Daily
Span Timer, add an “Output (Duration)” Action, then configure the Output by selecting your
Grow Light Output, setting State to “On”, then save. Next, set the Start Time to “6:00” and End
Time to “20:00”, then save. For the Daily Point Timer, also add an “Output (Duration)” Action,
and set the Output to your Grow Light Output and State to “Off”, then save. Next, set the Start
Time to “20:01”, then save. Activate both controllers and your grow light output will remain
on between 6 AM and 8 PM and turn off at 8:01 PM. If using a Photo Light output, select the
Grow Light on Only output instead of the Grow Light output.
Electrical Conductivity and pH Regulation
Add a “Controller: Conditional” function. Set the Period to “300”, then add two “Measurement
(Single, Last)” Conditions, four “Output (Duration)” Actions, and one “Email (Single)” Action.
Set one Measurement Condition to the pH measurement, and the other to the electrical
conductivity measurement. Set the first Output to the first pump, with the State “on” and the
Duration “1”, which will dispense the acid. Set the second Output to the second pump, with the
State “on” and the Duration “1”, which will dispense the base. Set the third Output to the third
pump, with the State “on” and the Duration “3”, which will dispense the nutrient A. Set the
fourth Output to the fourth pump, with the State “on” and the Duration “3”, which will dispense
the nutrient B. Last, set the Email Action to the email you would want to get alert
notifications if the measurements aren’t working. Next, copy the following code to the
Conditional Statement, changing “PH_MEASURE_ID” to the ID associated with the pH
measurement, “EC_MEASURE_ID” to the ID associated with the electrical conductivity
measurement, “IDpump1” to the pump 1 ID, “PUMP_2_ID” to the pump 2 ID, “PUMP_3_ID” to
the pump 3 ID, “PUMP_4_ID” to the pump 4 ID, and “EMAIL_ID” to the email action ID. These
IDs are found under each Condition or Action that was created earlier.

1. ### Edit below to set desired ranges for pH and electrical

conductivity ###
 2.
3. # Desired range for electrical conductivity
4. range_ec_high = 1300
5. range_ec_low = 1000
6.
7. # Desired range for pH
8. range_ph_high = 6.2
9. range_ph_low = 5.5
10.
11. # pH range that will immediately cause a pH correction
12. range_ph_high_danger = 7.0
13. range_ph_low_danger = 5.0
14.
15. ### Edit below to set the IDs for Conditions and Actions ###
16.
17. condition_id_measurement_ph_id = "{PH_MEASURE_ID}" # Condition:
measurement, last, pH Input
18. condition_id_measurement_ec_id = "{EC_MEASURE_ID}" # Condition:
measurement, last, EC Input
19. action_id_pump_1_acid = "{PUMP_1_ID}" # Action: Pump 1 (Acid)
20. action_id_pump_2_base = "{PUMP_2_ID}" # Action: Pump 2 (Base)
21. action_id_pump_3_nutrient_a = "{PUMP_3_ID}" # Action: Pump 3
(Nutrient A)
22. action_id_pump_4_nutrient_b = "{PUMP_4_ID}" # Action: Pump 4
(Nutrient B)
23. action_id_email_notification = "{EMAIL_ID}" # Action: Email
Notification
24.
25. ### DO NOT EDIT BELOW THIS LINE UNLESS YOU KNOW WHAT YOU ARE DOING ###
26.
27. import time
28.
29. if 'notify_ec' not in self.variables: # Initiate EC notification
timer
30. self.variables['notify_ec'] = 0
31. if 'notify_ph' not in self.variables: # Initiate pH notification
timer
32. self.variables['notify_ph'] = 0
33. if 'notify_none' not in self.variables: # Initiate None measurement
notification timer
34. self.variables['notify_none'] = 0
35.
36. measure_ec = self.condition(condition_id_measurement_ec_id)
37. measure_ph = self.condition(condition_id_measurement_ph_id)
38. self.logger.debug("Conditional check. EC: {}, pH:
{}".format(measure_ec, measure_ph))
39.
40. if None in [measure_ec, measure_ph]:
41. if measure_ec is None:
42. self.message += "\nWarning: No EC Measurement! Check sensor!"
 if measure_ph is None:
43.
44. self.message += "\nWarning: No pH Measurement! Check sensor!"
45. if self.variables['notify_none'] < time.time(): # Only notify
every 12 hours
46. self.variables['notify_none'] = time.time() + 43200 # 12
hours
47. self.run_action(action_id_email_notification,
message=self.message) # Email alert
48. return
49.
50. # First check if pH is dangerously low or high, and adjust if it is
51. if measure_ph < range_ph_low_danger: # pH dangerously low, add base
(pH up)
52. msg = "pH is dangerously low: {}. Should be > {}. Dispensing 1 ml
base".format(measure_ph, range_ph_low_danger)
53. self.logger.debug(msg)
54. self.message += msg
55. self.run_action(action_id_pump_2_base) # Dispense 1 ml base (pH
up)
56. if self.variables['notify_ph'] < time.time(): # Only notify every
12 hours
57. self.variables['notify_ph'] = time.time() + 43200 # 12 hours
58. self.run_action(action_id_email_notification,
message=self.message) # Email alert
59. elif measure_ph > range_ph_high_danger: # pH dangerously high, add
acid (pH down)
60. msg = "pH is dangerously high: {}. Should be < {}. Dispensing 1 ml
acid".format(measure_ph, range_ph_high_danger)
61. self.logger.debug(msg)
62. self.message += msg
63. self.run_action(action_id_pump_1_acid) # Dispense 1 ml acid (pH
down)
64. if self.variables['notify_ph'] < time.time(): # Only notify every
12 hours
65. self.variables['notify_ph'] = time.time() + 43200 # 12 hours
66. self.run_action(action_id_email_notification,
message=self.message) # Email alert
67.
68. # If pH isn't dangerously low or high, check if EC is within range
69. elif measure_ec < range_ec_low: # EC too low, add nutrient
70. self.logger.debug("EC: {}. Should be > {}. Dosing 3 ml Nut A, 3 ml
Nut B".format(measure_ec, range_ec_low))
71. self.run_action(action_id_pump_3_nutrient_a) # Dispense 3 ml
nutrient A
72. self.run_action(action_id_pump_4_nutrient_b) # Dispense 3 ml
nutrient B
73. elif measure_ec > range_ec_high: # EC too high, add nutrient
74. msg = "EC: {}. Should be < {}. Need to add water to
dilute!".format(measure_ec, range_ec_high)
75. self.logger.debug(msg)
76. if self.variables['notify_ec'] < time.time(): # Only notify every
12 h
 12 hours
self.variables['notify_ec'] = time.time() + 43200 # 12 hours
77.
78. self.message += msg
79. self.run_action(action_id_email_notification,
message=self.message) # Email alert
80.
81. # If EC is in range, make sure pH is within range
82. elif measure_ph < range_ph_low: # pH too low, add base (pH up)
83. self.logger.debug("pH is {}. Should be > {}. Dispensing 1 ml
base".format(measure_ph, range_ph_low))
84. self.run_action(action_id_pump_2_base) # Dispense 1 ml base (pH
up)
85. elif measure_ph > range_ph_high: # pH too high, add acid (pH down)
86. self.logger.debug("pH is {}. Should be < {}. Dispensing 1 ml
acid".format(measure_ph, range_ph_high))
87. self.run_action(action_id_pump_1_acid) # Dispense 1 ml acid (pH
down)

Activate the controller and it will run the code every 5 minutes, which will check the pH and
electrical conductivity and compare them to the first 6 variables which define the minima and
maxima of the allowable ranges for pH and electrical conductivity. A pH “danger” range is first
checked, which if the pH is outside this range, an email notification will be set and acid or base
will be dispensed until the pH returns to within this range. Next, if the pH is within the danger
range, but the pH is out of the normal range and the electrical conductivity is low, it will first
add nutrients A and B at an equal ratio (1:1) in 3 ml increments until the electrical
conductivity rises above the minimum that’s set. Only after the electrical conductivity is
within the desired range will acid or base be added to bring the pH into the desired normal
range.

These order of events were chosen because nutrients are typically acidic and will lower the
pH when added. If nutrients were added after the pH was adjusted, the pH could be brought
back out of the desired normal range and require readjustment. By adjusting the electrical
conductivity first, this issue is avoided. The only exception is if the pH becomes so low or high
that it falls outside the designated danger range, which will trigger an email notification and
cause the pH to be immediately adjusted with acid or base until it falls back within the danger
range. Once back within this range, electrical conductivity can be adjusted again. For
instance, this prevents the pH from becoming dangerously acidic if a large amount of water is
replaced in the system, requiring a large amount of nutrient solution to be dispensed to being
the electrical conductivity back to the desired range. This large amount of nutrient solution
will significantly lower the pH, which doesn’t usually occur under normal operating
conditions, where small amounts of nutrients are added at a time. Another benefit of this is
the email notification that the pH is found to be at an extreme, which could also indicate
something is wrong with the system.
Vapor Pressure Deficit Regulation
To effectively adjust the vapor pressure deficit, a humidifier and exhaust fan is required.
Essentially, we will change the vapor pressure deficit to an ideal range for plant health by
either adding humidity with a humidifier or removing humidity and heat with an exhaust fan.
Add a “Controller: PID” function, then set the following options: Measurement to “vapor
pressure deficit” (this should be a measurement included with a temperature/humidity
sensor), Direction to “Both”, Setpoint to “850”, Band to “350”, Kp Gain to “0.05”, Ki Gain to “0.05”,
Kd Gain to “0.01”, Integrator Min to “-1000”, Integrator Max to “1000”, Output (Raise) to
“Humidifier Output”, Min On Duration (Raise) to “5”, Output (Lower) to the “Exhaust Fan
Output”, and Min On Duration (Lower) to “5”. Save, then activate the controller.

Configure Email Notification/Alert Settings

For email notifications to work, you’ll have to first configure an email address that will be used
to send the emails. Gmail is a good option, but you will have to Enable Less-Secure Apps for
the account first. You may want to create a new Gmail address only to be used with sending
Mycodo notification emails. Your email credentials should be entered on the Configure ->
Alerts page. Once you have configured the required settings, enter your email address and
click “Send Test Email” to test whether your settings worked.

There are many alerts that can be created to notify you by email when a measurement
deviates from a set range, but I’ll document how to create two, and these examples can be
used to create ones for other measurements. Of the notifications I’ve used, the most
important are knowing when a measurement isn’t being acquired, which indicates a sensor
issue, knowing when the water level is low so I can replenish the water reservoir, and knowing
that the water and air temperatures are within an acceptable range.

Email Notification for Air Temperature

Add a “Controller: Conditional” function. Add the Condition “Measurement (Single, Last), and
set it to the measurement you want to monitor (e.g. air temperature). Add the Action “E-Mail
(Single)” and set it to the email address you will want to be notified. If you want to edit or add
email addresses, this can be done from the user configuration menu. Copy the following code
to the Conditional Statement, then replace “asdf1234” with the ID associated with your
Measurement Condition, and “qwer5678” with the ID associated with your Email Action.

1. air_temperature_C_high = 35
2. air_temperature_C_low = 5
3.
4. air_temperature = self.condition("{asdf1234}")
5.
6. if air_temperature is not None:
7. if air_temperature > air_temperature_C_high:
8. self.message += "Air temperature is too high! ({} C)
\n".format(air_temperature)
9. elif air_temperature < air_temperature_C_low:
10. self.message += "Air temperature is too low! ({} C)
\n".format(air_temperature)
11. else:
12. return
13. self.run_action("{qwer5678}", message=self.message)
14. else:
15. self.message += "Cold not find an air temperature measurement.
Check your sensor!"
16. self.run_action("{qwer5678}", message=self.message)

The two variables at the top set the minimum and maximum allowed temperatures. If the
measured temperature falls outside this range, a message is generated and emailed informing
the user the temperature is too high or too low. An email notification is also sent if a
measurement cannot be found, indicating there may be an issue with the sensor.
Email Notification for Water Level
Add a “Controller: Conditional” function. Add the Condition “Measurement (Single, Last), and
set it to the GPIO State Input measurement. Add the Action “E-Mail (Single)” and set it to the
email address you will want to be notified. Copy the following code to the Conditional
Statement, then replace “asdf1234” with the ID associated with your Measurement
Condition, and “qwer5678” with the ID associated with your Email Action.

1. bool_water_float = self.condition("{asdf1234}")
2.
3. if bool_water_float is not None:
4. if bool_water_float == 1:
5. self.message += "The water level is low. Add water.\n"
 6. self.run_action("{qwer5678}", message=self.message)
7. else:
8. self.message += "No water float measurement found!"
9. self.run_action("{qwer5678}", message=self.message)

This code is configured to send an email notification that the water level is low when the float
sensor GPIO is pulled to 3.3-volts (high). If, however, your float sensor circuit pulls the GPIO
pin low (0 volts) when the water level is low, change line 4 from this:

if bool_water_float == 1:

to this:

if bool_water_float == 0:

Exhaust Timer
Add a “Trigger: Timer (Duration)” function, then add the Action “Output (Duration)” and set it
to the exhaust output and the State to “on” and the Duration to “60” seconds, then save the
Action. Set the function Period to “3600”, save, then activate. This will turn the exhaust fan on
for 60 seconds every 3600 seconds (1 hour).

DSLR Camera Timelapse

The following configures a DSLR camera to acquire photos at a regular interval using a
secondary neutral light described previously in this publication. This configuration also makes
use of two additional outputs that control a remote shutter cable connected to the camera by
activating one output to actuate the focus (DSLR Focus) and another output to activate the
shutter (DSLR Shutter). You may need to adjust your configuration to suit your needs if your
system operates differently than this.

Add a “Controller: Conditional” function and set Period (Seconds) to 7200 to capture photos
every 2 hours (or any other desired period). Add the Action “Output (Duration)” and set
Controller ID to the “Photo Light On Only” output and the State to “on”. Add the Action
“Output (Duration)” and set Controller ID to the “Photo Light On Only” output and the State to
“off”. Add the Action “Output (Duration)” and set Controller ID to the “DSLR Focus” output and
the State to “on”. Add the Action “Output (Duration)” and set Controller ID to the “DSLR Focus”
output and the State to “off”. Add the Action “Output (Duration)” and set Controller ID to the
“DSLR Shutter” output and the State to “on”. Add the Action “Output (Duration)” and set
Controller ID to the “DSLR Shutter” output and the State to “off”. Copy the following code to
the Conditional Statement, then replace “PHOTO_LIGHT_ON_ONLY_ON_ID” with the ID
associated with the Condition turning Photo Light On Only on,
“PHOTO_LIGHT_ON_ONLY_OFF_ID” with the ID associated with the Condition turning Photo
Light On Only off, “DSLR_FOCUS_ON_ID” with the ID associated with the Condition turning
DSLR Focus on, “DSLR_FOCUS_OFF_ID” with the ID associated with the Condition turning
DSLR Focus off, “DSLR_SHUTTER_ON_ID” with the ID associated with the Condition turning
DSLR Shutter on, and “DSLR_SHUTTER_OFF_ID” with the ID associated with the Condition
turning DSLR Shutter off.

1. import time
2.
3. self.run_action("{PHOTO_LIGHT_ON_ONLY_ON_ID}") # Photo Light On Only
(ON)
4. time.sleep(3)
5.
6. self.run_action("{DSLR_FOCUS_ON_ID}") # Stage 1 On
7. time.sleep(1)
8. self.run_action("{DSLR_SHUTTER_ON_ID}") # Stage 2 On
9. time.sleep(1)
10. self.run_action("{DSLR_SHUTTER_OFF_ID}") # Stage 2 Off
11. time.sleep(1)
12. self.run_action("{DSLR_FOCUS_OFF_ID}") # Stage 1 Off
13.
14. self.run_action("{PHOTO_LIGHT_ON_ONLY_OFF_ID}") # Photo Light On Only
(OFF)

This will turn only the Photo Light on during the photo capture. For a more detailed
description of how this works, see the previous two sections discussing Lighting and Outputs.

Adding and Configuring the LCD

On the Setup -> Function page, add a “20×4 Generic LCD” (or “16×2 Generic LCD” if using a
16 character, 2 line LCD). Set the Address (I2C) to the address of your LCD and the Period to
how often you want the LCD to update or display the next set of lines. For configuring what is
displayed on the LCD, there are display sets. By default, there is one display set, which is a set
of data you can set to be displayed on the lines of the LCD. If you would like to display more
data, you can change “Number of Line Sets” from 1 to your desired number of line sets to be
displayed. You can create as many display sets as desired. The Period will determine how often
the display cycles between these display sets. Activate the LCD and verify the LCD starts
displaying data.

LCD Function – Newer versions of the software may look different.

Adding and Configuring the Camera

If using a Raspberry Pi camera, ensure it’s properly connected and it has been enabled
through the menu accessible with the command sudo raspi-config or from the Configure
-> Raspberry Pi menu. Make sure you reboot after enabling the camera for your changes to
take effect.

Navigate to the More -> Camera page, select the “PiCamera” Library (or “opencv” or
“fswebcam” if using a USB camera), then click Add. Click “Capture Still” to see if an image can
be acquired from the camera. There are a number of options that can be configured, including
Output, which is the output to be turned on prior to acquiring an image. This is useful for
taking photos at night by turning on a light while the photo is being acquired. See the previous
Lighting and Output sections for descriptions of a custom lighting configuration that enables
activating a photo light if you desire more neutral lighting during photos than your grow light
can provide. The setting Duration (Output) is the number of seconds to turn the output on
prior to taking the photo, which is useful if your lights take several seconds to turn on or
warm up (such as with fluorescent bulbs). There are also options for setting the still image
and live streaming resolutions. It’s advised to first test with the default settings, as certain
settings combinations may prevent images from being acquired properly.
Adding and Configuring a Dashboard
On the Data -> Dashboard page, there are many types of widgets that can be added,
organized by dragging, and resized. Perhaps the most useful widget is the Graph. Graphs
allow you to select any number of Inputs, Outputs, PID controllers, and other measurements
to be displayed on a historical graph. Graphs are updated automatically with new data, so you
always see the latest measurements. There are also a number of settings to tune the graph to
your liking, including the x-axis duration, series colors, and range selector, among others. You
can also create multiple dashboards to organize different views, or to prevent one dashboard
from becoming too cluttered. I won’t go into too much detail here, but suffice to say,
configuring a dashboard and exploring the various widgets is one of the more fun experiences
in Mycodo.
Automation Epilogue
Now that automation has been set up and is running, you’ll want to run your system for at
least a few days before entrusting it to care for your plants. This will give you a chance to find
potential issues and correct them. You’ll want to ensure the system can properly adjust
nutrients and pH. You can test this by removing water from the reservoir and replacing it with
fresh water, which should raise the pH and lower the electrical conductivity. You should also
ensure lighting, exhaust, and other functions are operating as expected.

Plant Cultivation
Once Mycodo has been configured and monitored for a sufficient amount of time to
determine it’s operating as expected, it’s safe to begin growing plants.

Biocontrol
With the system operational, it’s only a matter of when, not if, undesirable microorganisms
will find the large amount of nutrient-rich water and begin to feast, potentially causing plant
disease and crop loss. It’s far easier to initially establish desirable microorganisms in the
system at the beginning to competitively-exclude undesirable microbes, than to try to
exclude undesirable microbes later if they’ve been allowed established themselves.
Therefore, I recommend inoculating a new water system with a healthy community of
microorganisms.

There is debate about the best method for doing this, but what’s worked for me is introducing
a small soil sample as well as a commercial biocontrol agent. The soil will introduce a large
number of soil-associated microorganisms and the biocontrol agent will inoculate the system
with a known beneficial microorganism known to combats pathogens. To inoculate with soil, I
placed 20 g of potting soil in a small mesh bag and steeped it in the water reservoir for 30
minutes. For the biocontrol agent, I chose to use the bacterium Streptomyces griseoviridis,
which has been observed to combat pathogens and increase yields. This bacterium is sold
under the commercial name Mycostop. This microorganism develops a mutualistic
relationship with plant roots, known as the rhizophere, providing a benefit to both the
bacterium and the plant. Bacteria of the genus Streptomyces are known to produce numerous
antimicrobial compounds, and this is believed to be how this organism confers protection to
and improves the health of plants it colonizes. This organism is of a natural origin (non-GMO),
food-safe, used in commercial food production, and is certified organic. Mycostop can be
purchased in different amounts, from 2 grams to several kilograms. A two-gram packet is
more than enough for several inoculations of a small water system.

It should be noted that deep-cleaning to remove potential contaminants is not advised

(unless you had a particularly bad contamination occur), as this could negatively impact the
established S. griseoviridis and other beneficial microbes. Rather, it is recommended not to use
any chemical or mechanical cleaning methods following inoculation of biocontrol agents. This
is the beauty of biocontrol. Rather than decontaminating undesired microorganisms
following an infection, we’ve purposefully established a beneficial community that will
compete with any pathogens trying to establish themselves. Although this is will not be able
to combat every pathogen, as long as we support the health of our beneficial microorganisms,
we’ve conferred reasonable protection to our system. You should still monitor the health of
your plants and respond quickly to any decreases in health or yields by assessing the
underlying cause and developing actions to mitigate the issue.

Planting Seeds
Before planting, you should consider the quantity and rate you would like to harvest. If you
plant to the maximum capacity of your system, you will harvest a large amount at once, with
long periods between harvests. However, if you stagger your planting, your harvests will be
staggered as well, with smaller harvests at a more frequent rate. Below is one example
schedule for growing lettuce in a system using 6 channels with 10 planting holes per channel,
that will continually yield a 10-plant harvest every week, starting at the 8th week. Note that
every week 10 seeds are planted, however you might consider planting 11 or 12 seeds to
ensure you have enough to fill the entire channel should some seedlings not perform well.

Week Actions

1 Start 10 seeds germinating (set A)

2 Continue growing seeds in nursery tray (set A).

Start 10 seeds germinating (set B).

3 Plant 10 seedlings (set A) in channel 1.

Continue growing seeds in nursery tray (set B).

Start 10 seeds germinating (set C).

 Week Actions

4 Plant 10 seedlings (set B) in channel 2.

Continue growing seeds in nursery tray (set C).

Start 10 seeds germinating (set D).

5 Plant 10 seedlings (set C) in channel 3.

Continue growing seeds in nursery tray (set D).

Start 10 seeds germinating (set E).

6 Plant 10 seedlings (set D) in channel 4.

Continue growing seeds in nursery tray (set E).

Start 10 seeds germinating (set F).

7 Plant 10 seedlings (set E) in channel 5.

Continue growing seeds in nursery tray (set F).

Start 10 seeds germinating (set G).

8 Harvest plants from channel 1.

Plant 10 seedlings (set F) in channel 6.

Continue growing seeds in nursery tray (set G).

Start 10 seeds germinating (set H).

9 Harvest plants from channel 2.

Plant 10 seedlings (set G) in channel 1.

Continue growing seeds in nursery tray (set H).

Start 10 seeds germinating (set I).

… Continue this cycle and you will have a harvest of 10 mature lettuce plants every week.

For hydroponic systems, seeds are planted in supporting substrates rather than nutritive
substrates like soil. These are typically in the form of small cubes with a hole in the center to
plant a seed. Place one or two seeds in each cube’s hole, then place your substrates in a
germination tray. Hydrate the cubes with pure water. Don’t use the water from the
hydroponic system, as the fertilizer concentration will likely be too high and germination may
be disrupted or the seedling may become injured or have its growth stunted. If there’s a
significant amount of excess water that drained from the cubes while hydrating them, pour it
out from the tray to reduce the risk of microbial growth. Leave only the water in the cubes
and a small amount on the bottom of the tray. Place a lid on the tray to retain humidify and
place the tray in a warm area or on top of a heating mat. Check periodically that the cubes are
still moist and add water as needed.
It will generally take a few days to a week for your plants to germinate, depending on the
seed. Once the seeds have germinated, start watering them with a dilute nutrient solution,
roughly 1/4 to 1/2 the concentration of your hydroponic system (300 – 500 μS/cm), and
expose them to light. Grow the seedlings in the nursery tray for a week or two until the first
sets of true leaves form. This acclimates the plants to handle the more intense growing
conditions of the hydroponic system. Make sure you supply your seedlings with ample light to
prevent excessive etiolation (Josse and Halliday 2008). If long, weak stem growth is observed,
increase the light intensity. You can also slowly increase the nutrient concentration to
acclimate them to a higher concentration prior to transferring to the hydroponic system.
Once the seedlings are about 2 to 3 inches tall, break apart the individual cubes and place one
cube in each hole of the hydroponic channels.

Planted in this tutorial were the Salanova Premier Collection pelleted lettuce seeds (red
oakleaf, red butter, green oakleaf, and green butter lettuces) and Genovese pelleted basil
seeds.

Maturation
Over the next few weeks, as your plants mature, you may need to adjust nutrient
concentration, light intensity, air/water temperature, vapor pressure deficit, or other
parameters. Pay attention to the health of your plants and respond to any irregularities by
researching and investigating the symptoms you observe to develop an understanding of the
potential causes and to develop corrective measures. It may take several attempts before you
develop the skill and knowledge to consistently yield bountiful harvests.
For these lettuce varieties, be cautious of using too much light or too high of a nutrient
concentration, as this can cause your plants to grow too fast, resulting in dense center leaf
crowding, curled leaf tips, and tip burn. Expect that some degree of tuning will be required to
develop a configuration for optimal plant growth in your particular system. I recommend
searching the scientific literature for optimal parameters for your particular crop in order to
develop a good base from which to begin tuning. There is already ample literature on
hydroponic lettuce cultivation and optimal parameters for nutrients (Gent 2003; Cometti et
al. 2013), temperature (Thompson et al. 1998; Gent 2016), and light (Korkmaz et al. 1999;
Gent 2014; Zhang et al. 2018), among others.

Harvest
The lettuces grown in this article will typically be ready to harvest in approximately 4 to 6
weeks after planting, depending on the light intensity, nutrient concentration, and
temperature. To harvest, simply pull the plant out of the growth channel. These lettuces are
highly perishable, so it’s advised to harvest as close to the time you’re going to use them. If
storing for later use, place them in a refrigerator at or below 40 °F. These varieties of lettuce
are convenient to prepare because they can be broken into individual leaves by removing the
base of the stem from the bottom, either by twisting with your hand or cutting with a small
knife.
Future Directions
There’s a lot of possible modifications that can be made to improve or expand this system.
Here are just a few that came to mind:

Add a dissolved oxygen sensor to monitor the water dissolved oxygen (DO)
concentration or couple it with an air pump and air stone to regulate the amount of
DO in the water system.
Add an oxidation-reduction potential sensor to monitor the water oxidation-reduction
potential (ORP) or couple it with peristaltic pumps and solutions to regulate ORP.
Add ducting to the lighting fixtures to reduce heat buildup in the grow tent by
exhausting it outside the tent.
Add a solenoid valve to automatically open the flow of a fresh water source to the
reservoir when the level gets low.
Add an air or water heater to maintain an optimal temperature in cold environments.
Add a dehumidifier or additional exhaust mechanisms to regulate vapor pressure
deficit in high humidity environments.
 Add a thermal pad with a temperature sensor to regulate a consistent temperature for
germinating trays of seeds in a nursery area prior to planting in the channels.

Log In to View My System

My system was open to guest logins for the first month after this article’s release, and has
now been disabled while I work on my next project. Check back soon for my next article.

References
Barbosa GL, Gadelha FDA, Kublik N, Proctor A, Reichelm L, Weissinger E, Wohlleb
GM, Halden RU. 2015. Comparison of Land, Water, and Energy Requirements of
Lettuce Grown Using Hydroponic vs. Conventional Agricultural Methods.
International Journal of Environmental Research and Public Health 12:6879–6891.
DOI 10.3390/ijerph120606879
Cometti NN, Bremenkamp DM, Galon K, Hell LR, Zanotelli MF. 2013. Cooling and
concentration of nutrient solution in hydroponic lettuce crop. Horticultura Brasileira
31:287–292. DOI 10.1590/S0102-05362013000200018
Gent MP. 2012. Composition of hydroponic lettuce: effect of time of day, plant size,
and season. Journal of the Science of Food and Agriculture 92:542–550. DOI
10.1002/jsfa.4604
Gent MPN. 2014. Effect of Daily Light Integral on Composition of Hydroponic Lettuce.
HortScience 49:173–179. DOI 10.21273/HORTSCI.49.2.173
Gent MPN. 2016. Effect of temperature on composition of hydroponic lettuce. Acta
Hortic 95–100. DOI 10.17660/ActaHortic.2016.1123.13
Gent MPN. 2003. Solution Electrical Conductivity and Ratio of Nitrate to Other
Nutrients Affect Accumulation of Nitrate in Hydroponic Lettuce. HortScience
38:222–227. DOI 10.21273/HORTSCI.38.2.222
Josse E-M, Halliday KJ. 2008. Skotomorphogenesis: The Dark Side of Light Signalling.
Current Biology 18:R1144–R1146. DOI 10.1016/j.cub.2008.10.034
Korkmaz A, Pill WG, Cobb BB. 1999. Rate and Synchrony of Seed Germination
Influence Growth of Hydroponic Lettuce. HortScience 34:100–103. DOI
10.21273/HORTSCI.34.1.100
Mycostop Research Results. Retrieved March 8, 2020. <http://www.agbio-
inc.com/mycostop-research-results.html>
 Thompson HC, Langhans RW, Both A-J, Albright LD. 1998. Shoot and Root
Temperature Effects on Lettuce Growth in a Floating Hydroponic System. Journal of
the American Society for Horticultural Science 123:361–364. DOI
10.21273/JASHS.123.3.361
Zhang X, He D, Niu G, Yan Z, Song J. 2018. Effects of environment lighting on the
growth, photosynthesis, and quality of hydroponic lettuce in a plant factory.
International Journal of Agricultural and Biological Engineering 11:33–40. DOI
10.25165/j.ijabe.20181102.3420

Latest Articles

Mushroom Cultivation Automation: From Foraging to Fruiting

September 3, 2021

Mushrooms are a unique and often mysterious organism. They’re not quite plants
and not quite animals. As a fungus, mushrooms represent a distinct evolutionary
lineage. Fungi grow as a single-cell yeast form (used in beer, wine, and bread
making) and multi-cellular, thread-like mycelia that often produce fruiting bodies
(mushrooms) and sclerotia (e.g. truffles), and are responsible for the decomposition
of the vast majority of dead plant matter. Understanding life cycle of mushroom
fungi is a valuable skill if you wish to harness this knowledge for mushroom
cultivation.

Automated Hydroponic System Build

June 3, 2020

Hydroponic farming is a method of growing crops without soil, with the main
benefits of environmental and nutrient control, water conservation, and reduction
of labor. This technique relies on a number of technologies that the principles of
automation can be applied in order to improve yield and consistency. In this article
and accompanying video, I’ll show you how to built a hydroponic system with basic
components and automate it using the open source software Mycodo running on a
Raspberry Pi single board computer.

Remote Radiation Monitoring

 Remote Radiation Monitoring
August 19, 2019

Radiation is the transmission of energy through space in the form of a particle or

wave. These can be generated by instruments or […]

LoRaWAN Tracker and Mapper

April 4, 2019

LoRaWAN is a great long-distance, low-power transmission protocol. One of the

questions that often comes up after deploying a gateway and nodes is […]

Outdoor LoRaWAN Gateway

March 20, 2019

I’ve been heavily involved with sensors and remote sensing for many years.
Recently, I’ve become intrigued with remote sensing, and began my exploration […]

Bringing a Vintage Telegraph into the Digital Age

February 18, 2016

I recently inherited a key on board (KOB) telegraph that my late grandfather used
to practice Morse code with when he was a […]

← Remote Radiation Mushroom Cultivation

Monitoring Automation: From
Foraging to Fruiting →

154 thoughts on “Automated Hydroponic System Build”

Rafael Mata says:

June 4, 2020 at 9:08 pm
WOW !! I dream Come True,, Congratulations…..
Reply

Jason Greeley says:

June 6, 2020 at 10:08 pm
I’m wanting to start on making one of these, but due to the cost of components I’m
wondering if I can start with only one or two of the functions running to start, and add
further sensors and functionality down the road? Does the software go all crazy if it’s
not receiving input from all the specified sensors? I want to start with just ph sensor,
and ph control via the dosing pump, probably add temperature too since I know that
can be a factor in PH sensing accuracy.

Reply

Kyle Gabriel says:

June 6, 2020 at 10:28 pm
The software is scalable so you can add as few or as many inputs, outputs, and
functions as you desire. There’s no conflict from having too few devices.

Reply

retumile sedange says:

February 5, 2021 at 5:25 pm
dude thats impressive..so inspired

Reply

Sean Sullivan says:

January 28, 2022 at 4:21 pm
Awesome video and literature. If I add1 more peristaltic pump in the
system do I have to code it in t Mycodo or does it automatically
recognize it? How about the water fill and water dump solenoids? If I
added that does it require coding or is it tunable in Mycodo?

Reply

Kyle Gabriel says:

January 28, 2022 at 4:26 pm
 All of what you described can be done through the web interface,
no coding necessary. There are generic peristaltic pump and
generic on/off outputs you can add and configure with a few
clicks.

Reply

Sean Sullivan says:

January 28, 2022 at 5:56 pm
I had sent you a email asking about this and 1 more thing. If
I want to run 2 reservoirs like in your video except that
each reservoir is feeding 2 different trays at different EC.
Will that require programming? Im super excited to build
mine!

Kyle Gabriel says:

January 28, 2022 at 6:14 pm
No more programming than is mentioned in this build
article. For however many reservoirs/sensor arrays you
have, just multiply the steps for setting up nutrient/pH
regulation.

Sal says:
June 9, 2020 at 6:47 am
amazing job mate! congrats.

Reply
Pingback: AUTOMATED HYDROPONIC SYSTEM BUILD « Adafruit Industries –
Makers, hackers, artists, designers and engineers!

John Otsuki says:

June 11, 2020 at 10:05 am
This is one of the best written and physically solid presentations of a high end DIY
build. I am looking forward to making this as my next major project. Thank you for
doing such a great job and I look forward to hearing more.
Reply

David Kerrins says:

June 11, 2020 at 11:03 am
AWESOME presentation, truly. This is the best prototype system I have seen online.

Do you have any thoughts on aquaculture automation?

Any estimate on how automated (physical non-intervention) your solutions might

become?

Reply

Kyle Gabriel says:

June 11, 2020 at 11:07 am
I don’t have much experience with automating aquaculture (which I hope to
change soon), but I don’t think the principles would be much different than this.

Could you clarify your second question? So far, I’ve only had to replenish water
in the reservoir, since I didn’t have a water source near to connect to a solenoid
valve. So, there’s very little intervention needed as it stands.

Reply

Fredrick Juma Ouma says:

June 11, 2020 at 3:39 pm
I need a system like this with proposed modification. I have 15fish nursery each
producing 1000 litres of nutrient rich “waste” water. I need to convert this water into
resource for fruit n vegetable farming. Kindly let me know possible cost implication.
The Hydro Victoria Fish Farm also runs a RAS system(1200liters under circulation in a
20m by 15m godown roofed with ironsheets n walled by 60% greenhouse paper for
catfish n tilapia egg propagation (hatching n nursing).The farm is located along the
shores of Lake Victoria in Busia County, at Port Victoria town in Kenya (Africa).

Reply

Monkey House says:

June 11, 2020 at 6:35 pm
Looks like a fantastic setup. Great write-up too. Thanks for sharing it!

One thing that is missing IMHO is an itemised price indication for the BOM. This could
help prospective builders decide whether they want to pursue such an endeavour.

A lot of items are quite pricey (eg: Atlas Scientific). What is your estimated total cost
for the automation and sensing side of this project?

Reply

Kyle Gabriel says:

June 11, 2020 at 6:39 pm
Prices sometimes change and the Amazon Associate terms prohibit listing of
prices (because they change), but the total for everything except the tools
should be around $1200. You can save a few hundred $ by using generic
peristaltic pumps and run them for short durations of time. You could also use
less expensive water sensors, but non-lab-grade probes can’t handle being
submerged indefinitely like the ones produced by Atlas Scientific, so YMMV.

Reply

8BitSky says:
June 11, 2020 at 8:12 pm
If this were to be done in a classroom for an educational display (with the bonus of
green production), would the tent be required? From my understanding after reading
your post and some research, the tent helps to maintain humidity, temperature and
reflect light for maximum efficiency of LED’s. If a plexiglass front wall barrier was
constructed within a sealed “box” would the LED’s need to be on longer or would an
additional panel need to be added to account for the lost light? New to aquaponics,
have gardened before and have been teaching some courses on robotics and
automation lately. This seems like the perfect project to help tie everything together!

Reply

Kyle Gabriel says:

June 11, 2020 at 8:19 pm
I use the tent primarily for light retention and to be able to have more control
over the environment. If using grow lights, the tent is really useful, as the lights
are fairly intense and you wouldn’t want prolonged exposure to them. For
education, I would recommend building this against a window that gets a good
 amount of sunlight, with a slightly different configuration that has channels
stacked above one another, where one channel’s outflow goes into the next
channel, with the last channel returning the water to the reservoir. This open
design lets you see everything that’s happening. I believe this type of design is
called a greenwall, or is at least very similar to one.

Reply

8BitSky says:
June 11, 2020 at 9:09 pm
I work within the interior of the building, with no windows, only
skylights. I would be looking to partner this idea with a green wall using
vertical tubes, a reflective material on the backing wall, and the LED
grow lights positioned facing the wall (and plants). I have seen consumer
products set up within the classroom and the reflected light doesn’t
seem to be as harsh when the LED’s are not directly above or eye level.
The automation aspect of this project is what really caught my attention.
Thanks for the quick reply and sharing such an amazing project.

Reply

Chris Soukas says:

June 11, 2020 at 11:23 pm
Checking out your content, you should really learn how to use the mirror tool in fusion
360. I keep feeling so bad for you watching you recreate profiles!

You can also use the project function to make the geometry much faster.

Reply

Kyle Gabriel says:

June 11, 2020 at 11:29 pm
I’m familiar with mirror, I just chose not to use it here. I also did use project for
some of the construction. They’re simple parts, there’s no harm in taking my
time so the time-lapses come out longer than 2 seconds. The goal was to make
an entertaining video, not design the parts with the minimal number of actions.
Thanks for the tips!

Reply
 Kenny Tan says:
June 12, 2020 at 1:01 am
This looks awesome, anywhere to buy the smart system? I have a basic setup and like
to start automation. It would be great to start with your hardware and build top of it. :)
Thanks.

Reply

Kyle Gabriel says:

June 12, 2020 at 7:04 am
There is not. I provide the software and general guidance. Some assembly is
required.

Reply

Pree Kolari says:

June 17, 2020 at 12:34 pm
Awesome project. I just saw all the Mycodo work. Amazing! I will certainly try it out
the automation work in my aeroponic system. thanks also for open-sourcing this.

Reply

Anthony DeChiaro says:

June 17, 2020 at 2:22 pm
Fantastic work and effort to put this together, job very well done. I’ve seen some smart
hydro systems before, but never one fully automated with nutrient & acid pumps. Also
thanks for making me aware of Mycodo, looks super useful and have a few ideas for
use.

Out of curiosity, why go with a NFT system? And why no netpots?

Reply

Kyle Gabriel says:

June 17, 2020 at 2:51 pm
Thank you. I chose NFT because I had some familiarity with it and because of its
simplicity. With netpots, you need extra materials (pots) and a longer seedling
 growth period (for longer roots) before introduction into the system.

Reply

. says:
June 21, 2020 at 9:57 am
Have you looked at other studies on other beneficial bacteria other than Mycostop?

What did not make the cut and may be a waste of money?

Popular ones are Hydroguard, Liquid Ocra, Great White, UC Roots, BooBloom

Opinions also on roots stimulators / enzymes?

Reply

Derek Jacobs says:

June 23, 2020 at 11:56 am
You’ve noted controls for a humidifier, but there didn’t appear to be anything to
generate humidity in the Materials and Parts list. What did you use for this?

Reply

Kyle Gabriel says:

June 23, 2020 at 12:43 pm
Any consumer humidifier should be fine. I ended up using an ultrasonic
humidifier transducer with a float to keep the proper water level over the
transducer. This was set inside a 2 gallon jug with the lid cut off, with a laundry
detergent cup cut and placed over the top of the transducer to prevent water
from splashing out of the container. I didn’t record this being built, but I may
modify the article with a build section and add to the parts list. I’ll try to
remember to reply to this thread when I do.

Reply

Adam M says:
August 11, 2020 at 6:59 pm
Hi Kyle, Great project, I think I will be making one similar using your guides. I plan to
have a mix of 5V and 3.3V devices using I2C. I am confused about your section
“Connecting the Sensors,…” and especially the Note 1 about level shifting. You
mention all your devices are 3.3-5V compatible, so it seems like you are using 3.3V to
run SDA and SCL back to the pi, and would also be using 3.3V to power your devices.
However in all of your schematics you have the power for these devices connected to
the 5V pin on the RPi. Is this a mistake in the diagrams or am I missing something? I am
thinking of using a level shifter and powering my 5V devices with a separate voltage
source (or the pi 5V pin 2/4) and all the 3.3V devices from the pi 3.3V pin 1.

Reply

Kyle Gabriel says:

August 11, 2020 at 7:09 pm
It’s not an error, the schematics are how I actually wired it. The only
circumstance where using 5 volts on the I2C lines of the Pi is acceptable is if you
don’t have the SDA or SCL lines pulled to 5 volts. In reviewing my schematic, my
LCD requires 5 volts, but the rest of my I2C devices can operate on either 3.3 or
5 volts. My LCD doesn’t pull the I2C lines to the input voltage, it only pulls them
low for communication. I didn’t have a level shifter available when wiring, so it’s
what I went with. I should really go back and add one for the LCD and move the
other devices to the 3.3 volt line in the schematic, or add a note explaining this
anomaly in the schematics.

See:

1: https://www.raspberrypi.org/forums/viewtopic.php?
p=924630&sid=a312daa2503b9ff3f7a9c388ab1cbc24#p924630

2: https://raspberrypi.stackexchange.com/a/68200

Edit: I just updated the build documentation to include “Note 2” explaining the
anomaly.

Reply

Pierre says:
August 30, 2020 at 12:41 am
Yow this is fantastic. I might build & install this winter.

Question, how easy can we add a “nutrients schedule”?

From clone to grow room, we all start with a low EC and increase up to a “max” value
that is considered safe for ~90% of the strains. Worst case if there’s some burning,
hopefully you were around to notice and lower. When switch nutrients from grow to
bloom, we can again start to increase the EC level. Also consider some growing
method requires ~7 days flushing w/ clean water to “clean” the fruits.

Quickly I couldn’t tell if such schedule was possible.

Switching from grow to bloom and then to “flush” mode is often a decision taken by the
operator. Does Mycodo support such a thing? Quickly I’d say … maybe something like a
down box that allows operator to chose the profile/mode the system is running? Ex. :
Clone, Vegetable/grow, Flower/bloom, Flush. The scroll down menu would support
“create new profile” with different settings and be able to save it. Someone then create
the following profiles :

Clone-Week1

Clone-Week2

Grow-Week1

Grow-Weekn

Bloom-Week1

Bloom-Weekn

Flush

But once switching from grow to bloom, the system should be able to automatically
knows it juste entered Bloom-Week1 – keep note of this info somewhere. When it
reaches Week2, the application should automatically switch to week2 w/o the need of
the operator.

Also… when switching from grow to bloom nutrients, someone needs to empty the A &
B pumps’ tubes from the liquid currently in place. I would run this system w/ to 2 more
pumps.

Reply

Seann says:
September 21, 2020 at 6:57 am
Hi I would really love to build this and support you at the same time by purchasing
items from your list the problem is that I do not have enough experience and your
video goes way to fast for me as I have no experience. If you could make a video
showing where you are plugging each wire Or have more close ups images of the
wiring and bread boards so a novice can understand and build.
Thanks

Seann

Reply

Kyle Gabriel says:

October 5, 2020 at 6:22 pm
What I produce is not meant for someone with no experience (although that
shouldn’t discourage anyone from trying). There are a lot of other resources to
learn about hardware and software for a beginner. Best of luck.

Reply
Pingback: An automated hydroponic system #PiDay #RaspberryPi @Raspberry_Pi «
Adafruit Industries – Makers, hackers, artists, designers and engineers!

Claude Bugeja says:

September 27, 2020 at 11:44 am
Hi Kyle, very great project. Its going to help me for my secondary school project. One
thing is that the peralistic pumps are expensive can you tell me how to use normal
peralistic pumps(I mean by the ones that are controlled by positive and negative
wiring only) in the project. Thank you

Reply

Kyle Gabriel says:

September 27, 2020 at 11:49 am
Claude,

Generic Peristaltic pumps could always be used with Mycodo. You would simply
wire the pump to its power supply with a relay in-line with the positive wire.
Then, in Mycodo you would create an On/Off GPIO Output and set it to the pin
connected to control the relay. This allows the pump to be turned on for
durations of time. Additionally, since the release of this article, I’ve created a
Generic Peristaltic Pump Output module in Mycodo, which allows for
outputting specific volumes in addition to durations of time (following the
measurement of your particular pump’s flow rate in ml/min and setting this
value in the output options).

Reply
 Claude Bugeja says:
September 28, 2020 at 7:53 am
thanks a lot Kyle ,1) can you please send me a small simple schematic so i could
understand it more pls.

2)also the Ethernet cable wiring in the power box is mentioned to connect:

One Ethernet wire to the ground (-) input of each relay module.

One Ethernet wire to the positive (+) input of each relay module.

One Ethernet wire to the channel 1 signal input of the mechanical relay.

One Ethernet wire to the channel 2 signal input of the mechanical relay.

One Ethernet wire to the channel 1 signal input of the solid state relay.

One Ethernet wire to the channel 2 signal input of the solid state relay.

is there specific wires of the Ethernet cable that i need to connect or any cable is
acceptable??

3) and can apply the atlas scientific wifi hydroponics kit instead of the individual kits??
(it can connect raspberry pi via 4 way pbt connector that has scl,sda,grd,vcc)

Reply

Kyle Gabriel says:

October 5, 2020 at 6:18 pm
1. Sorry I am not able to generate schematics by request.

2. The color of wire does not matter, so long as each end is connected to the
proper place.
3. I’m not familiar with the kit you reference.

Reply

Karen Smith says:

September 28, 2020 at 9:38 pm
Did you have to change the water or add water regularly?

Reply

Kyle Gabriel says:

September 28, 2020 at 9:43 pm
 That depends on how many plants, the plant species, the air humidity, the
temperature, whether there has been a contamination, among other factors.
Water will be removed from the system, either by plants or evaporation, and
need to be replenished. Certain nutrients may build up and require refreshing
the reservoir. Water can be set to automatically be refilled with a water level
sensor and solenoid connected to a secondary water source.

Reply

David says:
October 4, 2020 at 10:10 am
Great video and instructions. I just set mine up with only ph so far and I love it.
Wondering if you were planning on adding any additional logging parse ability or a way
to upload it into powerbi or a like program. I’m interested in the trends over time for
my stats (once I have all the other probes) and the ability to review historical data
without loading the pi up. Something that could like upload/dump logs to a cloud
source would be cool. Otherwise this project is as perfect as could be. Im looking
forward to the ORP add on as well.

Reply

Kyle Gabriel says:

October 5, 2020 at 6:15 pm
There exists the ability to create your own Input, Output, and Function
modules, so exporting data can be accomplished any way you see fit. I’ve
created several modules myself for personal projects to transfer data over
LoRaWAN, The Things Network, public radiation monitoring maps, TTN
Mapper-Tracker, and others.

Reply

Dylan russell says:

October 6, 2020 at 9:06 pm
Awesome! Looking forward to giving it a try!

Reply

Dan300 says:
October 11, 2020 at 3:47 pm
You did an incredible job. I am in the process of carrying out the same project. To
facilitate the wiring I use the “Whitebox Labs Tentacle T3”.

I just found a small error, the default address of the atlas pump is 103 (0x67), it’s EC
which is 100. So I changed the address of the wrong module.

Keep up this great job. If you could add the CO2 Atlas that would be cool.

Reply

Claude Bugeja says:

October 13, 2020 at 3:00 pm
the probes that i have chose are all analog signaled can i attach an analog to digital
converter so i could connected them to scl and sda

Reply

Kyle Gabriel says:

October 14, 2020 at 7:40 am
Your answer can be found in the manual:
https://kizniche.github.io/Mycodo/Analog-To-Digital-Converters/

Reply

José Lobo says:

October 14, 2020 at 7:34 am
Hi, fantastic job. I have arduino project on the way but coding is not easy . Would like
to test yot system. Plese teel me one thig: I have here a Raspberry Pi model B+ with
512Mb. What is your opinion, will it be enough, even without camera or wifi? or
Model3 is really necessary? I wish i could find a use for the model 1 i have here :) Thank
you, and congrats

Reply

Kyle Gabriel says:

October 14, 2020 at 7:36 am
The version 1 Pi will likely work, but will lack performance compared to v2, 3,
and 4. The last v1 I ran Mycodo on was in 2015, and that was a much different
 piece of software than it is today. That Pi now sits in my history museum, aka
project graveyard.

Reply

José Lobo says:

October 14, 2020 at 6:35 pm
Thank you for quick answer. I live in front of a graveyard :) but still hope
to use the RPI before sending to the other side of the street:) For now is
working as an internet radio and works reasonably … I have ordered a
RPi3 to work with your program since i´ve an hydroponic system and i’m
sure i´ll learn a lot. Stay safe :)

Reply

Karen Smith says:

October 18, 2020 at 5:41 am
How many fans and other equipment did you use and for what purposes? Sorry I was
having trouble understanding the temperature and humidity regulation part

Reply

Kyle Gabriel says:

October 18, 2020 at 6:44 pm
There is one fan and it’s used to lower both temperature and humidity if either
rise above a certain threshold. More precisely, a simple conditional modulates
the fan to regulate temperature and a PID controller modulates the fan to
regulate vapor pressure deficit, which is based on both temperature and
humidity.

Reply

Karen Smith says:

October 19, 2020 at 5:40 am
Please disregard my first reply, what I meant was what if the temp and
humidity decreased at a certain threshold, does your system have a way
to increase these?
 Reply

Kyle Gabriel says:

November 26, 2020 at 11:28 am
Not included in the build log is the addition of a humidifier. You
could also add a heater. Any condition can be modulated for
which there is both a sensor that an measure it and an electrical
device or devices that can affect it.

Reply

Shreyas says:
November 19, 2020 at 12:38 am
Hey, Crazy read, one of the best automation experiments out there. I love doing these
experiments as well. One suggestion for an add in feature can be a ppfd/par sensor,
connected to an inexpensive hydraulic system/motorized system, that can lower and
rise the light according to the ppfd required each cycle of the plant. As for when the
plants grows in size, you can add an infrared sensor connected/glued to the ppfd
sensor which can be lowered and raised to be always above the plant. I love over
engineering stuff, automation is the way to go. The only thing left is a robotic butler to
cook my meals.

Reply

Doom says:
January 23, 2021 at 4:14 pm
instead of raising and lowering it might be better to dim and undim. less moving
parts.

dim and undim would also work better in situation where you have multiple
shelfs of a rack in a vertical situation.

as the plants grow taller.. you can dim.

or if you need it to be brighter you can turn on an additional lamp.

the camera feed can go to a simple cognitive vision module running on the PI

which calculates the distance of the light from the plant height.

when the plants are ready to move from veg to flowering, the robot can move
the plant to the other grow tent.

 sorry youll still have to cook for yourself :)

Reply

Kyle Gabriel says:

April 1, 2021 at 6:43 pm
Had my light been designed to dim, I would have. I do know of others
who have used Mycodo with methods to simulate sunrise and sunset as
well as adjusting light output based on photon measurements at canopy
of the growing plants.

Reply

Ali says:
November 22, 2020 at 4:50 pm
I am really impressed. I have been thinking about hydroponics for the last few months ,
and this project is exactly what I need. I am seriously considering support via patreon.
However, I have a question, will this work in regions with AC of 220~240 volts. The
“Iot Relay – Enclosed High-power Power Relay for Raspberry Pi” does not support the
mentioned voltage. Any thoughts about that. Thanks

Reply

Kyle Gabriel says:

November 22, 2020 at 4:53 pm
Ali, there are many relays available to choose from. You’ll have to select an
appropriate relay based on the control and switched voltages as well as the
anticipated load. There are a lot of pre-built relay modules that should also be
appropriate, as you are using a common AC voltage.

Reply

Kenny Campbell says:

November 25, 2020 at 4:34 pm
Will this work with other probes? I already have a ph and ec probe and ec/ph
transmitter.

Reply
 Kyle Gabriel says:
November 25, 2020 at 4:55 pm
Maybe. There are only a few that are currently supported, but users can create
their own custom Input modules. If you can provide a product link, I’ll check it
out and let you know if I think it can be easily integrated into Mycodo.

Reply

Samuel Chiang says:

December 2, 2020 at 3:20 am
I am working on my reef tank automation project. It’s very similar with your project. I
had built cabinet, glue glass, setup the PVC pipe, pH and temperature sensor box,
power control and meter box. I am going to build a dosing pump, auto water change
(Remove salt water, and using EC to determine add fresh water or salt water), even
Alkalinity meter by myself.

If you come to Taipei, welcome to my house to see my DIY projects.

Reply

Kyle Gabriel says:

January 10, 2021 at 1:19 pm
Excellent! Thank you.

Reply

Cody says:
December 6, 2020 at 5:32 pm
Do you have a video on how you built Mycodo? Or know of any resources that would
tech you how to build something similar to it? I think that would be very interesting to
watch for those wanting to create their own web based UI.

Reply

Kyle Gabriel says:

January 10, 2021 at 1:18 pm
 I don’t, but combining Flask and Bootstrap isn’t terribly difficult. I’ve had a lot of
fun learning and building Mycodo’s UI using these,and there’s a lot of
documentation out there on both.

Reply

joshua castro says:

December 13, 2020 at 10:40 am
you’re a beast!.. one of the best diy, congrats and thank you for the open source..
starting to build one.

Reply

Brad Viernow says:

December 20, 2020 at 1:18 am
Hello and great job! I am starting to purchase all of the components for this project. Do
you know of any less expensive EC probes that would work?

Reply

Kyle Gabriel says:

December 20, 2020 at 8:08 am
Thank you. Unfortunately, quality EC sensors are rare. If you happen to find one
start a topic on the forum and we can discuss it. kylegabriel.com/forum

Reply

Miguel says:
December 21, 2020 at 5:38 pm
Hello Kyle, great job, the pH and conductivity sensor can be submerged all the time?
and the other query if the tank is of greater capacity does not affect that you have a
smaller sample from the total values for example in a 2000 liter tank?

Reply

Kyle Gabriel says:

January 10, 2021 at 1:15 pm
 The water sensor probes are lab grade and can be submerged indefinitely. The
sample size for measuring water parameters can remain the same, as this is
simply a subsamples of the entire volume to determine if the chemicals have
adequately mixed. The subsample serves to separate the sensors from the main
reservoir, which is more susceptible to concentration gradients. The sampling
chamber draws a water sample from the same line that discharges into the
channels, after the pump, which is located on the opposite side of the reservoir
where chemicals are added.

Reply

Sung Shin says:

December 22, 2020 at 1:10 am
What a great job!

I’ve got lots of information and help from your article, and thanks a lot for your hard
work.

Reply

dan says:
January 2, 2021 at 3:26 am
How long did it take for you to finish the project?

Reply

Kyle Gabriel says:

January 10, 2021 at 1:11 pm
I worked on this for several months. I conceived it in December, began working
on it in January, and was finished with everything (building, demo grows,
filming, production, article) by June.

Reply

JeanNoumsi says:
January 3, 2021 at 11:33 pm
Hi Kyle,

for the grow light, I’m not quite sure why you referenced us to 1000W grow light Led
for a 4×2′ grow tent? I see you used a 300W led initially. And correct me if I am wong,
we should go for 32W per foot meter square, right? and if that is the case, then we are
only growing our plant on 24”x23” , which should be a 2×2′. So, 4 multiply by 32W
should give us 128W led needed. Can you please share with me your reference for
choosing your LED ?

thanks!

Reply

Kyle Gabriel says:

January 10, 2021 at 1:09 pm
The inverse-square law states for a point light source, every doubling of
distance reduces light intensity to 1/4. Intensity loss is inevitable, and I found
my 300W light was insufficient to cover the entire grow space. Regardless, I
mention that I would not be investing much time discussing lighting and this
should be researched on your own to determine your specific lighting needs,
therefore the light I included in the BOM was just a good light with decent
output that happened to be the same price that I bought my 300W light for
several years ago, so I thought it was a good deal.

Reply

Antonio Iantosca says:

January 9, 2021 at 5:03 pm
Nice work….going to try this project. Thank you

Reply

Juan Suero says:

January 10, 2021 at 12:30 pm
nice.

the thing with the frame and using the Maker Pipe connectors i dont understand is…
how do you get that slight angle.

i took a look at their website and it seems the only angle possible is 45 degrees.

i guess what i have to do is make a slightly higher hole on one side, use a T connector
for the side pipes that will hold the inlet

then use another T connector slightly rotated downward with a pipe headed towards
the drain outlet. ok i think i got it.
Reply

Kyle Gabriel says:

January 10, 2021 at 1:03 pm
I think I can picture what you’re describing. This is how I would do it (although
there are probably multiple methods): with 4 pipes vertical as legs, two will be
designated as the inlet side and two as the outlet side. T adapters will connect a
bar that spans across the inlet pipes and anither T and bar for the outlet pipes.
With the Ts, these bars can be moved up and down the vertical pipes, allowing
you to adjust the height and tightening the T to lock it in place.

Reply

Victor M. Romero II says:

January 11, 2021 at 6:36 pm
Hi this is an awesome read. I am also looking at trying to develop an automated
hydroponic system as a special project during these times where mobility and
recreation is limited. As I am new to all these things, may I ask if it is possible to
implement this project using Raspberry Pi 2 model B? What changes need to be done?

Reply

Kyle Gabriel says:

January 11, 2021 at 6:57 pm
Mycodo runs on all Pi versions, though it will be a bit slower on a Pi 2.

Reply

Suraj says:
January 12, 2021 at 6:07 am
Hi Kyle, Great work mate!!! Thanks.

I had a question. I’m from India here most of the kits and electronic circuits which you
mentioned to connect with the Raspberry pie are not available. Is it possible to run the
same process and integrate it with locally available circuitries and chips .

Thanks,

Reply
 Kyle Gabriel says:
January 12, 2021 at 1:35 pm
Yes. There are a number of sensors and devices with built-in support, but there
is also the ability to import custom user-created Inputs, Outputs, and Functions
to add support for previously-unsupported hardware a d add new functionality.

Reply

Jackson Spilsbury says:

January 12, 2021 at 1:24 pm
Hey Kyle, after forwarding my raspberrypi IP over port 443 I am getting SSL certificate
error, basically “this connection is not private”. Any idea how to fix this?

I know I can circumvent and visit the site anyways, but would like for others to be able
to access the site without having to do this workaround. I’ve also tried disabling the
“force https” option in the general mycodo configuration page, then forwarding port
80 (http) instead of 443, but the IP is still automatically redirected to HTTPS… any
pointers would be greatly appreciated. Thanks!

Reply

Kyle Gabriel says:

January 12, 2021 at 1:33 pm
I might have forced HTTPS with Flask-SSLify and forgot to remove the option in
the config to disable SSL. I’ll look into that. If you’re using a domain (such as a
dynamic DNS service), you can always get a free SSL certificate from
letsencrypt and overwrite the files in your SSL directory of Mycodo. I’ve done
this for some of my Mycodo instances that I shared to the public and didn’t
want to scare anyone.

Reply

Jackson Spilsbury says:

January 14, 2021 at 11:08 am
That did the trick! For Dynamic DNS I wrote a bash script scheduled
with crontab to keep a type A DNS zone record updated with the
current public IP. For the SSL cert I used letsencrypt, and completed the
 manual DNS challenge, verifying control over the hosted zone by
inserting the key into a TXT DNS record. Thanks Kyle.

Reply

Antonio Eggberg says:

January 12, 2021 at 8:58 pm
Hi Kyle,

This is super cool! Many thanks for sharing. I want to build this on my own.

I think the hardware setup – I need to get extra help.

May I ask – How long did it took you to do the actual setup? – (excluding all
github/programming part of the work).

Are you still running the system as is? or have you made further improvement?

Cheers

Reply

Kyle Gabriel says:

January 12, 2021 at 9:01 pm
You could probably build and set it all up in a weekend. My system is exactly as I
described it in the article.

Reply

Umesh Satavase says:

January 23, 2021 at 10:04 pm
Kyle

Awesome work and dedication. I have already started using your

mycodo on my rpi4. Can you enhance your code to include inputs from
UI to control instruments. I am looking for controlling lights in some
circumstances from website.

Reply

James says:
January 18, 2021 at 6:34 am
Hello Kyle,
Love the project and the detail you have put into it.

Am I right in thinking that as you are measuring and adding the nutrients etc as
required that are needed that you do not need to flush the system on a regular basis?
Was there ever a need to flush the system?

Did you dilute the liquid solution which is in the 4 bottles or is that as purchased?

Kind regards

James

Reply

Kyle Gabriel says:

April 1, 2021 at 6:47 pm
Hi, James. Yes, pH solutions and A/B nutrient solutions are only added when
needed and they are undiluted in the bottles. Having such tight control of
nutrient/pH solutions means there’s very little excess being used. I have rarely
needed to flush the tank. It’s somewhat difficult to determine if any particular
nutrient is becoming excessive in the system over time, so refreshing the
reservoir every so often may be beneficial.

Reply

Olga says:
January 20, 2021 at 1:36 pm
Hi Kyle! Great project! Is it possible to buy your system or maybe you can recommend
some videos to learn the basics?

Reply

Kyle Gabriel says:

February 12, 2021 at 11:12 am
Sorry, I don’t sell any kits. Here’s a good introductory video on the Raspberry Pi
and sensors: https://youtu.be/gnE4v-PcYKQ

Reply

nicola macchitella says:

January 26, 2021 at 8:14 pm
Hi Kyle, this project is sick! Really inspirational!

I want to try to replicate what you did but I don’t want to buy ALL the parts at once
mostly for economical reasons. I am starting by buying a peristaltic pump (besides the
raspberry pi). Do you think mycodo would work even if I only have this one widget
connected?

Also, do you think you could point me at the python code you are using in mycodo to
control the peristaltic pump? I hoping I could try to re-use that in a stand-alone script I
can write and run through the terminal only.

Best,

Nicola

PS. Really, best project I have seen on the web ever.

Reply

Kyle Gabriel says:

January 26, 2021 at 8:19 pm
Nicola, yes, Mycodo can operate a single pump using one of the several built-in
outputs for controlling pumps.

Reply

Leandro says:
February 7, 2021 at 10:11 am
Hi Kyle,

great article!

I had see at Atlas website that they have the tentacle t3 for raspberry, is that work
with Mycodo?

How many sensors we need at all?

Mycodo can talk with MQTT broker?

Looking forward your kindly reply.

Many thanks.

Leandro

Reply
 Kyle Gabriel says:
April 1, 2021 at 6:40 pm
The tentacle t3 merely connects all like connections and provides pins to easily
connect to each sensor’s power, ground, and communication pins, so yes it will
work with Mycodo.

Mycodo can publish (Output) and subscribe (Input) to MQTT topics.

Reply

Juan Suero says:

February 11, 2021 at 10:45 am
thanks. just one small but dangerous? point. im looking at what you actually wired at
7:36 and it doesnt jive with your schematic at https://kylegabriel.com/projects/wp-
content/uploads/sites/3/2020/01/Power-Control-Box-wiring-01.png in the schematic
you have the incoming power from the IEC 320 Panel Mount Power Socket ( L )
denoted by the red lines and they generally wire up to the extreme left screw
terminals of both relays. whereas power going from the relays leaves through the
extreme right screw terminals. but in the video you use the middle screw terminal. you
dont use the middle screw terminals in your schematic though. i am afraid to continue
because of this discrepancy.

Reply

Kyle Gabriel says:

February 11, 2021 at 2:19 pm
You’re correct. My schematic is inaccurate and the first wire should be
connected to the center terminal. The second wire is correct and connected to
the normally open terminal. I’ll update the figure.

Reply

Robert Stebler says:

February 12, 2021 at 10:52 am
Amazing setup! I’ve been working on this build for an ebb and flow hydroponic system
for the last month and I’m almost done. I added an air pump, water heater and reverse
osmosis water system to the build. Too much fun!!
I’ve been really surprised on how much time I’ve sunk into the reverse osmosis build
and the issues I’ve run into there. I’ve found that automated RO water with a float
valve that tops off the reservoir has a lot of TDS issues due to TDS creep. The RO
system needs to work in large batches at once so that it has time to flush the high TDS
water sitting in the membrane and continue to produce low TDS water. When it’s just
topping off the reservoir with small amounts of water then it’s topping off with the
high TDS water sitting in the RO membrane. Also made me wonder about home
systems with a pressure tank. After the first clean fill they start topping the tank with
higher TDS water. I just ordered an auto flush flow valve that will flush the RO
membrane everytime it gets used and also flush it every hour. Hopefully that takes
care of things. I have high TDS well water at 400 to 450 ppm. After RO it’s around 20.

Keep up the amazing work!! I’m following the Mushroom build on Patreon and will be
doing that in the future as well.

Reply

Robert Stebler says:

February 12, 2021 at 10:54 am
I ran into that issue as well but through trial and error found the middle screw terminal
worked.

Reply
Pingback: Phil on growing chili peppers with hydroponics | Craft Hot Sauce Podcast

Robert says:
February 18, 2021 at 5:12 pm
Hello! First off, must say im really impressed with the build. It really got me interested
in Hydroponics!

Im trying to follow the guide and setup Mycodo now, but i cant find the ezo-pmp when
following the guide.

I have followed it up to List_addr but sometimes i get a list of numbers, but no pump
showing, sometimes i get no list and just “Enter Command” again.

The pump are set to I2C and showing blue led. Am i missing something?

Thanks for answer!

Sincerly,
Reply

Kyle Gabriel says:

April 1, 2021 at 6:37 pm
Hi, Robert. First, enable I2C with “sudo raspi-config”, then reboot. After
changing the pump to I2C mode, connect 3.3 volts, ground, SDA, and SCL from
the Pi to the Pump. Then run “sudo i2cdetect -y 1” to scan your I2C bus 1 for
devices. You should see a device listed on the table of possible device
addresses. You can then disconnect your pump and rerun the i2cdetect
command and you should then see the address has disappeared.

Reply

Alex Wormuth says:

February 20, 2021 at 8:31 pm
Hey Kyle, thank you for documenting all of this! Would it be possible to use
https://atlas-scientific.com/kits/wi-fi-hydroponics-kit/ instead of buying the individual
probes?

Reply

Kyle Gabriel says:

April 1, 2021 at 6:32 pm
Yes, but you would still be connecting the Pi to the board all of the sensors
connect to anyway, so you’re merely adding another piece of equipment to
hold/organize the EZO sensor boards. You could use the wifi kit to upload to the
cloud, then use MQTT to then subscribe to the remote server receiving the
measurements, but you’re then adding several more components that could fail
(router, internet service provider, remote server, etc.). It’s far more reliable to
directly connect your Pi to the sensors.

Reply

Michael says:
February 23, 2021 at 1:00 pm
I don’t have much knowledge of EE or Coding, but I would really like to try this project
out. This project is awesome! Thank you so much for the through explanations. Only
issue is that the project isn’t quite budget friendly. I did the math and it seems like it
will be more than $1000 with the sensors and the pumps costing more than half. Are
there more budget friendly products out there that can give similar results?

Thank you

Reply

Kyle Gabriel says:

April 1, 2021 at 6:30 pm
Thank you. Unfortunately, for reliable water sensing, you will need to invest in
quality sensors. You can save money by using generic peristaltic pumps, but you
will need to also connect them to relays for Mycodo to be able to control them.

Reply

Federico Cordoba says:

March 16, 2021 at 11:09 am
Great job Kyle! I know Mycodo can handle few inputs, but what if I want it to handle
two or more zones? For example, if I want to control an enclosure for tomatoes and
another one for lettuces? Both enclosures with their own sensors but different light,
nutrient and stages…

Reply

Kyle Gabriel says:

March 16, 2021 at 11:12 am
Absolutely. You can create multiple functions that utilize different inputs and
outputs to regulate different environments.

Reply

Achmed says:
March 17, 2021 at 2:47 am
Super wonderful project Kyle … I am really happy found your works.

I see your project has many function to actuate many I/O peripherals. But i wonder
how to automate those functions regarding with time and specific purpose, schedule
and parameters. Is there any local or global variables or addressing or function
paramaters which can set to these requirements maybe with reading scheduled and
parameters files ?

Reply

Kyle Gabriel says:

April 1, 2021 at 6:22 pm
There are many ways to automate with Mycodo. There are timers (daily, time
point, duration, etc.), methods (changing variable over time), built-in functions,
and the ability to create and import your own functions. Mycodo is built on
Python, so you can incorporate any Python code and libraries into your
functions. You also have access to store variables in the Mycodo database or
temporally while the function controller is running. There is also the ability to
create and import your own Inputs and Outputs.

Reply

Jackson Spilsbury says:

March 17, 2021 at 7:25 pm
Hey Kyle question for you about temperature control in combination with a VPD PID
controller. I have a great VPD PID operating now with LED lights on 25% output, the
temperature never increases past ~72 degrees F. My objective is to increase the light
output to 50%, maintain VPD range between 800-1100, while also limiting the
temperature to 72 degrees F. However, temps are now up to 80 degrees with lights
turned up to 50%. I have tried creating a temperature PID leveraging the same exhaust
system (to lower only), but the two PID controllers seem to conflict with each other
(probably because they use one of the same outputs) and I end up breaching both of
the optimal zones. Any ideas how to regulate the VPD while also setting a max on
temp, through the Mycodo program?

Reply

Kyle Gabriel says:

March 17, 2021 at 7:31 pm
What you describe is a situation where software alone cannot solve. It is
unrealistic to both exhaust and raise humidity unless you humidify the air
coming in. It’s far easier to find a way to only exhaust the hot air from the light
 (blast exhaust) or introduce a cooling device, than to try to work with only
exhaust to lower temperature while keeping humidity high.

Reply

Jerome Hermoso says:

March 21, 2021 at 10:09 am
Hi i love the setup! i was wondering if this setup works on tomatoes. was planning to
build this for my thesis.

Reply

Kyle Gabriel says:

March 21, 2021 at 10:11 am
Thanks. Tomatoes grow better in a flood and drain system. Mycodo can be set
up to periodically operate a pump to do this rather than constantly running a
pump as is done in an NFT system.

Reply

John says:
March 27, 2021 at 9:05 pm
I was wondering if you had returned to consider adding a solenoid valve to
automatically open the flow of a fresh water source to the reservoir when the level
gets low?

Reply

Kyle Gabriel says:

April 1, 2021 at 6:19 pm
Had this been a larger reservoir or if there was easy access to a water supply
line, I would have done this.

Reply

Ken Tew says:

March 31, 2021 at 11:07 am
Thanks for posting this. It was fun to read and fun to attempt on my own too. I have
gotten first phase rolling. I have a seed tray sitting on top of downspouts right now.
https://www.dropbox.com/s/h2n9fmo0xf75zs6/SeedTray-Mycodo.png?dl=0

Reply

Gareth says:
April 9, 2021 at 8:13 am
Love what you’ve done here and am attempting to build a system like yours for CV
control of plant stress.

I have DFRobot EC and pH sensors (SKU:DFR0300 and SKU:SEN0161-V2) and the

RPi AD/DA High Precision Board with the ADS1256. I was wondering if you know if
the two would be compatible with your mycodo system. I see there is support for the
ADS1256 but not the sensors and before I attempt to make a custom input module
(I’m not the strongest coder), I would like to check with you.

Thanks for any advice in advance.

Reply

Kyle Gabriel says:

April 9, 2021 at 8:21 am
It should be possible. Check out https://forum.kylegabriel.com/t/atlas-
scientific-ph-sensor-analog/73 and come on over to the forum to start a
discussion.

Reply

Tri Vo says:
April 11, 2021 at 12:32 pm
You are a fucking legend. You even leave a simplified, step-by-step manual on how to
achieve this set-up. Someone please nominate this man for a Nobel, please! Thank you,
you scholar and gentleman.

Best,

Tri

Reply
 Lucas Villarino says:
April 23, 2021 at 1:01 pm
Hi, Kyle I really love your work, I think that is awesome!

I would like to know if instead of using the “Raspberry Pi 4 single board computer”,
could I use an Arduino board? I don’t know if the programming language changes but I
really want to know to try it in an Automated Aquaponic System.

Regards.

LV

Reply

Kyle Gabriel says:

April 23, 2021 at 1:22 pm
Hi Lucas. Thanks. Arduinos aren’t compatible, they have a different CPU
architecture and can’t support running Linux like a SBC such as a Raspberry Pi
can.

Reply

Markus Ganz says:

April 24, 2021 at 2:55 pm
Hey Kyle,

I also have just to say that you did awesome work. I will try to apapt your solution for
my own doings. Just one question. What screws did you use for mounting the
hardware components to the HDPE panel? Do m3 screws fit?

Reply

Kyle Gabriel says:

April 24, 2021 at 3:00 pm
Thanks. M3 are a bit too large for many PCB holes. I used the #1 pan head
screws in the materials list.

Reply
 Peter Beck says:
May 5, 2021 at 10:05 pm
Howdy Kyle!

I am Korean American living currently living in the Philippines, and have been studying
commercial feasibility of the Hydroponic business here, because during dry season
April-June where temp. goes upto 37-38, then June to September it turns to a wet
season and many many typhoons. Thank you so much for posting your wonderful
work, and I have been studying your work and the system, and could adapt it to my
system without spending a fortune on buying from Holland or Chinese (all suppliers
kinda look shady to me). Thank you again.

Reply

Rai Bur says:

May 14, 2021 at 4:26 am
Hi Kyle,

this is truly inspiring, congrats for the whole project. Was really excited about seeing
that on Youtube but now reading it its even better.

All the bests,

Rai

Reply

Axel says:
June 23, 2021 at 3:14 pm
Really cool been looking for something like this, or try to create something like this.
You did a great job I’ve only hooked up pH sensor and created a couple widgets but I’m
super excited spending lots of time with Mycodo. Thanks. I plan on running your pH
and EC code. It looks like I edit the first few lines to identify my devices but the rest of
the code is all there not to be changed unless you know what you’re doing.

Reply

Max Forbes says:

July 23, 2021 at 4:21 pm
Hey Kyle,

Holy cow, mad respect. This project is incredible… even just bridging the
hardware/software(/bioware) domains, but growing *food* too?! Also, I love those
circular meters in the dashboard.

Just wanted to send a tiny note that I wrote a quick tutorial for how you might improve
this page’s layout on mobile! tl;dr: without forcing word breaks in the comments,
pasting URLs can stretch the whole page! In case you’re interested, the walkthrough is
here:

https://maxwellforbes.com/posts/fixing-mobile-page-layouts#hydroponic

Cheers,

Max

Reply

Kyle Gabriel says:

July 25, 2021 at 9:29 am
Thanks, Max. I applied the fix. Normally I’m pretty good about fixing display
issues like this. I appreciate the help. Great article, too. I found some issue on it
though: “…stylistic issues on these ages for the sake…”, “…our extra wide
content has scroll bars them?”. Now we’ve both helped each other out to
improve our work!

Cheers.

Reply

Nick NY says:
July 27, 2021 at 12:16 am
Hate to sound like a broken record, but wow!! This is by far the most complete and
polished tutorial on this subject by a super large margin. Others may have started
down this path but I have yet to see full-blown nutrient automation done so well and
so detailed.

Thank you for taking the time to not only build something amazing but to share it in
this way, very much appreciated.

Reply
 Norbert says:
August 11, 2021 at 12:30 pm
Hi Kyle, kudos for this amazing system and the very well written overview!

Have you considered adding an automated light dimmer to mimic natural sun
behavior? I’ve heard this is a growing trend and that all plants really appreciate it.

Thanks!

Reply

Kyle Gabriel says:

August 11, 2021 at 12:59 pm
Thank you, and yes, I’ve considered it. The software supports the ability to do it
(and many others have done it), but the light I had on hand did not have the
ability.

Reply

Michael Johnston says:

August 18, 2021 at 11:53 pm
Hi Kyle, awesome design, I can’t wait to begin the build. Before I do anything this
extensive, I look over all the schematics and images to see where I might run into
trouble. I noticed the Flow Meter Totalizer isn’t included in the schematic and the
analog to digital converter connection is ambiguous to me as a nub. Is there a more
detailed schematic with exact pin connections? I’m looking at all the details but having
a difficult time connecting the dots. Thanks!

Reply

Kyle Gabriel says:

September 4, 2021 at 10:14 am
Both the flow meter and ADC connect to the I2C bus like the other I2C devices.
There are only 4 wires, power, ground, SDA, and SCL.

Reply

Jackson says:
August 21, 2021 at 6:15 pm
Hey Kyle, is there a way to run the program off of the network? I would like to run it
continuously off a tv in the grow room through the raspberry pi hdmi port, but do not
want to have it constantly using the local network. Thanks again.

Reply

Kyle Gabriel says:

September 4, 2021 at 10:13 am
A network is not required; you can set up a WiFi access point and connect to
the web interface that way or connect a keyboard and mouse. A real time clock
is needed if the Pi is not connected to the internet for accurate time-keeping.

Reply

John Rounds says:

September 16, 2021 at 2:14 pm
Can you clarify the connection of the AC current sensor and conversion board you
have listed in the parts list?

Reply

Kyle Gabriel says:

September 16, 2021 at 2:16 pm
The connection depends on what current sensor you use. If using just the
transformer listed, you only need to connect it to the ADC to measure the
returned voltage. If using the DFRobot sensor, this needs to be connected to
the DFRobot board, which is then connected to the ADC to measure the
voltage.

Reply

John Rounds says:

September 17, 2021 at 12:17 pm
It is the DFRobot sensor, connected to its board. It is connecting the
DFRobot board to the ADC that I need and the Pi that I need
clarification on.
 Reply

Kyle Gabriel says:

September 17, 2021 at 12:21 pm
See the manufacturer’s information:
https://wiki.dfrobot.com/Gravity_Analog_AC_Current_Sensor__S
KU_SEN0211_

Reply

John Rounds says:

September 17, 2021 at 2:03 pm
I looked at that before asking for your help. Its all Arduino
and does not involve the digital to analog conversion
board.

Kyle Gabriel says:

September 17, 2021 at 2:33 pm
The ADC has sets of positive and negative pins that it
measures the voltage. You essentially connect ground to
the negative and the signal wire to positive.

iDJMic says:
September 18, 2021 at 6:06 pm
Amazing job! I setup like this truly comes down to what your time is worth, taking
away much manual tasks.

Reply

Liam says:
September 19, 2021 at 10:11 pm
Thanks Kyle this is an amazing project that you have shared with the world!

I was just wondering If I wanted to have several levels of lights (each controlled
separately) and different bed systems (NFT, Flood drain and deep water) would I need
to use several raspberry pi?

In addition I would like to add some of the mods you discussed including
dehumidification, DO with O2 pump and a CO2 monitor with pump.

Thanks again.

Reply

Kyle Gabriel says:

September 19, 2021 at 10:18 pm
Thanks! No, you can run dozens of Outputs, Inputs, and Functions on a single Pi.

Reply

sourav says:
December 23, 2021 at 1:57 pm
Thank you to make such a detail article of your work i really appreciate your work it is
very helpful to understand this topic.

Thank you so much.

Reply

Karel says:
January 19, 2022 at 4:54 pm
This Co2 sensor is useless. It’s not dual beam. So you need to expose it to outside air
periodically to calibrate it. Otherwise, it starts to decrease in sensitivity and shows
much lower readings from a few weeks. Ideally the sensor should be replaced with a
Sensirion SCD30.

Reply

Kyle Gabriel says:

January 19, 2022 at 6:45 pm
You’re incorrect. You should familiarize yourself with the MH-Z19B before
making such claims.

Reply

Robert says:
February 1, 2022 at 3:33 pm
Hi Kyle,

Love what you’ve done here. I’m planning on mimicking it in the near future. Quick
question though…..would it be suitable to run for the most part, as is, for a DWC setup?

Thanks

Reply

Kyle Gabriel says:

February 1, 2022 at 7:17 pm
Sure, you would just have to reconfigure the plumbing, but the principles still
apply for water circulation and nutrient/pH regulation.

Reply

Justin Gravett says:

February 28, 2022 at 6:52 pm
Hi Kyle, thanks for sharing such an awesome project! Not sure if this is the best place
to post about it, but I recently discovered something I just wanted to share in case
anyone else experiences the same issue. Our Raspberry Pi was having intermittent i2c
issues where “sudo i2cdetect -y 1” would sometimes not see all of the devices, while
other times it would. Obviously this caused issues running Mycodo as well and the
Python library provided my Atlas Scientific. Adding a 10K pull up on the SDA and SCL
i2c lines completely fixed the intermittent connection issues. We just had to include a
bidirectional level shifter as well to prevent 5V input to the Pi. Hope this helps if
anyone experiences a similar issue!

Reply

Noah Kramer says:

April 11, 2022 at 1:06 pm
Amazing, inspiring project. Thanks for putting all the documentation together. I’m
going to build a similar system for an undergrad research project, I’d like to
incorporate a reinforcement learning model into mycodo to experiment with
optimizing the inputs. Have you heard of anyone doing something similar with
mycodo? Curious if you have thought about doing something similar as well.

Reply
 Kyle Gabriel says:
April 11, 2022 at 10:50 pm
Thank you. I’ve used Mycodo to teach graduate research methods courses, but
these were only 3 session modules, so optimization wasn’t able to be explored.

Reply

Leave a Reply
Your email address will not be published. Required fields are marked *

Comment *

Name *

Email *

Website

POST COMMENT

Projects | Forum | About | Contact | Privacy

Copyright © 2005-2022 Kyle Gabriel