In this project, we're going to build a small weather-logging node using a Feather and a temperature sensor. The captured data will then be sent over LoRaWAN to The Things Network.

Wait, this sounds similar to the a previous guide we have. What's different?

Good catch - the difference is that this guide is for use with CircuitPython, meaning you can get your project up and running on the Things Network quicker than ever. Also, since we're using CircuitPython, this guide is compatible with Python Linux boards, like the Raspberry Pi family. 

What is The Things Network?

The Things Network is a project dedicated to building a network for the Internet of Things. While WiFi is used in most Internet of Things devices, The Things Network uses a protocol called LoRaWAN which allows devices to talk to the internet without cellular or WiFi connectivity. This means you don't need to worry about protected wireless hotspots, cellular data plans, or spotty WiFi connectivity. 

It's ideal for most internet of things projects, and unlike cellular plans or WiFi - it's free to use.

There are many gateways available to connect your CircuitPython device to - if you'd like to find a gateway in your area, check the Gateway Map.

Parts

Our Feather M0 RFM9x is an all-in-one Feather with an onboard RFM9x radio module cooked in, built-in USB, and battery charging (highly useful for deploying LoRa nodes). 

Already have a CircuitPython-Compatible board (or a Raspberry Pi)? There are multiple ways to connect - the all-in-one Feather M0 LoRa and the stack-able Radio FeatherWing:

We'll also use a Si7021 temperature and humidity sensor to log and record whether it's sweltering or freezing.

Materials

You'll want to pick these up from the Adafruit Store if you don't have them on-hand already:

If you're running a weather logger, you'll want to include a LiPo Battery in your build. LoRa has less power-draw than WiFi, so your node should be able to run unattended for a while.

There's quite a few ways to connect to The Things Network with CircuitPython. You can use a Feather RFM9x (an all-in-one), a breakout or FeatherWing and a CircuitPython-compatible board, or a Python Linux computer/board. 

Wiring with a Feather M0 RFM9x

Attaching an Antenna

Wiring

Wiring with a Radio Featherwing

Wiring a LoRa Radio Transceiver Breakout

Attaching an Antenna

Wiring a CircuitPython-Compatible Board with a LoRa Breakout

Python Computer Wiring with a LoRa Breakout

Since there's dozens of Linux computers/boards you can use we will show wiring for Raspberry Pi. For other platforms, please visit the guide for CircuitPython on Linux to see whether your platform is supported. 

Here's the Raspberry Pi wired with a RFM9x Breakout and a SI7021:

All wired up? Let's move on to setting up The Things Network for TinyLoRa.

Installing CircuitPython on a Feather RFM9x

Some of the CircuitPython compatible boards come with CircuitPython installed. Others are CircuitPython-ready, but need to have it installed. 

If you're using a Feather M0 RFM9x, you'll need to install CircuitPython on your board.

• If you're using a RFM9x breakout and a Feather Express or a Python Linux board, you can ignore this page and continue to the library installation.

If you have Arduino IDE downloaded and installed, this process is as simple as uploading a sketch to your board then dragging and dropping a file.

First, we need to flash the UF2 bootloader onto the Feather.

Navigate to the latest release page for the UF2 bootloader on the Adafruit GitHub.

Click update-bootloader-feather_m0-v2.x.x-adafruit.7.ino to download the file.

Double click the file to open it in the Arduino IDE.

The .ino sketch should download to your computer. Open the file in the Arduino IDE by double-clicking the file. The Arduino IDE should load and the sketch should appear. The sketch will have a lot of hexadecimal numbers and might look daunting if you scroll down - don't worry, we're not touching any of it.

Download the latest CircuitPython version

Navigate to the CircuitPython release page and select the latest build for the Feather M0 RFM9x. 

Next, we're going to load CircuitPython (UF2 file) onto our board.

Start the UF2 Bootloader

Now find the UF2 file you downloaded. Drag that file to the FEATHERBOOT drive on your computer.

The lights should flash again, FEATHERBOOT will disappear and a new drive will show up on your computer called CIRCUITPY.

Congratulations! You've successfully installed CircuitPython!

Bundle Installation

We're constantly building/updating/improving our libraries, so we don't ship all the CircuitPython libraries with your board.

You'll want to grab a copy of the CircuitPython Library Bundle - a .zip file containing a bunch of useful CircuitPython libraries.  The TinyLoRa library is separate repository, but it's easier to install everything at once using the bundle.

Grab the latest version of the library bundle by clicking this button and make sure you download the version that matches your CircuitPython library version:

Unzip the library. You'll see a two folders: lib and examples. Double click the lib folder to see a listing of all the CircuitPython libraries. 

If you're using an Express board

On Express boards (the board will say 'EXPRESS' on it), the lib and examples directories can be copied directly to the CIRCUITPY drive.

Just drag the entire lib and examples (optional) folders into the CIRCUITPY drive, and 'replace' any old files if your operating system prompts you.

If you're using a non-Express board

The RFM9x Feather M0 (all-in-one) has less space available compared to express boards. To save space, we're only going to install the libraries required for this guide. 

From the bundle folder, navigate to the lib folder and find the required folders: adafruit_bus_device, adafruit_tinylora and adafruit_si7021.

Copy these three folders/files over to the CIRCUITPY drive's lib directory.

Before continuing make sure your board's lib folder has at the adafruit_tinylora, adafruit_bus_device, and adafruit_si7021 libraries copied.

Installation on Python Linux Computers

If you're following along with a Raspberry Pi, Beaglebone or any other supported small Linux computer, we'll use a special library called adafruit_blinka (named after Blinka, the CircuitPython mascot) to provide the layer that translates the CircuitPython hardware API to whatever library the Linux board provides. It's CircuitPython, on Pi!

If you haven't set up Blinka and the Adafruit IO Python Library yet on your Raspberry Pi, follow our guide:

• Blinka and Adafruit IO Setup

Enable I2C

We use two pins on the Pi (SDA/SCL) to communicate over I2C with the Si7021. You only have to do this step once per Raspberry Pi, the I2C interface is disabled by default.

• Enabling I2C

Once you're done with this and have rebooted, verify you have the I2C devices with the command:

sudo i2cdetect -y 1

The output from running this command should look like the following:

Enable SPI

You'll also need to enable SPI to communicate over SPI with the RFM breakout. You only have to do this step once per Raspberry Pi, the SPI interface is disabled by default.

• Enabling SPI

Once you're done enabling SPI and have rebooted, verify you have the SPI interface by typing the following command into the terminal:

ls -l /dev/spidev*

You should see the following output:

Installing required CircuitPython Libraries

Next, you'll need to install libraries to communicate with the RFM9x breakout and the Si7021. Since we're using Adafruit Blinka (CircuitPython), we can install CircuitPython libraries straight to our Raspberry Pi, instead of installing from separate files. 

Run the following command from your terminal to install the Adafruit_CircuitPython_Si7021 Library:

sudo pip3 install adafruit-circuitpython-si7021

Then, install the Adafruit_CircuitPython_TinyLoRa library

sudo pip3 install adafruit-circuitpython-tinylora

Writing code With Mu

Adafruit recommends the Mu editor to type your code into and interact with your project. You can find out how to install Mu on mac, PC, and Linux in this guide page on the Adafruit Learning System.

The advantage to using Mu is it has an editor and added features like direct save to the board's REPL interactive command line and print output. Mu can even plot values for you.

The Code

The code below will use the adafruit_tinylora library which provides high-level access to your radio transceiver's features. 

The code for sending data using a Si7021 sensor to a Things Network Gateway is shown below:

# SPDX-FileCopyrightText: 2021 ladyada for Adafruit Industries
# SPDX-License-Identifier: MIT

"""Using TinyLoRa with a Si7021 Sensor."""

import time

import adafruit_si7021
import board
import busio
import digitalio

from adafruit_tinylora.adafruit_tinylora import TTN, TinyLoRa

try:  # typing
    from typing import Annotated, TypeAlias

    bytearray4: TypeAlias = Annotated[bytearray, 4]
    bytearray16: TypeAlias = Annotated[bytearray, 16]
except ImportError:
    pass

# Board LED
led = digitalio.DigitalInOut(board.D13)
led.direction = digitalio.Direction.OUTPUT

# Create library object using our bus i2c port for si7021
i2c = busio.I2C(board.SCL, board.SDA)
sensor = adafruit_si7021.SI7021(i2c)

# Create library object using our bus SPI port for radio
spi = busio.SPI(board.SCK, MOSI=board.MOSI, MISO=board.MISO)

# RFM9x Breakout Pinouts
cs = digitalio.DigitalInOut(board.D5)
irq = digitalio.DigitalInOut(board.D6)
rst = digitalio.DigitalInOut(board.D4)

# Feather M0 RFM9x Pinouts
# cs = digitalio.DigitalInOut(board.RFM9X_CS)
# irq = digitalio.DigitalInOut(board.RFM9X_D0)
# rst = digitalio.DigitalInOut(board.RFM9X_RST)

# TTN Device Address, 4 Bytes, MSB
devaddr: bytearray4 = bytearray([0x00, 0x00, 0x00, 0x00])

# TTN Network Key, 16 Bytes, MSB
nwkey: bytearray16 = bytearray(
    [
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
    ]
)

# TTN Application Key, 16 Bytess, MSB
app: bytearray16 = bytearray(
    [
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
    ]
)

ttn_config = TTN(devaddr, nwkey, app, country="US")

lora = TinyLoRa(spi, cs, irq, rst, ttn_config)

# Data Packet to send to TTN
data: bytearray4 = bytearray(4)

while True:
    temp_val = sensor.temperature
    humid_val = sensor.relative_humidity
    print(f"Temperature: {temp_val:0.2f} C")
    print(f"relative humidity: {humid_val:0.1f} %")

    # Encode float as int
    temp_val = int(temp_val * 100)
    humid_val = int(humid_val * 100)

    # Encode payload as bytes
    data[0] = (temp_val >> 8) & 0xFF
    data[1] = temp_val & 0xFF
    data[2] = (humid_val >> 8) & 0xFF
    data[3] = humid_val & 0xFF

    # Send data packet
    print("Sending packet...")
    lora.send_data(data, len(data), lora.frame_counter)
    print("Packet Sent!")
    led.value = True
    lora.frame_counter += 1
    time.sleep(2)
    led.value = False

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121

# SPDX-FileCopyrightText: 2021 ladyada for Adafruit Industries
# SPDX-License-Identifier: MIT

"""Using TinyLoRa with a Si7021 Sensor."""

import time

import adafruit_si7021
import board
import busio
import digitalio

from adafruit_tinylora.adafruit_tinylora import TTN, TinyLoRa

try:  # typing
    from typing import Annotated, TypeAlias

    bytearray4: TypeAlias = Annotated[bytearray, 4]
    bytearray16: TypeAlias = Annotated[bytearray, 16]
except ImportError:
    pass

# Board LED
led = digitalio.DigitalInOut(board.D13)
led.direction = digitalio.Direction.OUTPUT

# Create library object using our bus i2c port for si7021
i2c = busio.I2C(board.SCL, board.SDA)
sensor = adafruit_si7021.SI7021(i2c)

# Create library object using our bus SPI port for radio
spi = busio.SPI(board.SCK, MOSI=board.MOSI, MISO=board.MISO)

# RFM9x Breakout Pinouts
cs = digitalio.DigitalInOut(board.D5)
irq = digitalio.DigitalInOut(board.D6)
rst = digitalio.DigitalInOut(board.D4)

# Feather M0 RFM9x Pinouts
# cs = digitalio.DigitalInOut(board.RFM9X_CS)
# irq = digitalio.DigitalInOut(board.RFM9X_D0)
# rst = digitalio.DigitalInOut(board.RFM9X_RST)

# TTN Device Address, 4 Bytes, MSB
devaddr: bytearray4 = bytearray([0x00, 0x00, 0x00, 0x00])

# TTN Network Key, 16 Bytes, MSB
nwkey: bytearray16 = bytearray(
    [
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
    ]
)

# TTN Application Key, 16 Bytess, MSB
app: bytearray16 = bytearray(
    [
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
        0x00,
    ]
)

ttn_config = TTN(devaddr, nwkey, app, country="US")

lora = TinyLoRa(spi, cs, irq, rst, ttn_config)

# Data Packet to send to TTN
data: bytearray4 = bytearray(4)

while True:
    temp_val = sensor.temperature
    humid_val = sensor.relative_humidity
    print(f"Temperature: {temp_val:0.2f} C")
    print(f"relative humidity: {humid_val:0.1f} %")

    # Encode float as int
    temp_val = int(temp_val * 100)
    humid_val = int(humid_val * 100)

    # Encode payload as bytes
    data[0] = (temp_val >> 8) & 0xFF
    data[1] = temp_val & 0xFF
    data[2] = (humid_val >> 8) & 0xFF
    data[3] = humid_val & 0xFF

    # Send data packet
    print("Sending packet...")
    lora.send_data(data, len(data), lora.frame_counter)
    print("Packet Sent!")
    led.value = True
    lora.frame_counter += 1
    time.sleep(2)
    led.value = False

Using Mu (or a text editor), copy this code and save it to your computer, naming it code.py.

Note: The code is configured to use a RFM9x breakout by default. If you are using a Feather M0 RFM9x, delete the following lines:

# RFM9x Breakout Pinouts

cs = digitalio.DigitalInOut(board.D5)

irq = digitalio.DigitalInOut(board.D6)

rst = digitalio.DigitalInOut(board.D4)

Then, uncomment (remove the #) the following lines to enable the RFM9x's builtin D0 and CS pins

Feather M0 RFM9x Pinouts

irq = digitalio.DigitalInOut(board.RFM9X_D0)

cs = digitalio.DigitalInOut(board.RFM9X_CS)

rst = digitalio.DigitalInOut(board.RFM9X_RST)

Installing Code on CircuitPython

Navigate to where you saved your code.py file. Copy (by dragging the file and dropping it) the file to the CIRCUITPY drive.  

While the code will appear to run, it is not yet set up for your Application or Device. We'll do that next.

Setting up the code for The Things Network

While we can send data to The Things Network, and our gateway might notice it, it isn't registered to a device yet. To register your device with The Things Network using ABP, you'll need to set three unique identifiers in the code.py file: the Network Session Key, the Device Address, and the Application Session Key.

Before adding the unique identifiers to our sketch, we'll need to first expand them by clicking the <> icon.

These are your keys. We're going to enter them into our code.py, but we need to be careful - the keys on the Things Network console use parentheses or curly braces { } instead of brackets [ ]. 

First, copy the Device Address from the TTN console to the devaddr variable in the code. 

Then, remove the braces { } from the device address.

A device address copied from The Things Network console would look like: { 0x26, 0x02, 0x1F, 0x07 }. In the code.py, it'd look like: devaddr = bytearray([0x26, 0x02, 0x1F, 0x07]).

Then, copy the Network Session Key from the TTN console to the NwkSkey variable in the code. Make sure to modify the code to remove the parentheses/curly braces { }. 

Finally, copy the Application Session Key from the TTN console to the AppSkey variable in the code. Make sure to modify the code to remove the parentheses/curly braces { }. 

That's all for now - we're ready to run our code! 

Running the Code using Mu Editor

Upon saving (ctrl/cmd +s) your code, the board will refresh. 

But where is the output? The serial monitor is hidden by default in Mu Editor.

In the Mu Editor, click the Serial button on the top icon-bar to bring up the Serial REPL.

You should see the temperature and humidity values printed to the console. The packet's status (sending, or timed out and unable to send) should appear shortly after. The LED on your Feather should also blink the onboard LED when it successfully sends a packet to The Things Network.

If you see Packet Sent!, rejoice - your packet has sent to the Things Network. But, while it's been sent, we want to check and make sure it's been received. 

If you're using a Raspberry Pi: The Pi does not have an onboard LED located at pin D13, so it will not blink when the packet is sent. In the code, comment out every line which uses the LED, like the following:

led = digitalio.DigitalInOut(board.D13) to # led = digitalio.DigitalInOut(board.D13)

Checking Data on The Things Network Console

We want to make sure the data has been received on the other end. To do this, we'll navigate to the The Things Network Console and select the application. From the menu on the right hand side of the page, Click Data.

Decoding the Payload

If you're sending packets in strange formats or encodings (like we are!), The Things Network Console has a programmable data decoder to decode the packets, and assign useful labels to the data.

Copy and paste the decoder script below into the decoder's integrated text editor and click save. 

Then, click the data tab. Next to the raw payload data, you should see the decoded data for humidity and temperature.

// TinyLoRa - Si7021 Decoder
function Decoder(bytes, port) {
  var decoded = {};

  // Decode bytes to int
  var celciusInt = (bytes[0] << 8) | bytes[1];
  var humidInt = (bytes[2] << 8) | bytes[3];
  
  // Decode int to float
  decoded.celcius = celciusInt / 100;
  decoded.humid = humidInt / 100;

  return decoded;
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14

// TinyLoRa - Si7021 Decoder
function Decoder(bytes, port) {
  var decoded = {};

  // Decode bytes to int
  var celciusInt = (bytes[0] << 8) | bytes[1];
  var humidInt = (bytes[2] << 8) | bytes[3];
  
  // Decode int to float
  decoded.celcius = celciusInt / 100;
  decoded.humid = humidInt / 100;

  return decoded;
}

I'm located in a region other than the United States

All the software modifications you'd need to make are within one line of code.py. The country defaults to the US (adafruit is located in this region), but you can change it to any of the four supported regions:

• US - USA, Canada, and South America
• EU - Europe
• AS - Asia
• AU - Australia

If you wanted to switch to an European frequency plan, you'd configure the ttn_config object with the following parameters:

ttn_config = TTN(devaddr, nwkey, app, country = 'EU')

We've supported a subset of the major frequency plans in the initial release of TinyLoRa for CircuitPython. If you don't see your region listed here, you can submit an issue and we'll add it to the library.

If you'd like to use a specific channel...

By default, TinyLoRa for CircuitPython broadcasts on multiple channels, randomly hopping between them. Since most gateways only listen for a transmission on a few channels at a time, this is a way to increase the probability of the gateway "seeing" the packet. 

You can specify a channel by feeding the LoRa object constructor a channel as a keyword argument. For example, if you'd like to send data on channel 3, the object would be created like the following:

lora = TinyLoRa(spi, cs, irq, ttn_config, channel=3)

Want to change channels during the event loop, between sending different data packets?  We included a set_channel function which will set up the radio to use the specified channel:

lora.set_channel(3)

If you'd like to use a specific data rate...

By default, TinyLoRa uses the datarate SF7BW125. That is, a spreading factor of 7 and a bandwidth of 125kHz. Similar to setting a channel, TinyLoRa can set the datarate given a specific bandwidth and spreading factor. If you'd like to set the transmission datarate to a spreading factor of 10 and a bandwidth of 125kHz, call set_datarate() like the following: 

lora.set_datarate("SF10BW125")

TinyLoRa has the following datarates available for use by this method:

• SF7BW125, SF7BW250, SF8BW125, SF9BW125, SF10BW125, SF11BW125, SF12BW125

Want to learn more about datarates and LoRa? Check out this blog post here..

My packet never successfully sends

If you're hitting an error, RuntimeError: Timeout during packet send , it's likely a gateway is not in range, or connected properly. Ensure that there are gateways near you and in-range (click here for a map of all the registered gateways).

If you're a Things Network Gateway operator, check from the Things Network Console if the gateway has been seen recently. If the Last Seen value is greater than a few minutes, reset your gateway. 

My project has disconnected/been reset and I don't see any data in my application

If you haven't disabled frame counter checks from within your device (instructions on how do this are on the TinyLoRa TTN Setup page), you'll encounter this issue when the device is reset/loses power.

Why would this happen? There's a mismatch between the frame counter on your device and the frame counter on The Things Network. Most likely, the frame counter on your device is reset to zero, and the frame counter on your device's console is set to the last known value. 

There are ways around this issue, such as using a different activation method such as OTAA (currently unsupported by TinyLoRa).

However, we can still reset both the node and the frame counter on the application.

What's the deal with the frame counter? Do I need it in my code?

Disabling frame counter checks allows you to transmit data to The Things Network without requiring a match between your device's frame counter and the console's frame counter.

If you're making a project and doing a lot of prototyping/iteration to the code, disabling these checks is okay as it'll let you reset the device and not re-register it to the application (it'll also prevent counter overflows).

If you're deploying a project, you'll want to re-activate the frame counter for security purposes. With the frame counter disabled, one could re-transmit the messages sent to TTN using a replay attack. 

The Things Network's documentation page has a full explanation about the role of the Frame Counter.