This repo contains a sample app to show how you can monitor CO2 levels using a Raspberry Pi and C#, and send the data to Azure Event Grids. It also shows how to build an alert device that listens to the CO2 readings and uses these to color an indicator light.
This repo contains 2 apps - a Sensor app, and a Monitor app. The sensor app detects CO2 levels and sends them over MQTT to an Azure Event Grid. The Monitor app listens to these value also over MQTT and lights LEDs depending on the CO2 level:
| Level | Color |
|---|---|
| <800 | Green |
| 800-1200 | Yellow |
| 1200-1800 | Orange |
| >1800 | Red |
The devices are authenticated using x.509 certificates.
This device is built around a Raspberry Pi. I'm using a Raspberry Pi 5 with 8GB of RAM, an active cooling fan and a 256GB SSD because I can, but you can run this on much lower power devices, anything from a Pi 2 onwards.
The CO2 sensor is a SCD41. I'm using the one from Seeed Studio with a Grove connector, and a Grove base hat. I also use a Blinkt from Pimoroni as an indicator. This is a strip of 8 RGB LEDs that plugs into the entire GPIO pin set.
The SCD41 uses I2C, so needs to be connected to one of the ports labelled I2C - there are 3 on the opposite side to the GPIO pins.
You can also get an SCD41 without the Grove connection. This can be connected to the relevant GPIO I2C pins.
The Raspberry Pi is running the latest Paspberry Pi OS. My apps are C# app, so it has .NET 8 installed. There is also some drivers for the Grove base hat to install that turns on the I2C interface.
To set up the Azure services, you can use the Azure CLI. This can also be installed on the Raspberry Pi if you want. The devices will connect to Azure using X.509 certificates, so if you don't already have one, you can use Step to create them. Step doesn't support Arm64, so you can't run it on a Raspberry Pi.
Finally, clone this repo to your Raspberry Pi, and open it in something like VS Code.
This setup uses an Azure Event Grid as the MQTT broker.
All Azure respources live in a resource group. Use the Azure CLI to create a resource group:
az group create --location westus3 \
--resource-group rg-co2-sensorThis will create a resource group called rg-co2-sensor in the West US 3 region. Use your own name and preferred location if you want.
To create the event grid:
az eventgrid namespace create --name <name> \
--resource-group rg-co2-sensor \
--topic-spaces-configuration "{state:Enabled}"Replace <name> with the name of your event grid. The name needs to be globally unique as this makes up part of the URL for the MQTT broker.
If you called your resource group something other than rg-co2-sensor, replace this too.
If you are asked to install the event grid extension, select Y to install this. This is needed to create event grids using the Azure CLI. The flag --topic-spaces-configuration "{state:Enabled}" will turn on the MQTT broker.
When this is created, the ourput will include a hostname for the MQTT server in the topicSpacesConfiguration section, as well as the location:
{
"topicSpacesConfiguration": {
"hostname": "<name>.<location>.ts.eventgrid.azure.net",
"maximumClientSessionsPerAuthenticationName": 1,
"maximumSessionExpiryInHours": 1,
"state": "Enabled"
},
}Make a note of this hostname. as you will need this to configure the apps.
Topic spaces are groupings of MQTT topics. By using spaces, you can manage access to groups of topics at once. To send and recieve data over MQTT, you need to define at least one topic space.
az eventgrid namespace topic-space create --name co2-topics \
--namespace-name <event grid name> \
--resource-group rg-co2-sensor \
--topic-templates ["co2-readings"]Set the --namespace-name to the name of your event grid namespace. If you called your resource group something other than rg-co2-sensor, replace this too.
This will create a topic space called co2-topics that allows you to publish and subscribe to co2-readings.
To authenticate the device, you will need an X.509 certificate. Each device can have its own cert, or can share a cert. In this case, to make things easier, they sensor and monitor apps will share a cert.
From another device (not your Raspberry Pi), make sure you have the Step CLI installed. This doesn't support ARM64, so cannot run on a Pi.
-
Create a root and intermediate certificate:
step ca init --deployment-type standalone \ --name MqttAppSamplesCA \ --dns localhost \ --address 127.0.0.1:443 \ --provisioner MqttAppSamplesCAProvisionerThis creates the certificate with the name
MqttAppSamplesCAand provisionerMqttAppSamplesCAProvisioner- you can change these if you want. You will be asked for a password. -
Once you have the root CA, generate a certificate:
step certificate create sensor-cert sensor-cert.pem sensor-cert.key --ca .step/certs/intermediate_ca.crt \ --ca-key .step/secrets/intermediate_ca_key \ --no-password \ --insecure \ --not-after 2400hYou will be asked for the password you set in the previous step.
-
Get the thumbprint for the cert:
step certificate fingerprint sensor-cert.pem
-
Copy the generated
sensor.pemandsensor.keyfiles to your Raspberry Pi. Put them in the root folder of this repo.
For a client, such as a sensor device, to connect, it needs to be enabled as a client on the event grid. This client is configured with the thumbprint from the certificate to authenticate it.
-
Create the sensor client:
az eventgrid namespace client create --namespace-name <event grid name> \ --resource-group rg-co2-sensor \ --name co2-sensor \ --client-certificate-authentication "{validationScheme:ThumbprintMatch,allowed-thumbprints:[<sensor thumbprint>]}"
Set the
--namespace-nameto the name of your event grid namespace. If you called your resource group something other thanrg-co2-sensor, replace this too.Replace
<sensor thumbprint>in the--client-certificate-authenticationvalue with the thumbprint for the certificate. This is an array of values, so you need the surrounding square brackets. -
Create the monitor client in the same way:
az eventgrid namespace client create --namespace-name <event grid name> \ --resource-group rg-co2-sensor \ --name co2-monitor \ --client-certificate-authentication "{validationScheme:ThumbprintMatch,allowed-thumbprints:[<sensor thumbprint>]}"
Set the
--namespace-nameto the name of your event grid namespace. If you called your resource group something other thanrg-co2-sensor, replace this too.Replace
<sensor thumbprint>in the--client-certificate-authenticationvalue with the thumbprint for the certificate, the same as in the previous step.
The last step is to grant permissions to clients to publish or subscribe on different topic spaces. Permissions are set on client groups, rather than on individual clients. There is a client group name called $all that includes all clients, so it's easier for a simple app like this to grant permission on this client group to publish and subscribe to the topic space.
Permission bindings have to be created for either publish or subscribe, so you will need to create 2.
-
Create the permission binding for publishing:
az eventgrid namespace permission-binding create --namespace-name <event grid name> \ --resource-group rg-co2-sensor \ --name c02-sensor-binding-publisher \ --client-group-name '$all' \ --permission publisher \ --topic-space-name co2-topics
-
Create the permission binding for subscribing:
az eventgrid namespace permission-binding create --namespace-name <event grid name> \ --resource-group rg-co2-sensor \ --name c02-sensor-binding-subscriber \ --client-group-name '$all' \ --permission subscriber \ --topic-space-name co2-topics
You will need to set some configuration in the sensor and monitor apps. These are in the src/Sensor, and src/Monitor folders.
-
Copy the
appsettings.json.examplefile toappsettings.jsonin each folder. -
In these files, set the following:
{ "Hostname": <the hostname of the MQTT server>, "Topic": "co2-readings", "X509Pem": "../../sensor-cert.pem", "X509Key": "../../sensor-cert.key", }Set
"Hostname"to the hostname of the MQTT server. This is thehostnamefrom the output of the event grid creation. The paths to thesensor-cert.pemandsensor-cert.keyfiles assume you placed these files in the root of where you cloned this repo. -
In the
appsettings.jsonfile in theSensorfolder, set the following:{ ... "ClientId": "co2-sensor", } -
In the
appsettings.jsonfile in theMonitorfolder, set the following:{ ... "ClientId": "co2-monitor", }
Once the appsettings.json files are configured, you can run the sensor and monitor.
-
To run the sensor:
cd src/Sensor dotnet run -
To run the monitor:
cd src/Monitor dotnet run
You will see the output from the sensor showing the CO2 values, and the output from the monitor reading these values. The LEDs will also light up to show the CO2 level.
You can show the CO2 levels changing by breathing on the sensor - your breath is high in CO2, so by gently breathing on the sensor, the CO2 levels will go up and the LEDs will change to red.