Manage cookies

We use cookies to improve our services. Read more about how we use cookies and how you can refuse them.

The necessary cookies help to make the website work properly.

Anonymous statistical cookies help to understand how visitors use the website.

  • Cloud Platforms, M5Stack, MicroPython
    12 min | 223308

    #M5Stack: Fresh air checker can help you to stay safe from #COVID-19

    Cloud Platforms, M5Stack, MicroPython | 12 min | 223308

    Winter is coming and the air-quality inside a closed room could be an index used to estimate the amount of aerosols in it, and hence the risk of Covid-19 infection. In other words, a system of CO2 traffic lights could help to spot potential corona-aerosols!

    Where there are many people in a room, there is a lot of exhalation (approx. eight liters of air per minute per nose). This air contains CO2 (carbon dioxide). Long before Covid-19, in Germany, the German Environment Agency suggested that a concentration of CO2 in a classroom should not exceed 1000 ppm (parts per million) (Guidelines for Indoor Air Hygiene in School Buildings - 2008). Therefore, every measurement over this threshold was/is considered "hygienically evident" (1000-2000 ppm) or "hygienically unacceptable" (>2000 ppm). For your reference, the fresh air outside contains a CO2 concentration between 350-500 ppm.

    The idea of setting up CO2 measuring devices during the Covid-19 pandemic is not about measuring such extreme values. A CO2 measuring device would be an orientation when the air in the room should be exchanged i.e. ventilated (with the windows wide open for several minutes).

    VM f1-micro instance running
    Fig. 1: Sensor unit (M5Stack ATOM Echo and SGP30 Sensor)

    This tutorial helps you to build DIY CO2 measuring devices (see Fig. 1) and connect them to your smartphone. The measurements can be an aid for Covid-19 prevention. The idea is that if there is a lot of exhaled CO2 in the air, there are also a lot of aerosols in the air. These smallest particles, to which Sars-CoV-2 viruses from the respiratory tract attach themselves, can float in the room air and can spread over several meters. A high concentration of aerosols increases the risk for everyone else in the room to become infected.

    You can read more about the correlation between CO2-concentration and the potential risk of Covid-19 infection in this paper: Risk assessment of aerosols loaded with virus based on CO2-concentration.

    Hardware and Software

    The following hardware and software will be used in this tutorial:

    DIY Sections

    I divide this project into three parts:

    • The sensor that acquires the data;
    • The server microservice that concentrates and publishes (eventually analyzes, processes, and saves) the data;
    • Website to read the data.

    As I used the M5Stack ATOM Echo, 4 possibilities to communicate the measurement are available:

    1. RGB Led
    2. Audio
    3. RestAPI (Local IP)
    4. Google Cloud Platform (GCP) Pub/Sub and RestAPI (Global IP).

    The last option enables you to integrate more than one sensor and process the data. But of course, it is more complicated to configure. I will use only certain services to remain under the "always free tier" limits on GCP. However, as a consequence, the performance of data streaming will be reduced, which is not a problem for this project. If you've decided to save the data, you'll need to pay some cents. The allows you to enable and disable the different options (RestAPI, audio, and GCP).

    DIY: Setup Sensor Firmware

    The M5Stack ATOM family devices are based on an ESP32 PICO. Therefore, it is possible to program them using MicroPython. I'm sure you've heard about MicroPython, but if you haven't, just refer to this article.

    Flashing MicroPython

    The M5Stack ATOM family has only 512 KB of RAM. So, to reduce the memory usage, I've frozen some modules (all the files under the folder frozen/), especially the ones related to keys, encryption, and webserver. Additionally, the ATOM Echo needs the I2S interface to send data to the audio amplifier IC for the speaker. The original MicroPython version doesn't include the driver yet. So, I've extended it using this repository. The resulting firmware is located here. You can download it and flash it following these instructions:

    1. Install esptool, if you haven't yet, follow these instructions. Usually, you only need to type:
      pip install esptool
    2. Connect the M5Stack ATOM (ECHO) to a USB port of your PC and erase the flash typing: --port /dev/ttyUSB0 erase_flash
    3. Flash the new firmware typing: --chip esp32 --port /dev/ttyUSB0 --baud 115200 write_flash -z 0x1000 micropython_b7883ce_esp32_idf4.x_i2s.bin 

    If you need detailed information about these steps, check out this tutorial: Tutorial: Getting Started with MicroPython on ESP32, M5Stack, and ESP8266. If you're running Windows just change the lines to:

    python3 -m esptool --chip esp32 --port COMXX [...]

    Sensor Application

    To measure the CO2 concentration, I used the SGP30. The sensor is still active, but it has a replacement, which is not easy to find integrated into a board :(. The SGP30 is inside the TVOC/eCO2 M5Stack unit and it is easy to integrate to the M5Stack Cores using the PH2.0-4P socket (Plug&Play).

    The SGP30, like all VOCs / gas sensors, has varying readings! For general environmental uses, it will give you a good idea of trends and comparison. The SGP30 does have built-in calibration capabilities. The eCO2 is calculated based on H2 concentration, therefore, it is not a 'true' CO2 sensor for laboratory use.

    I recommend using VSCode and the Pymakr extension to program and upload code to ESP32 boards running MicroPython firmware. A tutorial is available on this link: MicroPython: Visual Studio Code (VSCode) as IDE.

    Before you upload the application to the M5Stack, you need to modify some files and eventually create a .pem key. But, first of all, clone the repository typing:

    git clone

    RGB LED and Audio

    Rename the file to As I mentioned before, there are 4 possibilities to communicate the measurements. If you've decided to go offline, set the values:

    'gcp': False,
    'restapi': False,

    and eventually 'audio': True, to get the alarms visual and/or auditory. The thresholds are available with the labels:

    'warning': 600,
    'danger': 1000

    After getting the measurements, the M5Stack goes for deep sleep for about 10 seconds 'deepsleepms': 10000.

    RestAPI (Local IP)

    If you also want to go online, you need to add the Wi-Fi credentials of your router/access-point under:

    wifi_config = {

    If you want the local network version, just activate the flag 'restapi': True,. The M5Stack starts a web server and you can check the IP address of the M5Stack (reported on the VSCode terminal or your router) to get the measured values like Fig. 2 and Fig. 3.

    RestAPI website
    Fig. 2: Sensor website.
    Grant the service account permissions
    Fig. 3: Sensor RestAPI (local IP).

    The M5Stack doesn't go for a deep sleep in this case, because it loses the values and stops the webserver. Therefore, a "light" sleep is started.

    Google GCP (Pub/Sub) and RestAPI (Global IP)

    For the GCP version, you need to have a Google Account. Then, go to the Google Cloud Console.

    The CO2 sensor needs a .pem key to authenticate on Google IoT Core. To do create that, follow these steps:

    1. Type the following on a Terminal on your host PC to create a public/private key pair:
      openssl genrsa -out rsa_private.pem 2048
      openssl rsa -in rsa_private.pem -pubout -out rsa_public.pem
    2. Then, create a Registry and add a Device in the GCP, as described in the Cloud IoT Core documentation. Add to the device the public key (.pub) created in the previous step. I described these steps in detail in this tutorial.
    3. Using pip, install the rsa package on your host PC. This is needed to translate the RSA key for the device configuration.
      pip install rsa
    4. Run the following line inside the folder uPyCO2
      python utils/ >>

      The translated key will be appended at the end of the file.

    5. Update the file, move the key you've appended to the end of the file to the JSON key private_key (jwt_config['private_key']), and set also the Google Cloud Configuration section (google_cloud_config) of the file based on the Project ID, Registry ID and Device ID you set up in step 2. It should look like this example:
      google_cloud_config = {
      jwt_config = {
          'token_ttl': 43200, #12 hours
          # Use utiles/ to decode private pem to pkcs1.

    I have included some images and more details in an early tutorial this year. If you need more information, follow this link: MicroPython: GCP getting air-quality data from an M5Stack ATOM

    DIY: Setup Microservice Application

    The microservice application (source code inside the folder gcp/) will run in a VM instance on GCP.

    Google offers the f1-micro instance in the "always free tier" with a limitation of 720 hours/month, which means 30 days. This means you can run one instance of this VM machine within your account for free. The f1-micro offers 1 virtualized CPU core (with burst capability to help with sudden load spikes) and 614MB of memory. These specs don't sound great, but they are enough for this project if you can extend the RAM. This is possible using a SWAP file as RAM.

    I will be using the VM instance that I created for the project: GCP: Implementing Real-Time data pipelines - from ingest to datastore. Therefore, follow the instructions from that tutorial under the topics:

    • Create a VM instance
    • SWAP on VM instance
    • Docker on VM

    Subscription Pub/Sub

    To see the published messages, you need to subscribe to the topic that you've created while setting the device in the previous section. You can use the Google Pub/Sub API to make the subscription.

    To achieve that:

    1. Click on the Navigator Menu and scroll down to Pub/Sub and then activate the API.
    2. Click on the listed topic, e.g. projects/core-iot-sensors/topics/esp32-iot-tvoc-eco2-topic.
    3. Click on the Create Subscription button.
    4. Type a name for the subscription e.g. esp32-iot-tvoc and select the option Pull for the Delivery Type.
    5. Click on the Create button.

    You'll need the subscription name for the microservice to get the published messages. You can just rename the file .env.sample to .env and set the environmental variables PROJ_NAME= and PROJ_SUBS= with the "Project Name ID" and the "Subscription Name" values. In my case, the file looks like:


    Service Account (GCP Key)

    You may notice the GOOGLE_APPLICATION_CREDENTIALS env. variable in the .env file, so you will need also a Service Account key.

    To create a Service Account and get the JSON key file, follow these steps:

    1. Open the Service Accounts page in the Cloud Console.
    2. Click Select a project, choose your project, and click Open.
    3. Click Create Service Account.
    4. Enter a service account name (friendly display name) and then click on Create (see Fig. 4).
    5. Grant the service account the following permissions (see Fig. 5) and click on Done:
      • Pub/Sub Subscriber
    6. Click on the created Service Account
    7. Click on Add Key, Create new key and select JSON on the overlay and click on Create (see Fig. 6).
    8. Rename the downloaded file to service_account.json and move it inside the folder gcp/resources/.

    You can find more information about this topic here.

    Create a service account
    Fig. 4: Create a service account.
    Grant the service account permissions
    Fig. 5: Grant the service account permissions.
    Create a JSON key file
    Fig. 6: Create a JSON key file.

    Compiling the project

    In this subsection, you'll configure the project and set up an environment to upload the Docker files to the GCP to build the image using Cloud Build. You get 120 Build-minutes/day free to build images on GCP. Every extra minute will cost you USD0.003.

    Having said that, you'll need the Google Cloud SDK. To install it, follow the instructions presented here.

    To build the Docker image, you need to login to GCP using the SDK typing the following on a Terminal:

    gcloud auth login

    This opens a browser, in which you select your Google account and give the SDK permission to access your account. If everything works, you'll see a webpage saying "You are now authenticated with the Google Cloud SDK!"

    Then, you can select the project using:

    gcloud config set project <<project-name>>

    and submit the files to build the Docker image typing (inside the gcp/ folder):

    gcloud builds submit --config cloudbuild.yaml

    you'll get something like this:

    Creating temporary tarball archive of 11 file(s) totalling 10.0 KiB before compression.
    Uploading tarball of [.] to [gs://core-iot-sensors_cloudbuild/source/1[...]6dd.tgz]
    Fetching storage object: gs://core-iot-sensors_cloudbuild/source/[...]6dd.tgz#1604861150327172
    Copying gs://core-iot-sensors_cloudbuild/source/1[...]6dd.tgz#1604861150327172...
    / [1 files][  5.6 KiB/  5.6 KiB]                                                
    Operation completed over 1 objects/5.6 KiB.                                      
    Already have image (with digest):
    Sending build context to Docker daemon   21.5kB
    Step 1/13 : FROM python:3.7-buster
    3.7-buster: Pulling from library/python
    ID                                    CREATE_TIME                DURATION  SOURCE                                                                                          IMAGES                                      STATUS
    e1e95d81-f864-40a6-be4d-cc5c727cb12b  2020-11-08T18:45:50+00:00  1M18S     gs://core-iot-sensors_cloudbuild/source/1[...]6dd.tgz  SUCCESS

    As you can see in the log above, building the image took me about 78 seconds.

    To check the built images, go to the Container Registry page in the Cloud Console:

    DIY: Deploy the Microservice Application

    After creating the Docker images, you can use docker-compose to deploy the micro-service application. But first, you need to give the VM machine access to the Container Registry and set up environment variables for the Docker containers.

    Service Account (GCP key)

    As you may have already noticed, you'll need a Service Account to access the Container Registry. To create a service account and get a JSON key file, follow the steps described above. This time, just grant the Storage Object Viewer permission to the Service Account as shown in Fig. 7.

    Create service account
    Fig. 7: Create a service account to access the Container Registry.
    Upload files to the VM
    Fig. 8: Upload files to the f1-micro VM instance.
    VM f1-micro instance running
    Fig. 9: VM f1-micro instance running.

    After creating and downloading the key, follow these steps to upload the key to the f1-micro VM:

    1. Click on the SSH button from the f1-micro VM instance listed on the Google Compute Engine control panel (see Fig. 9).
    2. Upload the downloaded file by clicking on the Gear > Upload file option (see Fig. 8).
    3. Authenticate to the Container Registry typing the following on the SSH Terminal:
      docker login -u _json_key -p "$(cat [FILENAME].json)" https://[[HOSTNAME]]

      Replace [FILENAME] with the uploaded filename and [HOSTNAME] is,,, or, according to where your registry is located. You will get something like this:

      WARNING! Using --password via the CLI is insecure. Use --password-stdin.
      Login Succeeded

      More info here. If you prefer a secure way to store your credentials, check the following tutorial.

    Start the Application

    The microservice can offer a RestAPI over HTTP or HTTPS. To get the last one, you need to have a domain in which you can edit the A or AAA register of the DNS configuration. I've included two docker-compose.yml files:

    1. docker-compose.yml: it runs only the upyco2-service and offers the RestAPI a Website using the port 8080.
    2. docker-compose-https.yml: it runs two microservices (upyco2-service and Nginx). The Nginx service offers HTTPS and forward the queries using a revese proxy to the upyco2-service service. If you want this option, rename the file to docker-compose.yml and follow the instructions of this tutorial: Nginx: Docker, Flask and Let's encrypt (SSL).

    HTTP Microservice

    To launch the application you need to upload the docker-compose.yml and the .env file to the VM instance in GCP. Buf first,

    1. Modify the project name (core-iot-sensors) inside the docker-compose.yml to match your GCP project name.
    2. Create a folder e.g. co2-measurement inside the VM using the SSH terminal
      mkdir co2-measurement
    3. Upload the .env and docker-compose.yml to the folder using the Gear > Upload file option (see Fig. 9). If you don't see the .env file (linux), use the right-click on the file browser and activate the Show hidden Files option. The files will be uploaded to the $HOME folder, therefore you need to move them inside the folder:
    4. Move the files inside the co2-measurement folder:
      cd co2-measurement
      mv ~/.env .
      mv ~/docker-compose.yml .
    5. Start the microservice application by typing:
      docker-compose up -d 

      inside the folder where the docker-compose.yml file is located (see Fig. 10).

    Setup the Firewall

    To access the RestAPI, you should forward a port through the firewall to the VM instance. To do that, you need to set up a new rule (or use the rules: default-allow-http, default-allow-https):

    1. Open the Firewall Setting under the VPC network.
    2. Click on the Create Firewall Rule button.
    3. Enter:
      • Name e.g. default-http-alt
      • Target tags e.g http-allow-alt
      • Source IP ranges e.g. for all the internet, or your specified IP of your PC.
      • Under Specified protocols and ports, activate the TCP box and add the port that you want to forward to the VM instance, e.g. 8080.
    4. Click on Create.

    You can find more information about this topic here.

    Then, you need to edit the configuration of your VM instance to add Network tags to it. To do that:

    1. Open the VM instances menu
    2. Click on the VM instance's name.
    3. Click on the Edit button.
    4. Enter the name of the firewall rule that you've created before in the field Network tags (see Fig. 11).
    5. Scroll down and click on Save.
    Starting the microservice application
    Fig. 10: Starting the microservice application.
    VM instance settings
    Fig. 11: VM instance settings.

    DIY: Smartphone Report

    You can use your browser to check the website or the RestAPI like shown in Fig. 12 and 13. As I mentioned before, using the GCP enables you to integrate more than one sensor and process the data. Therefore, to get the data of a sensor, you need to visit the following URLs:

    • RestAPI: http://[IP-ADDRESS]/dev?deviceId=[DEVICE_ID]
    • Website: http://[IP-ADDRESS]/?deviceId=[DEVICE_ID]

    Replace the [IP-ADDRESS] with the External IP described in Fig. 9 and [DEVICE_ID] with the Device ID that you configured on the Google IoT Core.

    GCP Website
    Fig. 12: GCP Website.
    VM instance settings
    Fig. 13: GCP RestAPI.

    On your Android, you can use the Gauge Widget application, but it requires an HTTPS URL. On iOS, you can use Scriptable app. The following tutorials include the steps that you need to activate SSL and use Gauge Widget or Scriptable app:

    Double Sterilization: O2 and Ozone Cabinet

    If you are looking to disinfect your masks, I can recommend the following article: Xiaomi Smartda UV Disinfection Cabinet.


    It's from the Xiaomi family. It has two lamps (UV lamp and Ozone lamp) and an air-drying functionality up to 55°C. The cabinet is easy to use with touch buttons on the front.