Adafruit.IO nRF24L01 Windows 10 IoT Core Field Gateway with BorosRF2

A couple of BorosRF2 Dual nRF24L01 Hats arrived earlier in the week. After some testing with my nRF24L01 Test application I have added compile-time configuration options for the two nRF24L01 sockets to my Adafruit.IO nRF24L01 Field Gateway.

Boros RF2 with Dual nRF24L01 devices
public sealed class StartupTask : IBackgroundTask
{
   private const string ConfigurationFilename = "config.json";

   private const byte MessageHeaderPosition = 0;
   private const byte MessageHeaderLength = 1;

   // nRF24 Hardware interface configuration
#if CEECH_NRF24L01P_SHIELD
   private const byte RF24ModuleChipEnablePin = 25;
   private const byte RF24ModuleChipSelectPin = 0;
   private const byte RF24ModuleInterruptPin = 17;
#endif

#if BOROS_RF2_SHIELD_RADIO_0
   private const byte RF24ModuleChipEnablePin = 24;
   private const byte RF24ModuleChipSelectPin = 0;
   private const byte RF24ModuleInterruptPin = 27;
#endif

#if BOROS_RF2_SHIELD_RADIO_1
   private const byte RF24ModuleChipEnablePin = 25;
   private const byte RF24ModuleChipSelectPin = 1;
   private const byte RF24ModuleInterruptPin = 22;
#endif

private readonly LoggingChannel loggingChannel = new LoggingChannel("devMobile AdaFruit.IO nRF24L01 Field Gateway", null, new Guid("4bd2826e-54a1-4ba9-bf63-92b73ea1ac4a"));
private readonly RF24 rf24 = new RF24();

For this initial version only one nRF24L01 device socket active at a time is supported.

Windows 10 IoT Core BorosRf2 – Dual nRF24L01 pHat/Hat

I have a couple of nRF24L01P Raspberry PI projects (primarily my Adafruit.IO and Azure IoT Hubs/Central Windows 10 IoT Core telemetry field gateways) and recently Boros Lab a vendor of suitable Raspberry PI Hats opened a store on Tindie.com.

I ordered a couple of BorosRf2 – Dual nRF24L01 pHat/Hat + RTC for Pis (mine were without the Real-time clock(RTC)) for testing. The vendor’s github repository had details of the GPIO pins used so it was relatively quick and easy to modify my Windows 10 IoT nRF24L01 test harness to work with a single port on the hat.

Single port configuration

By setting a conditional compile option (CEECH_NRF24L01P_SHIELD, BOROS_RF2_SHIELD_RADIO_0 or BOROS_RF2_SHIELD_RADIO_1) my test application could be configured to support the Boros or Ceech (with a modification detailed here) shields.

namespace devmobile.IoTCore.nRF24L01BackGroundTask
{
	public sealed class StartupTask : IBackgroundTask
	{
		// nRF24 Hardware interface configuration
#if CEECH_NRF24L01P_SHIELD
      private const byte ChipEnablePin = 25;
      private const byte ChipSelectPin = 0;
      private const byte InterruptPin = 17;
#endif
#if BOROS_RF2_SHIELD_RADIO_0
      private const byte ChipEnablePin = 24;
      private const byte ChipSelectPin = 0;
      private const byte InterruptPin = 27;
#endif
#if BOROS_RF2_SHIELD_RADIO_1
      private const byte ChipEnablePin = 25;
      private const byte ChipSelectPin = 1;
      private const byte InterruptPin = 22;
#endif
      private const string BaseStationAddress = "Node1";
      private const byte nRF24Channel = 20;
      private RF24 Radio = new RF24();
      private BackgroundTaskDeferral deferral;
      private ThreadPoolTimer timer;

Both vendors’ shields worked well with my test application, the ceech shield (USD9.90 April 2019) is a little bit cheaper, but the Boros shield (USD15.90 April 2019 ) doesn’t require any modification and has a socket for a second nRF24 device.

Windows 10 IoT Core Time-Lapse Camera Azure IoT Hub Storage Revisited

In my previous post the application uploaded images to an Azure storage account associated with an Azure IoT Hub based on configuration file settings. The application didn’t use any of the Azure IoT Hub device management functionality like device twins and direct methods.

Time-lapse camera setup

In this version only the Azure IoT hub connection string and protocol to use are stored in the JSON configuration file.

{
  "AzureIoTHubConnectionString": "",
  "TransportType": "Mqtt",
} 

On startup the application uploads a selection of properties to the Azure IoT Hub to assist with support, fault finding etc.

// This is from the OS 
reportedProperties["Timezone"] = TimeZoneSettings.CurrentTimeZoneDisplayName;
reportedProperties["OSVersion"] = Environment.OSVersion.VersionString;
reportedProperties["MachineName"] = Environment.MachineName;
reportedProperties["ApplicationDisplayName"] = package.DisplayName;
reportedProperties["ApplicationName"] = packageId.Name;
reportedProperties["ApplicationVersion"] = string.Format($"{version.Major}.{version.Minor}.{version.Build}.{version.Revision}");

// Unique identifier from the hardware
SystemIdentificationInfo systemIdentificationInfo = SystemIdentification.GetSystemIdForPublisher();
using (DataReader reader = DataReader.FromBuffer(systemIdentificationInfo.Id))
{
   byte[] bytes = new byte[systemIdentificationInfo.Id.Length];
   reader.ReadBytes(bytes);
   reportedProperties["SystemId"] = BitConverter.ToString(bytes);
}

Azure Portal Device Properties

The Azure Storage file and folder name formats along with the image capture due and update periods are configured in the DeviceTwin properties. Initially I had some problems with the dynamic property types so had to .ToString and then Timespan.TryParse the periods.

Twin deviceTwin= azureIoTHubClient.GetTwinAsync().Result;

if (!deviceTwin.Properties.Desired.Contains("AzureImageFilenameLatestFormat"))
{
   this.logging.LogMessage("DeviceTwin.Properties AzureImageFilenameLatestFormat setting missing", LoggingLevel.Warning);
   return;
}
…
if (!deviceTwin.Properties.Desired.Contains("ImageUpdateDue") || !TimeSpan.TryParse(deviceTwin.Properties.Desired["ImageUpdateDue"].Value.ToString(), out imageUpdateDue))
{
   this.logging.LogMessage("DeviceTwin.Properties ImageUpdateDue setting missing or invalid format", LoggingLevel.Warning);
   return;
}
Azure Portal Device Settings

The application also supports two commands “ImageCapture’ and “DeviceReboot”. For testing I used Azure Device Explorer

After running the installer (available from GitHub) the application will create a default configuration file in

\User Folders\LocalAppData\PhotoTimerTriggerAzureIoTHubStorage-uwp_1.2.0.0_arm__nmn3tag1rpsaw\LocalState\

Which can be downloaded, modified then uploaded using the portal file explorer application. If you want to make the application run on device start-up the radio button below needs to be selected.

Windows 10 IoT Core Time-Lapse Camera Azure IoT Hub Storage

After building a couple of time lapse camera applications for Windows 10 IoT Core I built a version which uploads the images to the Azure storage account associated with an Azure IoT Hub.

I really wanted to be able to do a time-lapse video of a storm coming up the Canterbury Plains to Christchurch and combine it with the wind direction, windspeed, temperature and humidity data from my weather station which uploads data to Azure through my Azure IoT Hub LoRa field gateway.

Time-lapse camera setup

The application captures images with a configurable period after configurable start-up delay. The Azure storage root folder name is based on the device name in the Azure IoT Hub connection string. The folder(s) where the historic images are stored are configurable and the images can optionally be in monthly, daily, hourly etc. folders. The current image is stored in the root folder for the device and it’s name is configurable.

{
  "AzureIoTHubConnectionString": "",
  "TransportType": "Mqtt",
  "AzureImageFilenameFormatLatest": "latest.jpg",
  "AzureImageFilenameFormatHistory": "{0:yyMMdd}/{0:yyMMddHHmmss}.jpg",
  "ImageUpdateDueSeconds": 30,
  "ImageUpdatePeriodSeconds": 300
} 

With the above setup I have a folder for each device in the historic fiolder and the most recent image i.e. “latest.jpg” in the root folder. The file and folder names are assembled with a parameterised string.format . The parameter {0} is the current UTC time

Pay attention to your folder/file name formatting, I was tripped up by

  • mm – minutes vs. MM – months
  • hh – 12 hour clock vs. HH -24 hour clock

With 12 images every hour

The application logs events on start-up and every time a picture is taken

After running the installer (available from GitHub) the application will create a default configuration file in

User Folders\LocalAppData\PhotoTimerTriggerAzureIoTHubStorage-uwp_1.0.0.0_arm__nmn3tag1rpsaw\LocalState\

Which can be downloaded, modified then uploaded using the portal file explorer application. If you want to make the application run on device start-up the radio button below needs to be selected.

/*
    Copyright ® 2019 March devMobile Software, All Rights Reserved
 
    MIT License

…
*/
namespace devMobile.Windows10IotCore.IoT.PhotoTimerTriggerAzureIoTHubStorage
{
	using System;
	using System.IO;
	using System.Diagnostics;
	using System.Threading;

	using Microsoft.Azure.Devices.Client;
	using Microsoft.Extensions.Configuration;

	using Windows.ApplicationModel;
	using Windows.ApplicationModel.Background;
	using Windows.Foundation.Diagnostics;
	using Windows.Media.Capture;
	using Windows.Media.MediaProperties;
	using Windows.Storage;
	using Windows.System;
	
	public sealed class StartupTask : IBackgroundTask
	{
		private BackgroundTaskDeferral backgroundTaskDeferral = null;
		private readonly LoggingChannel logging = new LoggingChannel("devMobile Photo Timer Azure IoT Hub Storage", null, new Guid("4bd2826e-54a1-4ba9-bf63-92b73ea1ac4a"));
		private DeviceClient azureIoTHubClient = null;
		private const string ConfigurationFilename = "appsettings.json";
		private Timer ImageUpdatetimer;
		private MediaCapture mediaCapture;
		private string azureIoTHubConnectionString;
		private TransportType transportType;
		private string azureStorageimageFilenameLatestFormat;
		private string azureStorageImageFilenameHistoryFormat;
		private const string ImageFilenameLocal = "latest.jpg";
		private volatile bool cameraBusy = false;

		public void Run(IBackgroundTaskInstance taskInstance)
		{
			StorageFolder localFolder = ApplicationData.Current.LocalFolder;
			int imageUpdateDueSeconds;
			int imageUpdatePeriodSeconds;

			this.logging.LogEvent("Application starting");

			// Log the Application build, OS version information etc.
			LoggingFields startupInformation = new LoggingFields();
			startupInformation.AddString("Timezone", TimeZoneSettings.CurrentTimeZoneDisplayName);
			startupInformation.AddString("OSVersion", Environment.OSVersion.VersionString);
			startupInformation.AddString("MachineName", Environment.MachineName);

			// This is from the application manifest 
			Package package = Package.Current;
			PackageId packageId = package.Id;
			PackageVersion version = packageId.Version;
			startupInformation.AddString("ApplicationVersion", string.Format($"{version.Major}.{version.Minor}.{version.Build}.{version.Revision}"));

			try
			{
				// see if the configuration file is present if not copy minimal sample one from application directory
				if (localFolder.TryGetItemAsync(ConfigurationFilename).AsTask().Result == null)
				{
					StorageFile templateConfigurationfile = Package.Current.InstalledLocation.GetFileAsync(ConfigurationFilename).AsTask().Result;
					templateConfigurationfile.CopyAsync(localFolder, ConfigurationFilename).AsTask();

					this.logging.LogMessage("JSON configuration file missing, templated created", LoggingLevel.Warning);
					return;
				}

				IConfiguration configuration = new ConfigurationBuilder().AddJsonFile(Path.Combine(localFolder.Path, ConfigurationFilename), false, true).Build();

				azureIoTHubConnectionString = configuration.GetSection("AzureIoTHubConnectionString").Value;
				startupInformation.AddString("AzureIoTHubConnectionString", azureIoTHubConnectionString);

				transportType = (TransportType)Enum.Parse( typeof(TransportType), configuration.GetSection("TransportType").Value);
				startupInformation.AddString("TransportType", transportType.ToString());

				azureStorageimageFilenameLatestFormat = configuration.GetSection("AzureImageFilenameFormatLatest").Value;
				startupInformation.AddString("ImageFilenameLatestFormat", azureStorageimageFilenameLatestFormat);

				azureStorageImageFilenameHistoryFormat = configuration.GetSection("AzureImageFilenameFormatHistory").Value;
				startupInformation.AddString("ImageFilenameHistoryFormat", azureStorageImageFilenameHistoryFormat);

				imageUpdateDueSeconds = int.Parse(configuration.GetSection("ImageUpdateDueSeconds").Value);
				startupInformation.AddInt32("ImageUpdateDueSeconds", imageUpdateDueSeconds);

				imageUpdatePeriodSeconds = int.Parse(configuration.GetSection("ImageUpdatePeriodSeconds").Value);
				startupInformation.AddInt32("ImageUpdatePeriodSeconds", imageUpdatePeriodSeconds);
			}
			catch (Exception ex)
			{
				this.logging.LogMessage("JSON configuration file load or settings retrieval failed " + ex.Message, LoggingLevel.Error);
				return;
			}

			try
			{
				azureIoTHubClient = DeviceClient.CreateFromConnectionString(azureIoTHubConnectionString, transportType);
			}
			catch (Exception ex)
			{
				this.logging.LogMessage("AzureIOT Hub connection failed " + ex.Message, LoggingLevel.Error);
				return;
			}

			try
			{
				mediaCapture = new MediaCapture();
				mediaCapture.InitializeAsync().AsTask().Wait();
			}
			catch (Exception ex)
			{
				this.logging.LogMessage("Camera configuration failed " + ex.Message, LoggingLevel.Error);
				return;
			}

			ImageUpdatetimer = new Timer(ImageUpdateTimerCallback, null, new TimeSpan(0, 0, imageUpdateDueSeconds), new TimeSpan(0, 0, imageUpdatePeriodSeconds));

			this.logging.LogEvent("Application started", startupInformation);

			//enable task to continue running in background
			backgroundTaskDeferral = taskInstance.GetDeferral();
		}

		private async void ImageUpdateTimerCallback(object state)
		{
			DateTime currentTime = DateTime.UtcNow;
			Debug.WriteLine($"{DateTime.UtcNow.ToLongTimeString()} Timer triggered");

			// Just incase - stop code being called while photo already in progress
			if (cameraBusy)
			{
				return;
			}
			cameraBusy = true;

			try
			{
				using (Windows.Storage.Streams.InMemoryRandomAccessStream captureStream = new Windows.Storage.Streams.InMemoryRandomAccessStream())
				{
					await mediaCapture.CapturePhotoToStreamAsync(ImageEncodingProperties.CreateJpeg(), captureStream);
					await captureStream.FlushAsync();
#if DEBUG
					IStorageFile photoFile = await KnownFolders.PicturesLibrary.CreateFileAsync(ImageFilenameLocal, CreationCollisionOption.ReplaceExisting);
					ImageEncodingProperties imageProperties = ImageEncodingProperties.CreateJpeg();
					await mediaCapture.CapturePhotoToStorageFileAsync(imageProperties, photoFile);
#endif

					string azureFilenameLatest = string.Format(azureStorageimageFilenameLatestFormat, currentTime);
					string azureFilenameHistory = string.Format(azureStorageImageFilenameHistoryFormat, currentTime);

					LoggingFields imageInformation = new LoggingFields();
					imageInformation.AddDateTime("TakenAtUTC", currentTime);
#if DEBUG
					imageInformation.AddString("LocalFilename", photoFile.Path);
#endif
					imageInformation.AddString("AzureFilenameLatest", azureFilenameLatest);
					imageInformation.AddString("AzureFilenameHistory", azureFilenameHistory);
					this.logging.LogEvent("Saving image(s) to Azure storage", imageInformation);

					// Update the latest image in storage
					if (!string.IsNullOrWhiteSpace(azureFilenameLatest))
					{
						captureStream.Seek(0);
						Debug.WriteLine("AzureIoT Hub latest image upload start");
						await azureIoTHubClient.UploadToBlobAsync(azureFilenameLatest, captureStream.AsStreamForRead());
						Debug.WriteLine("AzureIoT Hub latest image upload done");
					}

					// Upload the historic image to storage
					if (!string.IsNullOrWhiteSpace(azureFilenameHistory))
					{
						captureStream.Seek(0);
						Debug.WriteLine("AzureIoT Hub historic image upload start");
						await azureIoTHubClient.UploadToBlobAsync(azureFilenameHistory, captureStream.AsStreamForRead());
						Debug.WriteLine("AzureIoT Hub historic image upload done");
					}
				}
			}
			catch (Exception ex)
			{
				this.logging.LogMessage("Camera photo save or AzureIoTHub storage upload failed " + ex.Message, LoggingLevel.Error);
			}
			finally
			{
				cameraBusy = false;
			}
		}
	}
}

The images in Azure Storage could then be assembled into a video using a tool like Time Lapse Creator or processed with Azure Custom Vision Service.

RFM9X.IoTCore Adafruit LoRa Radio Bonnet support

The RFM9X chip select line on the Adafruit LoRa Radio Bonnet 868 or 915MHz with OLED RFM95W is connected to pin 26(CS1), the reset line to pin 22(GPIO25) and the interrupt line to pin 15(GPIO22).

When I ran the RFM9XLoRaDeviceClient from my RFM9X.IoTCore library with the following configuration

#if ADAFRUIT_RADIO_BONNET
	private const byte ResetLine = 25;
	private const byte InterruptLine = 22;
	private Rfm9XDevice rfm9XDevice = new Rfm9XDevice(ChipSelectPin.CS1, ResetLine, InterruptLine);
#endif

public void Run(IBackgroundTaskInstance taskInstance)
{
	rfm9XDevice.Initialise(Frequency, paBoost: true, rxPayloadCrcOn : true);
#if DEBUG
	rfm9XDevice.RegisterDump();
#endif
	rfm9XDevice.OnReceive += Rfm9XDevice_OnReceive;
#if ADDRESSED_MESSAGES_PAYLOAD
	rfm9XDevice.Receive(UTF8Encoding.UTF8.GetBytes(Environment.MachineName));
#else
	rfm9XDevice.Receive();
#endif
	rfm9XDevice.OnTransmit += Rfm9XDevice_OnTransmit;

	Task.Delay(10000).Wait();

	while (true)
	{
		string messageText = string.Format("Hello from {0} ! {1}", Environment.MachineName, MessageCount);
		MessageCount -= 1;

		byte[] messageBytes = UTF8Encoding.UTF8.GetBytes(messageText);
		Debug.WriteLine("{0:HH:mm:ss}-TX {1} byte message {2}", DateTime.Now, messageBytes.Length, messageText);
#if ADDRESSED_MESSAGES_PAYLOAD
		this.rfm9XDevice.Send(UTF8Encoding.UTF8.GetBytes("AddressHere"), messageBytes);
#else
		this.rfm9XDevice.Send(messageBytes);
#endif
		Task.Delay(10000).Wait();
	}
}
#endif

I could see messages being sent and received in the debug output

Register 0x3e - Value 0X00 - Bits 00000000
Register 0x3f - Value 0X00 - Bits 00000000
Register 0x40 - Value 0X00 - Bits 00000000
Register 0x41 - Value 0X00 - Bits 00000000
Register 0x42 - Value 0X12 - Bits 00010010
...
The thread 0xec4 has exited with code 0 (0x0).
The thread 0x868 has exited with code 0 (0x0).
22:21:47-RX PacketSnr 9.8 Packet RSSI -80dBm RSSI -122dBm = 59 byte message "�LoRaIoT1Maduino2at 62.8,ah 77,wsa 1,wsg 3,wd 34.88,r 0.00,"
22:21:52-TX 31 byte message Hello from AdaFruitIOLoRa ! 255
22:21:52-TX Done
The thread 0xbf8 has exited with code 0 (0x0).
The program '[3380] backgroundTaskHost.exe' has exited with code -1 (0xffffffff).

Next step modify my Adafruit IO and Azure IoT Hub/Central field gateways.

Windows 10 IoT Core Field Gateways “less is more”

After looking back at the technical support interactions for my Azure IoT Hubs Windows 10 IoT Core Field Gateway & AdaFruit.IO LoRa Windows 10 IoT Core Field Gateway I think removing a “feature” might make it easier for first time users.

In an early version of the software I used to provide a sample configuration JSON file in the associated GitHub repository. Users had to download this file to a computer, update it with their Azure IOT Hub or Azure IoT Central connection string or AdafruitIO APIKey , frequency and device address, then upload to the field gateway.

In a later version of the software I added code which created an empty configuration file with defaults for all settings, many of which were a distraction as the majority of users would never change them.

More settings meant there was more scope for users to change settings which broke the device samples and the gateway.

I have removed the code to generate the full configuration file (starting with Azure IOT Hub field gateway) and included a sample configuration file with the minimum required settings in the GitHub repositories and installers.

I am assuming that if a user wants to change advanced settings they can look at the code and/or documentation and figure out the setting names and valid values.

The new sample configuration file for a Azure IoT Hub telemetry only gateway is

{
  "AzureIoTHubDeviceConnectionString": "Azure IOT Hub connection string",
  "AzureIoTHubTransportType": "amqp",
  "SensorIDIsDeviceIDSensorID": false,
  "Address": "Device address",
  "Frequency": 915000000.0
}

The prebuilt installers available on GitHub post version 1.0.13.0 (Azure IoT Hub) and 1.0.5.0 (Adafruit.IO) will implement this model.

Carbon Dioxide Sensor(MH-Z16) library comparison

The first library I looked at was for the DFRobot Gravity: UART Infrared CO2 Sensor (0-50000ppm). There was sample code provided on the associated wiki page. The code worked first time I ran it but I didn’t use this library due to the lack of checksum & packet header/footer validation.

/***************************************************
* Infrared CO2 Sensor 0-50000ppm(Wide Range)
* ****************************************************
* The follow example is used to detect CO2 concentration.
  
* @author lg.gang(lg.gang@qq.com)
* @version  V1.0
* @date  2016-6-6
  
* GNU Lesser General Public License.
* See <http://www.gnu.org/licenses/> for details.
* All above must be included in any redistribution
* ****************************************************/ 
#include <SoftwareSerial.h>
SoftwareSerial mySerial(10, 11); // RX, TX
unsigned char hexdata[9] = {0xFF,0x01,0x86,0x00,0x00,0x00,0x00,0x00,0x79}; //Read the gas density command /Don't change the order
void setup() {
  
  Serial.begin(9600);
  while (!Serial) {

  }
  mySerial.begin(9600);

}

void loop() {
   mySerial.write(hexdata,9);
   delay(500);

 for(int i=0,j=0;i<9;i++)
 {
  if (mySerial.available()>0)
  {
     long hi,lo,CO2;
     int ch=mySerial.read();

    if(i==2){     hi=ch;   }   //High concentration
    if(i==3){     lo=ch;   }   //Low concentration
    if(i==8) {
               CO2=hi*256+lo;  //CO2 concentration
      Serial.print("CO2 concentration: ");
      Serial.print(CO2);
      Serial.println("ppm");      
      }
    }   
  } 
}

After some GitHub searching the second library I looked at was abbozza_CO2_MHZ16_arduino by Michael Brinkmeier. This library appears to be “plug-in” module for the abbozza! framework. I didn’t use this library due to the lack of checksum & packet header/footer validation.

/**
 * @license
 * abbozza! Calliope plugin for the MH-Z16 CO2 sensor
 * 
 * The sensor has to be connected to a serial connection with 9600 baud.
 *
 * Copyright 2015 Michael Brinkmeier ( michael.brinkmeier@uni-osnabrueck.de )
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *   http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#include "SoftwareSerial.h"
#include "MHZ16.h"
#include "Arduino.h"

MHZ16::MHZ16(int tx, int rx) {
    this->serial = new SoftwareSerial(rx,tx,false);
    this->serial->begin(9600);
}


void MHZ16::calibrate() {
    int idx;
    for (idx = 0; idx < 9; idx++) {
        serial->write(cal[idx]);
    }
    delay(10);
}

void MHZ16::doMeasurement() {
    int idx;
    int bu;

    for (idx = 0; idx < 9; idx++) {
        serial->write(cmd[idx]);
    }
    delay(10);

    while (serial->available()) {
        do {
            bu = serial->read();
        } while (bu != 255);
        buf[0] = bu;

        idx = 1;
        while (serial->available() && (idx < 9)) {
            bu = serial->read();
            buf[idx] = bu;
            idx++;
        }

        if (idx == 9) {
            PPM = ((int) buf[2]) *256 + ((int) buf[3]);
        }
    }
}

int MHZ16::getPPM() {
    return PPM;
}

The third library was produced by Sandbox electronics for their selection of 10,000PPM thru 100,000PPM MH-Z16 sensors. Their datasheet looked similar(maybe newer?) to the Seeedstudio one and the packet format was the same.

Their library had checksum & packet header/footer validation but I didn’t use it because the carbon dioxide concentration was calculated using 4 bytes (maybe this was to support the different range sensors?)

/*
Description:
This is a example code for Sandbox Electronics NDIR CO2 sensor module.
You can get one of those products on
http://sandboxelectronics.com

Version:
V1.2

Release Date:
2019-01-10

Author:
Tiequan Shao          support@sandboxelectronics.com

Lisence:
CC BY-NC-SA 3.0

Please keep the above information when you use this code in your project.
*/

#include <SoftwareSerial.h>
#include <NDIR_SoftwareSerial.h>
#define  RECEIVE_TIMEOUT  (100)

#if ARDUINO >= 100
    #include "Arduino.h"
#else
    #include "WProgram.h"
#endif

class SoftwareSerial;

uint8_t NDIR_SoftwareSerial::cmd_measure[9]                = {0xFF,0x01,0x9C,0x00,0x00,0x00,0x00,0x00,0x63};
uint8_t NDIR_SoftwareSerial::cmd_calibrateZero[9]          = {0xFF,0x01,0x87,0x00,0x00,0x00,0x00,0x00,0x78};
uint8_t NDIR_SoftwareSerial::cmd_enableAutoCalibration[9]  = {0xFF,0x01,0x79,0xA0,0x00,0x00,0x00,0x00,0xE6};
uint8_t NDIR_SoftwareSerial::cmd_disableAutoCalibration[9] = {0xFF,0x01,0x79,0x00,0x00,0x00,0x00,0x00,0x86};

NDIR_SoftwareSerial::NDIR_SoftwareSerial(uint8_t rx_pin, uint8_t tx_pin) : serial(rx_pin, tx_pin, false)
{
}


uint8_t NDIR_SoftwareSerial::begin()
{
    serial.begin(9600);

    if (measure()) {
        return true;
    } else {
        return false;
    }
}

uint8_t NDIR_SoftwareSerial::measure()
{
    uint8_t i;
    uint8_t buf[9];
    uint32_t start = millis();

    serial.flush();

    for (i=0; i<9; i++) {
        serial.write(cmd_measure[i]);
    }

    for (i=0; i<9;) {
        if (serial.available()) {
            buf[i++] = serial.read();
        }

        if (millis() - start > RECEIVE_TIMEOUT) {
            return false;
        }
    }

    if (parse(buf)) {
        return true;
    }

    return false;
}


void NDIR_SoftwareSerial::calibrateZero()
{
    uint8_t i;

    for (i=0; i<9; i++) {
        serial.write(cmd_calibrateZero[i]);
    }
}


void NDIR_SoftwareSerial::enableAutoCalibration()
{
    uint8_t i;

    for (i=0; i<9; i++) {
        serial.write(cmd_enableAutoCalibration[i]);
    }
}


void NDIR_SoftwareSerial::disableAutoCalibration()
{
    uint8_t i;

    for (i=0; i<9; i++) {
        serial.write(cmd_disableAutoCalibration[i]);
    }
}


uint8_t NDIR_SoftwareSerial::parse(uint8_t *pbuf)
{
    uint8_t i;
    uint8_t checksum = 0;

    for (i=0; i<9; i++) {
        checksum += pbuf[i];
    }

    if (pbuf[0] == 0xFF && pbuf[1] == 0x9C && checksum == 0xFF) {
        ppm = (uint32_t)pbuf[2] << 24 | (uint32_t)pbuf[3] << 16 | (uint32_t)pbuf[4] << 8 | pbuf[5];
        return true;
    } else {
        return false;
    }
}

The forth library I looked at was MHZ-Z-C02-Sensors by Tobias Schürg this library was for different series of MHZ sensors. With re-synching, configurable timeouts and checksum validation it looked like the code could easily be adapted for the MH-Z16.

/* MHZ library

    By Tobias Schürg
*/

#include "MHZ.h"

const int MHZ14A = 14;
const int MHZ19B = 19;

const int MHZ14A_RESPONSE_TIME = 60;
const int MHZ19B_RESPONSE_TIME = 120;

const int STATUS_NO_RESPONSE = -2;
const int STATUS_CHECKSUM_MISMATCH = -3;
const int STATUS_INCOMPLETE = -4;
const int STATUS_NOT_READY = -5;

unsigned long lastRequest = 0;

MHZ::MHZ(uint8_t rxpin, uint8_t txpin, uint8_t pwmpin, uint8_t type)
    : co2Serial(rxpin, txpin) {
  _rxpin = rxpin;
  _txpin = txpin;
  _pwmpin = pwmpin;
  _type = type;

  co2Serial.begin(9600);
}

/**
 * Enables or disables the debug mode (more logging).
 */
void MHZ::setDebug(boolean enable) {
  debug = enable;
  if (debug) {
    Serial.println(F("MHZ: debug mode ENABLED"));
  } else {
    Serial.println(F("MHZ: debug mode DISABLED"));
  }
}

boolean MHZ::isPreHeating() {
  if (_type == MHZ14A) {
    return millis() < (3 * 60 * 1000);
  } else if (_type == MHZ19B) {
    return millis() < (3 * 60 * 1000);
  } else {
    Serial.println(F("MHZ::isPreHeating() => UNKNOWN SENSOR"));
    return false;
  }
}

boolean MHZ::isReady() {
  if (isPreHeating()) return false;
  if (_type == MHZ14A)
    return lastRequest < millis() - MHZ14A_RESPONSE_TIME;
  else if (_type == MHZ19B)
    return lastRequest < millis() - MHZ19B_RESPONSE_TIME;
  else {
    Serial.print(F("MHZ::isReady() => UNKNOWN SENSOR \""));
    Serial.print(_type);
    Serial.println(F("\""));
    return true;
  }
}

int MHZ::readCO2UART() {
  if (!isReady()) return STATUS_NOT_READY;
  if (debug) Serial.println(F("-- read CO2 uart ---"));
  byte cmd[9] = {0xFF, 0x01, 0x86, 0x00, 0x00, 0x00, 0x00, 0x00, 0x79};
  byte response[9];  // for answer

  if (debug) Serial.print(F("  >> Sending CO2 request"));
  co2Serial.write(cmd, 9);  // request PPM CO2
  lastRequest = millis();

  // clear the buffer
  memset(response, 0, 9);

  int waited = 0;
  while (co2Serial.available() == 0) {
    if (debug) Serial.print(".");
    delay(100);  // wait a short moment to avoid false reading
    if (waited++ > 10) {
      if (debug) Serial.println(F("No response after 10 seconds"));
      co2Serial.flush();
      return STATUS_NO_RESPONSE;
    }
  }
  if (debug) Serial.println();

  // The serial stream can get out of sync. The response starts with 0xff, try
  // to resync.
  // TODO: I think this might be wrong any only happens during initialization?
  boolean skip = false;
  while (co2Serial.available() > 0 && (unsigned char)co2Serial.peek() != 0xFF) {
    if (!skip) {
      Serial.print(F("MHZ: - skipping unexpected readings:"));
      skip = true;
    }
    Serial.print(" ");
    Serial.print(co2Serial.peek(), HEX);
    co2Serial.read();
  }
  if (skip) Serial.println();

  if (co2Serial.available() > 0) {
    int count = co2Serial.readBytes(response, 9);
    if (count < 9) {
      co2Serial.flush();
      return STATUS_INCOMPLETE;
    }
  } else {
    co2Serial.flush();
    return STATUS_INCOMPLETE;
  }

  if (debug) {
    // print out the response in hexa
    Serial.print(F("  << "));
    for (int i = 0; i < 9; i++) {
      Serial.print(response[i], HEX);
      Serial.print(F("  "));
    }
    Serial.println(F(""));
  }

  // checksum
  byte check = getCheckSum(response);
  if (response[8] != check) {
    Serial.println(F("MHZ: Checksum not OK!"));
    Serial.print(F("MHZ: Received: "));
    Serial.println(response[8], HEX);
    Serial.print(F("MHZ: Should be: "));
    Serial.println(check, HEX);
    temperature = STATUS_CHECKSUM_MISMATCH;
    co2Serial.flush();
    return STATUS_CHECKSUM_MISMATCH;
  }

  int ppm_uart = 256 * (int)response[2] + response[3];

  temperature = response[4] - 44;  // - 40;

  byte status = response[5];
  if (debug) {
    Serial.print(F(" # PPM UART: "));
    Serial.println(ppm_uart);
    Serial.print(F(" # Temperature? "));
    Serial.println(temperature);
  }

  // Is always 0 for version 14a  and 19b
  // Version 19a?: status != 0x40
  if (debug && status != 0) {
    Serial.print(F(" ! Status maybe not OK ! "));
    Serial.println(status, HEX);
  } else if (debug) {
    Serial.print(F(" Status  OK: "));
    Serial.println(status, HEX);
  }

  co2Serial.flush();
  return ppm_uart;
}

uint8_t MHZ::getLastTemperature() {
  if (isPreHeating()) return STATUS_NOT_READY;
  return temperature;
}

byte MHZ::getCheckSum(byte* packet) {
  if (debug) Serial.println(F("  getCheckSum()"));
  byte i;
  unsigned char checksum = 0;
  for (i = 1; i < 8; i++) {
    checksum += packet[i];
  }
  checksum = 0xff - checksum;
  checksum += 1;
  return checksum;
}

int MHZ::readCO2PWM() {
  // if (!isReady()) return STATUS_NOT_READY; not needed?
  if (debug) Serial.print(F("-- reading CO2 from pwm "));
  unsigned long th, tl, ppm_pwm = 0;
  do {
    if (debug) Serial.print(".");
    th = pulseIn(_pwmpin, HIGH, 1004000) / 1000;
    tl = 1004 - th;
    ppm_pwm = 5000 * (th - 2) / (th + tl - 4);
  } while (th == 0);
  if (debug) {
    Serial.print(F("\n # PPM PWM: "));
    Serial.println(ppm_pwm);
  }
  return ppm_pwm;
}

The forth library I looked at was MHZ16_uart by Intar it had been updated recently, was quite lightweight, had timeouts, checksum & packet header/footer validation.

/*
  MHZ16_uart.cpp - MH-Z16 CO2 sensor library for ESP-32
  by Intar BV
  version 0.1
  
  License MIT
*/

#include "MHZ16_uart.h"
#include "Arduino.h"


#define WAIT_READ_TIMES	100
#define WAIT_READ_DELAY	10

// public

MHZ16_uart::MHZ16_uart(){
}
MHZ16_uart::MHZ16_uart(int rx, int tx){
	begin(rx,tx);
}

MHZ16_uart::~MHZ16_uart(){
}

#ifdef ARDUINO_ARCH_ESP32
void MHZ16_uart::begin(int rx, int tx, int s){
	_rx_pin = rx;
	_tx_pin = tx;
	_start_millis = millis();
	_serialno = s;
}
#else
void MHZ16_uart::begin(int rx, int tx){
	_rx_pin = rx;
	_start_millis = millis();
	_tx_pin = tx;
}
#endif

void MHZ16_uart::calibrateZero() {
	writeCommand( zerocalib );
}

void MHZ16_uart::calibrateSpan(int ppm) {
	if( ppm < 1000 )	return;

	uint8_t com[MHZ16_uart::REQUEST_CNT];
	for(int i=0; i<MHZ16_uart::REQUEST_CNT; i++) {
		com[i] = spancalib[i];
	}
	com[3] = (uint8_t)(ppm/256);
	com[4] = (uint8_t)(ppm%256);
	writeCommand( com );
}

int MHZ16_uart::getPPM() {
	return getSerialData();
}

boolean MHZ16_uart::isWarming(){
	return millis() <= _start_millis + PREHEAT_MS;
}

//protected
void MHZ16_uart::writeCommand(uint8_t cmd[]) {
	writeCommand(cmd,NULL);
}

void MHZ16_uart::writeCommand(uint8_t cmd[], uint8_t* response) {
#ifdef ARDUINO_ARCH_ESP32
	HardwareSerial hserial(_serialno);
	hserial.begin(9600, SERIAL_8N1, _rx_pin, _tx_pin);
#else
	SoftwareSerial hserial(_rx_pin, _tx_pin);
	hserial.begin(9600);
#endif
    hserial.write(cmd, REQUEST_CNT);
	hserial.write(MHZ16_checksum(cmd));
	hserial.flush();
	
	if (response != NULL) {
		int i = 0;
		while(hserial.available() <= 0) {
			if( ++i > WAIT_READ_TIMES ) {
				Serial.println("error: can't get MH-Z16 response.");
				return;
			}
			yield();
			delay(WAIT_READ_DELAY);
		}
		hserial.readBytes(response, MHZ16_uart::RESPONSE_CNT);
	}

}

//private

int MHZ16_uart::getSerialData() {
	uint8_t buf[MHZ16_uart::RESPONSE_CNT];
	for( int i=0; i<MHZ16_uart::RESPONSE_CNT; i++){
		buf[i]=0x0;
	}

	writeCommand(getppm, buf);
	int co2 = 0, co2temp = 0, co2status =  0;

	// parse
	if (buf[0] == 0xff && buf[1] == 0x86 && MHZ16_checksum(buf) == buf[MHZ16_uart::RESPONSE_CNT-1]) {
		co2 = buf[2] * 256 + buf[3];
	} else {
		co2 = co2temp = co2status = -1;
	}
	return co2;
}	

uint8_t MHZ16_uart::MHZ16_checksum( uint8_t com[] ) {
	uint8_t sum = 0x00;
	for ( int i = 1; i < MHZ16_uart::REQUEST_CNT; i++) {
		sum += com[i];
	}
	sum = 0xff - sum + 0x01;
	return sum;
}

It ran second time on one of my Arduino devices (after I figured out how to configure the serial port pins) and though intended for an ESP8266 device this is the library I will field test.

#include <MHZ16_uart.h>

//Select 2 digital pins as SoftwareSerial's Rx and Tx. For example, Rx=2 Tx=3
MHZ16_uart mySensor(4,5);

void setup()
{
  Serial.begin(9600);

  mySensor.begin(4,5); 
}


void loop() 
{
  if ( !mySensor.isWarming())
  {
    Serial.print("CO2 Concentration is ");
    Serial.print(mySensor.getPPM());
    Serial.println("ppm");
  }
  else
{
    Serial.println("isWarming");
  }
  
  delay(10000);
}

This was just a sample of the libraries I found on GitHub if I missed a good a library contact me via the comments.