My impression was also that my Airlink is also under-reading, but it looks like it isn't that consistent. For lower levels it seems more noticable.
My Airlink PM10 data isn't that different from my Nova SDS011, but for PM2.5 there's quite some difference. And the Airlink values for PM10 and PM2.5 are much closer together than for the SDS011. Please note that the sensors aren't at the exact same position, and the SDS011 is reading the data every 2min.
pm2.5.png (81.04 KiB) Viewed 3757 times
pm10.png (98.53 KiB) Viewed 3757 times
pm10-w.png (63.71 KiB) Viewed 3757 times
Unfortunately there's no good calibrated sensor nearby, so I can't tell which one is more accurate.
Graham,
for integrate the NOVA SDS011 to MeteoBridge look at this Thread: viewtopic.php?f=61&t=15158
Don't forget to create a user defined Station after saved the script into MB-Scripts Folder.
best regards
Mike
woolfg wrote: ↑Sat Oct 10, 2020 10:30 pm
I had seen that but I dont know how to get the data from my raspberry pi (which has the sensor connected to a USB port ) into the meteobridge
How does the meteobridge know where to look
I have two scripts; one on the raspberry pi collecting the data and saving it in discrete files the other on the meteobridge nano in the /scripts folder.
There are three steps to make all this work:
1) Ensure that your Raspberry Pi is running a html service to provide a basic website. I used lighttpd (https://pimylifeup.com/raspberry-pi-lighttpd/)
2) Use whatever NOVA SDS011 script and add a few lines to output to the website some files accessible over http(s)
3) Use the user-defined station on the meteobridge with the second script to read the files created in #2 and use the data.
Here are the relevant parts of the two scripts:
1) The bits you add to your Nova SDS011 reading script. The PM2.5 and PM10 values in my script are stored in an array called values[]. Before I use them I ensure that they are multiplied by 10 and converted to strings. This is because I spotted that when Meteobridge reads them they are divided by 10.
pm25d = str(int(values[0]*10))
pm10d = str(int(values[1]*10))
with open("/var/www/html/aqi25.data","w") as aqifile:
aqifile.write(pm25d)
aqifile.close()
with open("/var/www/html/aqi10.data","w") as aqifile:
aqifile.write(pm10d)
aqifile.close()
2) The content of the aqijmf.plugin that you put on your meteobridge in the /scripts folder. Call yours whatever you want. My Raspberry PI can be accessed on the same LAN as the Meteobridge. Hence the call to 10.10.10.202. You could easily use a domain name depending on how you've setup your RPi
#!/bin/sh
#
while true;
do
aqi25=$(wget -qO- http://10.10.10.202/aqi25.data)
echo "air1 $aqi25"
aqi10=$(wget -qO- http://10.10.10.202/aqi10.data)
echo "air2 $aqi10"
sleep 300
done
Screenshot 2020-10-11 at 11 October 2020 13.32.11 pm.png (60.25 KiB) Viewed 3684 times
My live data looks like this:
Screenshot 2020-10-11 at 11 October 2020 13.27.58 pm.png (15.54 KiB) Viewed 3684 times
That's it. You now see your values in the live data.
If you want to use these values as your main AQI values, because say, you don't have a Davis Airlink or PurpleAir or any other device, then map these to Air0PM for PM10 and Air1PM for PM 2.5. These will then work using the AirQuality Davis modules that Brian has provided in his Weather34 template. I have both sensors running in comparison and the Airlink, which is additional sensor2 uses the mapping to Air0PM, Air1PM (and AirPM2) as you can see below:
Screenshot 2020-10-11 at 11 October 2020 13.46.53 pm.png (178.91 KiB) Viewed 3680 times
How are you collecting data today from your SDS011? Yes, it's connected to your RPi via USB, as is mine, but you must be running some kind of script (Python, Bash, etc.) to run the sensor and regularly capture the readings.
It is to this script that you add my sample code, not any of the lighttpd files. My full script (aqi.py) to run the sensor is below:
This does xx:00 and xx:30 minutes of every hour, day, week, month. It also outputs any messages and errors to a log file.
The meteobridge aqijmf.plugin runs continuously. It is the sleep(300) in the plugin code that waits 5-minutes (300 seconds) before taking the next reading. The plugin does not align to any specific time slot. I could wait much longer but if I do then I might end up with a much bigger delay after the RPi updates the data.
