4 ms·

We at AirGradient maintain popular indoor and outdoor open source air quality monitoring projects [1] and gained quite a lot of experience on air quality measur

We at AirGradient maintain popular indoor and outdoor open source air quality monitoring projects [1] and gained quite a lot of experience on air quality measurement. The hard part of these projects is not connecting electronic modules and get some air quality data out of it but to ensure that the data is accurate and this is where I have some questions. The success of these kind of projects stands and falls with the trust that people have in the data it generates. 1) The built uses the Sensirion SPS30 PM sensor. As per Sensirion assembly guideline the maximum allowed air flow around the sensor is one meter per second. This corresponds to 3.6 km/h or 2.2 m/h. A car normally travels a multitude of this speed. I did not see anywhere in the built instructions how they manage the airflow inside the enclosure to be within this limit. In my opinion detailed airflow simulations inside the enclosure are essential. 2) Temperature and humidity has strong impacts on the accuracy of PM measurements and it is important that these are accounted for. This is already a challenge for stationary low cost sensors but having this sensor mounted on the roof of a car will probably give a wide vaiety of different heat exposures. I did not see how this is accounted for. 3) I am not sure if there are large advantages with mobile sensors compared to a high density stationary network of low cost sensors. Air pollution, especially on roads fluctuates a lot with peak times. So a measurmenet taken at 8am at a busy junction might be quite different a few hours later. How long do you keep past measurements valid for a certain location? I applaud all open source projects in this space and it looks like a lot of effort went into this project so it would be great if above concerns could be addressed/clarified from somebody in the team. [1] https://www.airgradient.com/open-airgradient/kits/ https://www.airgradient.com/open-airgradient/kits/ [2] https://sensirion.com/media/documents/7990F04A/616544B0/Sensirion_Particulate_Matter_AppNotes_SPS30_Mechanical_Design_and_As.pdf https://sensirion.com/media/documents/7990F04A/616544B0/Sens...

jrockway 4y ago
> Temperature and humidity has strong impacts on the accuracy of PM measurements How do you guys accurately measure humidity electronically?
- ahaucnx 4y ago
  We mostly use the Sensirion SHT temperature and humidity sensors that are quite accurate.
  - jrockway 4y ago
    Ah interesting. I have a weather station that uses that chip (not sure if 40 or 45), and the instructions say to never expect great information from it until you calibrate it against a sling psychrometer. (It is certainly not the worst humidity sensor I have. The one in my humidifier is consistently 10% off from the other sensors. The BME280 seems to yield similar results to the SHT45. But without a calibrated reference, I'm just guessing, and assumed that was the state of the art of electronic humidity sensing. A man with two inaccurate watches that agree with each other is never on time, or something like that.)
stephen_g 4y ago
Yeah, the idea of mounting them on top of vehicle seemed a bit strange to me too, given how the particulate sensors work. Perhaps they could use the GPS and just sample for a bit when the vehicle has been stopped for a few seconds? I didn't think of the temperature issue - that does seem like a tricky problem too...

jrockway 4y ago

> Temperature and humidity has strong impacts on the accuracy of PM measurements How do you guys accurately measure humidity electronically?

ahaucnx 4y ago
We mostly use the Sensirion SHT temperature and humidity sensors that are quite accurate.
- jrockway 4y ago
  Ah interesting. I have a weather station that uses that chip (not sure if 40 or 45), and the instructions say to never expect great information from it until you calibrate it against a sling psychrometer. (It is certainly not the worst humidity sensor I have. The one in my humidifier is consistently 10% off from the other sensors. The BME280 seems to yield similar results to the SHT45. But without a calibrated reference, I'm just guessing, and assumed that was the state of the art of electronic humidity sensing. A man with two inaccurate watches that agree with each other is never on time, or something like that.)

stephen_g 4y ago

Yeah, the idea of mounting them on top of vehicle seemed a bit strange to me too, given how the particulate sensors work. Perhaps they could use the GPS and just sample for a bit when the vehicle has been stopped for a few seconds? I didn't think of the temperature issue - that does seem like a tricky problem too...