I'm building a particulate sensor using the Nova SDS011 PM sensor. It detects PM2.5 and PM 10. Papers like Potential and Limitations of the Low-Cost SDS011 Particle Sensor for Monitoring Urban Air Quality by Budde, et al., and "Field performance of a low-cost sensor in the monitoringof particulate matter in Santiago, Chile" by Tagle, et al., report that the sensor is more accurate for PM2.5 than PM10. I've also found this to be true of the two sensors that I have—comparing my readings against a professional PurpleAir sensor nearby, the PM2.5 readings I get are very similar to the PurpleAir sensor but the PM10 readings are often 1-8 µg/m^3 higher than the PurpleAir sensor gives. This makes sense: the Nova SDS011 seems to be a "good enough" small and inexpensive alternative to other more professional sensors.
However, does anyone know whether it's possible to correct for the Nova SDS011's inaccuracies?
Thank you for those links! I appreciate the information and will see what I'm able to pull off as a hobbyist and layperson. I'm not sure whether I can use an offset to correct my readings, since the sensor output has been always higher than an accepted value but not by a consistent margin. At least for now, I'll focus on calibration. Thanks again!
Loquai defined a ratio between the calibrated SDS readings and the reference Grimm sensor readings, but I'd like to be able to use that calibration on an independent SDS sensor. Is it possible to simply omit a reference value and put any PM2.5 measurements from an independent SDS011 through a function like -.509*ln(PM10 / PM25) + 1.2203?
Or is this calibration not actually applicable without having a reference sensor nearby to measure the same particulate matter?
Also, what did Loquai mean when he wrote
Is he indicating that the calibrated SDS readings may be expressed as sqrt(PM2.5_SDS * PM2.5_Grimm)?
