We did not calibrate the sensor, nor has the St Marys Lane installation as a whole ever been calibrated.
(We do however provide here a link to the
DEFRA graphs of PM2.5 etc. for Barnsley Road, against which our measurements can be compared).
The SDS011 sensor used is only specified as being accurate to the maximum of +/- 15% and 10μg/m3.
Furthermore, this specification only applies at 25C and 50% humidity, so what happens outside this range is not specified.
The definition of pm2.5 quoted by the UK government's assessment of affect size,
specifies "an inlet of a size selective sampler with a transmission efficiency of 50% at an aerodynamic diameter of 2.5 micrometres".
This implies that discrimination between pm2.5 and larger particles is mechanical and may not be perfect.
The laser detection system in the St Marys Lane sensor on the other hand implies discrimination on some kind of "optical size",
which may treat for example, liquid droplets or different coloured or shaped particles quite differently.
The maximum working humidity of the SDS011 is specified as 70%, and the device is thought to overestimate the pollution when the humidity is high.
The sensor has been reported to produce occasional spikes.
Indeed, on 2019-11-17 Luftdaten recorded a single-measurement spike from our sensor of around 261 pm10 and 161 pm2.5.
A Google search revealed this
(scroll down to page 30),
and the suggestion that such spikes corresponded to certain relative humidity "set-points".
While not understanding how such a (bug?) could arise, we assume that any large single-measurement spikes should be ignored.
Sensor life is specified at only 8,000 hours or around 1 year, and may drift before or after this period.
(Although the measuring interval is set to 145 seconds, and measurements are
said to take 20 seconds
which should increase its life somewhat).