Sds011 data

Set of code and utilities to read air particle metrics from SDS011 sensor.

The code includes a utility that writes raw data from the sensor in a set of files.

The code also includes a tool to upload all data to an InfluxDB instance, possibly also grouping it by specific duration.

Installing

go install github.com/wojciechka/sds011data/cmd/sds011-reader
go install github.com/wojciechka/sds011data/cmd/sds011-influxdb-writer

Usage

Retrieving data from sensor

$ sds011-reader \
  -device /dev/ttyUSB0 \
  -dataDirectory /path/to/datadir

The command above will write data to specified directory, using the file format described below.

The directory /path/to/datadir must already exist.

Uploading data to InfluxDB

$ sds011-influxdb-writer \
  -dataDirectory /path/to/datadir \
  -stateFile /path/to/datadir/.writer-state \
  -influxToken "(token)" \
  -influxOrg "(org)" \
  -influxBucket "(bucket)" \
  -groupBy "1m" \
  -influxTags "tag1=value1,tag2=value2,..." \
  -wait

File format

The data read from sensor is kept in a directory, using flat file structure, grouped in multiple files.

Data for each specific day is stored using YYYYMMDD filename format, such as 20060102.

Each file keeps data for a specific second as 4 bytes. The format of the data matches the format returned by SDS011 sensor. The first two bytes contain value of PM2.5 readout in low endian notation, multiplied by 10. The last two bytes contain value of PM10 readout in low endian notation, multiplied by 10.

For example for value of 04 01 fe 00, 0x0104 indicates current value of PM2.5 - the value being 260, therefore the real value for PM2.5 result is 26.0. The raw value for PM10 is 0x00fe, which is 254 in decimal notation. The real value of the readout for PM10 is 25.4 in this case.

Offset to each data for each second is 4 multiplied by number of seconds elapsed since midnight. An appropriate formula is offset.Second() + offset.Minute()*60 + offset.Hour()*3600)*4.

Any data that is missing is filled with the four bytes being 0xffffffff - as in certain cases a readout of 0.0 for both PM2.5 and PM10 is possible, but values of 6553.5 are very unlikely. Also, as that level would exceed limits by around 100 times, the concern at that point should not be storing of the data, but implications of the values being so high.