patricktokeeffe / 2022-biofiltro-logger Goto Github PK

If logger stops receiving new serial data (e.g. cable disconnected, user opens analyzer Setup menu, analyzer turned off), then most recent value is retained indefinitely. Program should populate destination variable with NAN upon failing to obtain new serial record instead.

N.B. since analyzers operate at approximately 1Hz (±1-2%) the logger main scan loop runs at twice the anticipated data rate (2 Hz = 500 msec) to avoid a scenario where main loop might begin after new serial data record is already received. Maybe drop this approach entirely and just run deal with occasional serial record being lost.

Sometimes, without explanation, the logger will show invalid data for a single variable even though all other serial data values were parsed correctly.

How could this be possible? To correctly interpret all values except for a single column? One possible explanation is corruption in the serial message - but to confirm or invalidate this would require logging the actual serial message as a string (~400 characters).

Example serial data stream:

Example screenshot, showing a spurious (garbage) value of 140671100 instead of the correct value, 140.671:

Variables for serial data analyzers are initialized to zero, but if analyzer is not sending data yet (e.g. turned off or cable disconnected), then those zeros are interpreted as valid data and get included in final data storage.

Table biofiltro_runs gets called from the main scan (500 ms) and Totalize() instructions apply a 1/2 scaling factor to compensate for main scan running faster than new data is queried. Except the weather station is only updated every five seconds so the scaling factor should actually be 1/10th (table called 5x2 times between successive measurements).

Alternatively, implement an inter-scan flag to ensure Totalize() is only called once, immediately after getting new measurements.