What place on the property gets the most total sunlight? As the sun moves across the sky there will be shadows created based upon trees, clouds, and other obstructions. As the season progresses and the sun angle changes the amount of sunlight hitting certain places where plants are being raised will also change.
Illumination sensors (TLS2591) were placed in three locations for a comparative evaluation. This sensor was selected because it includes the UV range which is used by plants and it has a high dynamic range. While it is interesting to see the raw sensor measurements, it does not address very well the question of total sunlight being received over time.
To make a total sunlight visualization the integral of the individual measurements are needed. When looking at the daily integral the answer becomes clear that for this specific time of year there is quite a bit of difference between the three locations and day to day total illumination also varies. The "greenhouse" location suffers from trees that can filter the sunlight, especially in the morning and evening. The "WestLux" location has the least blockage at this time of year, but as the season progresses I expect it to suffer as the sun angle changes.
To get the integral of the sensor data I use the mcsMQTT Edit tab feature to create Accum Device for this sensor. Two choices are provided for this use case. One is to reset at midnight and the other to let the integral run indefinitely. The no-reset option would provide the answer for the total sunlight that hit each of the three location throughout the entire season. The reset-at-midnight option gives better visibility on the daily behaviors and will require additional processing at the end of the season if manual visualization at various times of the season does not provide the answer.
I had considered to "calibrate" the three sensors so that they would provide equivalent illumination measurements starting now. That is the Expression textbox where the calibration is shown to use the raw reading (1.0 divisor). One calibration technique was to look at the maximum lux measurement for each of the three and then factor them so that the max values would all be the same. Another technique was to look at the typical measurement at a sunny solar noon where obstructions should not exist at all locations. In the end I elected to use the raw data as the three sensors were the same and used the same mounting technique.
The data was sampled at 5 minute rates using Wemos D1 Mini and Tasmota. Finer resolution could be used to improve the quality of the measurement as clouds pass and trees sway. For better visualization of the samples the Low Pass Filter could have been used to smooth out the peaks and noise, but visualization of the raw data was not important and adding a low pass filter ( values < 1.0) would tend to decrease the accuracy of the integral which is really what was of interest.
The data is being collected in InfluxDB for long term storage retention.
Illumination sensors (TLS2591) were placed in three locations for a comparative evaluation. This sensor was selected because it includes the UV range which is used by plants and it has a high dynamic range. While it is interesting to see the raw sensor measurements, it does not address very well the question of total sunlight being received over time.
To make a total sunlight visualization the integral of the individual measurements are needed. When looking at the daily integral the answer becomes clear that for this specific time of year there is quite a bit of difference between the three locations and day to day total illumination also varies. The "greenhouse" location suffers from trees that can filter the sunlight, especially in the morning and evening. The "WestLux" location has the least blockage at this time of year, but as the season progresses I expect it to suffer as the sun angle changes.
To get the integral of the sensor data I use the mcsMQTT Edit tab feature to create Accum Device for this sensor. Two choices are provided for this use case. One is to reset at midnight and the other to let the integral run indefinitely. The no-reset option would provide the answer for the total sunlight that hit each of the three location throughout the entire season. The reset-at-midnight option gives better visibility on the daily behaviors and will require additional processing at the end of the season if manual visualization at various times of the season does not provide the answer.
I had considered to "calibrate" the three sensors so that they would provide equivalent illumination measurements starting now. That is the Expression textbox where the calibration is shown to use the raw reading (1.0 divisor). One calibration technique was to look at the maximum lux measurement for each of the three and then factor them so that the max values would all be the same. Another technique was to look at the typical measurement at a sunny solar noon where obstructions should not exist at all locations. In the end I elected to use the raw data as the three sensors were the same and used the same mounting technique.
The data was sampled at 5 minute rates using Wemos D1 Mini and Tasmota. Finer resolution could be used to improve the quality of the measurement as clouds pass and trees sway. For better visualization of the samples the Low Pass Filter could have been used to smooth out the peaks and noise, but visualization of the raw data was not important and adding a low pass filter ( values < 1.0) would tend to decrease the accuracy of the integral which is really what was of interest.
The data is being collected in InfluxDB for long term storage retention.