Digital light sensors – Using R for Photobiology

Abstract

Brief description of modern multichannel digital light sensors as components suitable for the construction of complete weather-proof sensors.

Code

library(ggspectra)
library(photobiologyWavebands)
library(photobiologySensors)

energy_as_default()
theme_set(theme_bw())

1 Modern digital light sensors

In the past, and still currently, most light sensors used for quantification of photosynthetically active radiation, for specific bands within PAR and for FR are constructed based on discrete silicon photodiodes and optical filters, delivering a low voltage analogue signal.

In recent years integrated circuits (ICs) based on CMOS technology including photodiodes, analogue amplification, analogue to digital conversion (ADC) and supporting standardized serial digital communication protocols have been developed for different purposes. Some of them are good enough to replace with advantage discrete photodiodes.

One of the key advances is the deposition of interference filters directly on the IC chip, selectively on individual photodiodes. The incorporation of signal amplification makes it possible to rely on smaller photodiodes and by implementing gain adjustments increase the dynamic range. Furthermore, the implementation of an ADC in the same chip and relying on digital communication avoids noise pickup in the wiring and makes possible the use of much simpler and cheaper circuits to acquire and log the data. Miniaturization also reduces power requirements.

Integrated digital sensors, even if calibrations are frequently applied during manufacture, aim to be cheap and thus their specifications can have wider wavelength tolerances under test conditions. However, some of these new IC sensors have temperature compensation and a temperature sensor also integrated resulting in extremely low and consistent dark noise readings and a large dynamic range.

A key supplier of this type of digital sensors is ams OSRAM that offers many different variations aiming at different applications. Vishay has also developed some sensors with a limited number of channels. All these sensors are extremely small and encapsulated as surface mounted devices (SMD) in very small packages (e.g., $2 \times 3 \times 1$ mm) with a tiny window.

When bought as components in quantity they are very cheap (3 to 15 €). Given their size, for prototyping and custom assembly they are easier to use if bought as “break out boards” (10 to 25 €), i.e., ready soldered on a small printed circuit board (PCB) to easy the task of connecting them. They cannot be used on their own, they need to be connected to a micro-controller or single-board-computer to acquire the data and control their settings.

The Swiss company YoctoPuce has developed USB modules containing a microcontroller, memory, an isolated power supply and a USB interface. These modules have a built in data logger and a webserver-based user interface as well as a programming API. Through a hub they can be accessed remotely through the Internet or LAN. These modules cost more than the bare sensor ICs (40 to 100 €) but provide a readily usable solution. Their Yocto-Spectral USB module based on ams OSRAM’s AS743 digital spectral sensor when used for measuring irradiance require the addition of a diffuser. This module has built-in LEDs for measuring reflectance.

Aranet has just released a weather proof sensor with a cosine diffuser and a LoRa interface that connects through a base station, likely based on a digital spectral sensor. This sensor is sold for 1000 € + VAT, and although the easiest to use being a ready made field sensor, it lacks flexibility in the precomputed data summaries or access to raw data from individual channels.

For all these sensors, the wavelength and channel sensitivity guaranteed in specifications have rather loose tolerances. In most cases electronic components are closer to the typical specifications than the tolerance limits.

Examples of promising digital sensors that could be useful in agriculture and other fields are given below.

UV Sensor VML6075 from Vishay

The VML6075 from Vishay is a digital UV sensor with two channels, named UVB and UVA in the documentation but closer to UVA1 and UVA2 in reality. Vishay provides an algorithm to compute UVI. This sensor is still available from distributors but its manufacture has been discontinued by Vishay.

Code

autoplot(sensors.mspct$Vishay_VEML6075, 
         w.band = UV_bands("CIE"), range = c(280, 450))

UV Sensor AS7331 from ams OSRAM

The AS7331 from ams OSRAM is a digital UV-radiation sensor with three channels, named UVC, UVB and UVA in the documentation. The band separation is rather good for computing UVI in sunlight and the usual artificial UV radiation sources.

Code

autoplot(sensors.mspct$ams_AS7331, 
         w.band = UV_bands("CIE"), range = c(200, 450))

Sensitivity of Si photodiodes to UV radiation decreases with decreasing wavelength, but the design compensates for this using a photodiodes with larger area to compensate.

Spectral VIS and IR sensor AS7343 from ams OSRAM

The AS7343 from ams OSRAM is a digital VIS and NIR sensor with 13 channels. This sensor seems like the most useful for measuring light from the perspective of plants as it seems possible to estimate PAR, ePAR and FR.

Code

autoplot(sensors.mspct$ams_AS7343, 
         w.band = VIS_bands())

Spectral VIS and IR sensor AS7341 from ams OSRAM

The AS7341 from ams OSRAM is a digital spectral VIS and IR sensor with 10 channels. The channels have wavebands of similar width suggesting that it could work well as a spectrometer for visible light. From the perspective of measuring light as relevant to plants, it lacks a far-red channel.

Code

autoplot(sensors.mspct$ams_AS7341, 
         w.band = VIS_bands())

Spectral VIS and IR sensor TSL2591 from ams OSRAM

The TSL2591 from ams OSRAM is a digital illuminance sensor with two channels, named ch0 and ch1 in the documentation. Illuminance can be obtained by subtracting the NIR sensed by ch1 from ch0.

Code

autoplot(sensors.mspct$ams_TSL2591, 
         w.band = VIS_bands())

2 Conclusion

Some of these very cheap sensors could be adapted to use in the measurement or estimation of quantities used in both agriculture practice and in research. To realize this potential methods for the reliable retrieval of the quantities of interest consistently in time and space need to be developed.

I am currently exploring possible uses of the Yocto-Spectral USB module in agriculture production and plant science research.

Reuse

CC BY-SA 4.0

--- title: "Digital light sensors" subtitle: "Miniature multichannel integrated circuits" date: 2025-08-19 date-modified: 2026-06-03 author: "Pedro J. Aphalo" keywords: [light sensors, multichannel, digital] categories: [measurement and control] editor: markdown: wrap: 72 code-fold: true format: html: resource-path: - ./sensor-images code-link: true code-tools: true abstract: | Brief description of modern multichannel digital light sensors as components suitable for the construction of complete weather-proof sensors. draft: false --- ```{r, message=FALSE} library(ggspectra) library(photobiologyWavebands) library(photobiologySensors) energy_as_default() theme_set(theme_bw()) ``` ```{r, include=FALSE} library(knitr) opts_chunk$set(fig.align = 'center', fig.show = 'hold', fig.width = 7, fig.height = 4, cache = FALSE, message = FALSE) options(warnPartialMatchArgs = FALSE) ``` ## Modern digital light sensors In the past, and still currently, most light sensors used for quantification of photosynthetically active radiation, for specific bands within PAR and for FR are constructed based on discrete silicon photodiodes and optical filters, delivering a low voltage analogue signal. In recent years integrated circuits (ICs) based on CMOS technology including photodiodes, analogue amplification, analogue to digital conversion (ADC) and supporting standardized serial digital communication protocols have been developed for different purposes. Some of them are good enough to replace with advantage discrete photodiodes. One of the key advances is the deposition of interference filters directly on the IC chip, selectively on individual photodiodes. The incorporation of signal amplification makes it possible to rely on smaller photodiodes and by implementing gain adjustments increase the dynamic range. Furthermore, the implementation of an ADC in the same chip and relying on digital communication avoids noise pickup in the wiring and makes possible the use of much simpler and cheaper circuits to acquire and log the data. Miniaturization also reduces power requirements. Integrated digital sensors, even if calibrations are frequently applied during manufacture, aim to be cheap and thus their specifications can have wider wavelength tolerances under test conditions. However, some of these new IC sensors have temperature compensation and a temperature sensor also integrated resulting in extremely low and consistent dark noise readings and a large dynamic range. A key supplier of this type of digital sensors is [ams OSRAM](https://ams-osram.com/products/sensor-solutions/ambient-light-color-spectral-proximity-sensors/) that offers many different variations aiming at different applications. [Vishay](https://www.vishay.com/en/optoelectronics/) has also developed some sensors with a limited number of channels. All these sensors are extremely small and encapsulated as surface mounted devices (SMD) in very small packages (e.g., $2 \times 3 \times 1$ mm) with a tiny window. When bought as components in quantity they are very cheap (3 to 15 €). Given their size, for prototyping and custom assembly they are easier to use if bought as "break out boards" (10 to 25 €), i.e., ready soldered on a small printed circuit board (PCB) to easy the task of connecting them. They cannot be used on their own, they need to be connected to a micro-controller or single-board-computer to acquire the data and control their settings. The Swiss company [YoctoPuce](https://www.yoctopuce.com/) has developed USB modules containing a microcontroller, memory, an isolated power supply and a USB interface. These modules have a built in data logger and a webserver-based user interface as well as a programming API. Through a hub they can be accessed remotely through the Internet or LAN. These modules cost more than the bare sensor ICs (40 to 100 €) but provide a readily usable solution. Their [Yocto-Spectral USB module](https://www.yoctopuce.com/EN/products/usb-environmental-sensors/yocto-spectral) based on ams OSRAM's AS743 digital spectral sensor when used for measuring irradiance require the addition of a diffuser. This module has built-in LEDs for measuring reflectance. Aranet has just released a weather proof sensor with a cosine diffuser and a LoRa interface that connects through a base station, likely based on a digital spectral sensor. This sensor is sold for 1000 € + VAT, and although the easiest to use being a ready made field sensor, it lacks flexibility in the precomputed data summaries or access to raw data from individual channels. For all these sensors, the wavelength and channel sensitivity guaranteed in specifications have rather loose tolerances. In most cases electronic components are closer to the typical specifications than the tolerance limits. Examples of promising digital sensors that could be useful in agriculture and other fields are given below. ::: callout-note # UV Sensor VML6075 from Vishay The VML6075 from Vishay is a digital UV sensor with two channels, named UVB and UVA in the documentation but closer to UVA1 and UVA2 in reality. Vishay provides an algorithm to compute UVI. This sensor is still available from distributors but its manufacture has been discontinued by Vishay. ```{r} autoplot(sensors.mspct$Vishay_VEML6075, w.band = UV_bands("CIE"), range = c(280, 450)) ``` ::: ::: callout-note # UV Sensor AS7331 from ams OSRAM The AS7331 from ams OSRAM is a digital UV-radiation sensor with three channels, named UVC, UVB and UVA in the documentation. The band separation is rather good for computing UVI in sunlight and the usual artificial UV radiation sources. ```{r} autoplot(sensors.mspct$ams_AS7331, w.band = UV_bands("CIE"), range = c(200, 450)) ``` Sensitivity of Si photodiodes to UV radiation decreases with decreasing wavelength, but the design compensates for this using a photodiodes with larger area to compensate. ::: ::: callout-note # Spectral VIS and IR sensor AS7343 from ams OSRAM The AS7343 from ams OSRAM is a digital VIS and NIR sensor with 13 channels. This sensor seems like the most useful for measuring light from the perspective of plants as it seems possible to estimate PAR, ePAR and FR. ```{r} autoplot(sensors.mspct$ams_AS7343, w.band = VIS_bands()) ``` ::: ::: callout-note # Spectral VIS and IR sensor AS7341 from ams OSRAM The AS7341 from ams OSRAM is a digital spectral VIS and IR sensor with 10 channels. The channels have wavebands of similar width suggesting that it could work well as a spectrometer for visible light. From the perspective of measuring light as relevant to plants, it lacks a far-red channel. ```{r} autoplot(sensors.mspct$ams_AS7341, w.band = VIS_bands()) ``` ::: ::: callout-note # Spectral VIS and IR sensor TSL2591 from ams OSRAM The TSL2591 from ams OSRAM is a digital illuminance sensor with two channels, named ch0 and ch1 in the documentation. Illuminance can be obtained by subtracting the NIR sensed by ch1 from ch0. ```{r} autoplot(sensors.mspct$ams_TSL2591, w.band = VIS_bands()) ``` ::: ## Conclusion Some of these very cheap sensors could be adapted to use in the measurement or estimation of quantities used in both agriculture practice and in research. To realize this potential methods for the reliable retrieval of the quantities of interest consistently in time and space need to be developed. I am currently exploring possible uses of the Yocto-Spectral USB module in agriculture production and plant science research.