Internet Explorer is not a supported browser for TI.com. For the best experience, please use a different browser.
Video Player is loading.
Current Time 0:00
Duration 15:29
Loaded: 0.00%
Stream Type LIVE
Remaining Time 15:29
 
1x
  • Chapters
  • descriptions off, selected
  • en (Main), selected

Welcome to the TI Precision Lab series on light sensors. My name is Alex Bhandari-Young. And I am an applications engineer at Texas Instruments. In this video, we will classify the different types of common light sensors based on a number of parameters including spectral response and sensor output type. We will begin with the photodiode, which is the starting point for all light sensors.

A photodiode is a typical diode that is placed in a transparent package so that the PN junction is exposed to outside light. Photons of high enough energy striking the depletion region of the photo diode can transfer their energy to electrons, allowing the electron to escape the depletion region.

If the photo diode is reverse biased, the voltage causes the escaping electron to flow and can create a current that is measurable called the reverse current.

In the last video, we introduced the irradiance measurement. The reverse current created is proportional to this irradiance from the monochromatic light source shining on the photodiode sensitive area.

The simplest light sensor uses the reversed biased photodiode and a resistor to convert the photodiode current into a voltage. While this can work for very simple applications, there are some problems with this approach. The current produced by the photodiode is quite small, typically ranging from nanoamps to microamps. So the voltage across the resistor is small. The next slide will discuss how an amplifier may address this.

Additionally, photodiodes are not stable over temperature and drift can cause errors. This challenge will also be discussed further. A transimpedance amplifier or TIA is used to convert the photo current to voltage in many analog output light sensors. A simple circuit for a transimpedance amplifier is shown.

Here we see an integrated photodiode and amplifier example. Here the photodiode is the current source being amplified. And a reverse voltage bias is applied to the photodiode. Analog output light sensors can also incorporate temperature compensation to provide more stable readings across temperature.

There are still a number of challenges with analog output sensors. A microprocessor would require an analog to digital converter or ADC to read the voltage output from this type of sensor. Additionally, the analog signal needs to be carefully routed to the ADC to avoid noise from other signals on the board. This is further complicated if multiple sensors are needed or the signal needs to pass through a long cable. The next two sensor types address these challenges.

If the application only requires an indication of whether the light intensity is above or below a certain threshold, a threshold sensor may be used. Threshold sensors use a comparator to compare the light on the photodiode against a reference threshold and simply output 0 or 1 to indicate brightness above or below the reference threshold.

This sensor acts like an on/off switch, activating a response when the light level is above or below the threshold, such as switching on night vision or raising an alarm. The threshold is often set using external components.

With these sensors, the microprocessor can use a general purpose I/O pin to read the sensor output, eliminating the need for an analog to digital converter. The 0 or 1 digital value is also less susceptible to noise, though multiple sensors still require multiple signals to be routed. These sensors may also include temperature compensation.

For applications requiring an irradiance value rather than a threshold, there is another sensor type that solves a number of challenges with using an analog output sensor while still providing the irradiance value as output.

Digital output sensors integrate the analog to digital converter into the sensor and directly output a digital value. This type of sensor is convenient to use and can also save cost because it can directly connect to the host microprocessor without the need for additional components.

Digital sensors often use interfaces such as I squared C, which is also less sensitive to noise than an analog signal. Additionally, multiple sensors can be connected on the same bus, resulting in fewer signals to route. These sensors also often compensate for temperature variation in the photodiode. The ADC can integrate the reading over a period of time, which decreases noise in the measurement.

All of these processing steps result in a more accurate sensor compared to the discrete photodiode. These sensors can also allow thresholds and exposure time to be set on the device depending on the needs of the application, making them quite versatile.

So far, we have seen light sensors classified by the processing chain and output type. We now generalize the light sensors covered so far as shown in the diagram. These all involve a photo diode and a processing step that generates an output. In the following slides, we will build on this diagram to introduce other types of light sensors.

The most common photodiode types are silicon and indium gallium arsenide, abbreviated InGaAs. The characteristic spectral response of each is shown. We can see silicon photodiodes are sensitive over the visible IR and UV wavelengths. InGaAs photodiodes are sensitive to a much larger area of the IR region.

Often applications require specific sensitivity to certain types of light which cannot be achieved using bare photodiodes. In these cases, an optical filter is typically deposited on top of the photodiode using a special process. This changes the spectral response of the sensor to meet the application requirements. Typically, filters are used with silicon diodes for applications sensing visible and near IR light.

This architecture can also be duplicated in a multi-channel sensor which uses different filters to measure different wavelength bands of light. In the following slides, we will first discuss the ideal way to measure different spectra of light and the associated challenges that typically make the architecture shown the most practical to use in everyday devices. We will then explore the input side of this architecture consisting of filter and detector and why this is important for different types of light sensors. We will also introduce multi-channel sensor types.

Photodiodes measure irradiance, as seen earlier in this video. The most accurate way to measure light would be to measure the irradiance across the entire spectrum. This could be done by having many photo detectors each with a narrow bandwidth of detection. Another way this is done is by using a grading to split the light into individual wavelengths. This process is similar to the way a prism splits light and rainbows are formed in the sky.

The split light is directed onto a detector array. Since each wavelength falls on a different part of the detector, the power level for each wavelength can be measured within the bandwidth capabilities of the detector array. This arrangement is called the spectrometer.

The different spectra of light sources shown in previous videos were measured using a spectrometer. While spectrometers offer the best performance in terms of light sensing, this performance typically cost several thousand dollars. And spectrometers are bulky in size. Additionally, the grading is sensitive to the angle of incoming light.

In typical indoor and outdoor settings, light comes from many angles, meaning that spectrometers are mostly used in controlled laboratory environments where the light entry can be controlled. We will now discuss sensors that bring cost effective light sensing capabilities to everyday devices like phones, cars, and thermostats.

In order to measure many spectral bands in a small form factor, many channels with different band passed filters are typically integrated into a chip, eliminating the need for a diffraction grading. These spectral sensors typically have a wider bandwidth compared with spectrometers.

A hyperspectral sensor may have up to 20 channels, allowing the sensor to report power levels for up to 20 wavelength bands. This type of sensor could be used for differentiating between the different types of light sources covered in video two. However, the size and cost of the sensor will scale with the number of channels. Now we will look at application-specific sensors with fewer channels.

Multi-channel sensors will typically have between three and five channels. Common examples are XYZ and RGB sensors. The RGB color sensor is a three channel sensor that measures the red, green, and blue color content as defined by the RGB color space.

XYZ sensors match a CIE standard for defining the full space of colors able to be seen by the human eye. These sensors are useful for adjusting the color temperature of displays to match the environment for a more natural viewing experience.

The visible light sensors we introduced in the first video are single channel sensors with a photopic filter to match the human eye response. Here we see TI's OPT3001 sensor response introduced in video three as an example.

Since these sensors incorporate a photonic response, the output value is in lux. This response is very closely matching for all wavelengths and results in accurate lux measurements from the sensor. We have discussed in previous videos the application of this type of sensor in display brightness adjustment.

An expansion of the visible ambient light sensor is a two channel sensor measuring both visible and IR light. This sensor is useful for differentiating between different types of light sources without needing a mini channel sensor. For example, cars can use this sensor to differentiate between sunlight and artificial street lighting by comparing the ratio of light measurements between the two channels.

So far, we have seen application-specific light sensors where an added filter was needed. Adding engineered filters typically adds an extra processing step, which adds cost. As seen earlier, a bare photodiode has a broad spectral response, which is acceptable for many applications that do not require a targeted spectrum. An example response from TI's OPT3002 device is shown. This may be used in an ATM tamper detection application where the inside of the ATM is expected to be dark unless the ATM is opened.

In this video, we have seen different types of sensors for different applications. In following videos, we will discuss practical ways to integrate these sensors into a design. To find more light sensor technical resources and search TI products, please visit the link shown.

Thanks for taking the time to watch this video. Please try the following quiz. Question 1. Which of the following is a common material used for photodiodes? InGaAs or indium gallium arsenide is one of the two common types of photodiodes along with silicon mentioned earlier.

Question 2. What is the best way to measure the spectral output of a light bulb in the lab? In a lab setting, a spectrometer will give the most accurate spectral measurement with a narrower bandwidth compared with the other sensors listed.

Question 3. A smart watch application requires display brightness adjustment and differentiating sunlight from artificial light. What is the most practical sensor? A multi-channel device would be needed to differentiate sunlight from artificial light. Of the options, photopic plus IR is both the most cost effective and the best at differentiating sunlight from artificial light.

Question 4. If a photodiode spectral response does not meet an applications requirements, how is the response of the sensor chip typically varied to meet the requirements? A colored sheet made of a material such as glass can control light passing through to a degree but does not allow the same control as a chemically engineered filter. Additionally, off the shelf glass would not integrate into a sensor chip, as this requires a special process.

Photodiodes of different compositions do not give enough control over the spectrum to suit many applications as was shown earlier. Spectral response cannot be adjusted digitally as the photodiode outputs a current that integrates over the entire spectrum. Thus the correct answer is that a chemically engineered filter would be deposited on top of the photodiode using a special process.

This video is part of a series