Object detection commonly deploys optical sensors. They provide precise non-contact detection at short response times. Furthermore, they master both short and long distances and fit where space is a constraint. However, there is one drawback: they operate on visible light of the same spectral range as artificial light or sunlight. Consequently, such light may cause detection errors in light barriers and photoelectric sensors. Most tricky is the fact that in virtually no place will lighting conditions be the same. An optical sensor proving error-free detection at the manufacturer’s may struggle with different lighting conditions at the customer’s.
A common and conventional principle for optical distance measurement and object detection is triangulation. The principle is quite simple: the longer the distance to the object, the acuter the angle of the reflected sensor light. The sensor determines this angle based on the position at which the reflected light strikes the receiver, the photo diode array (PDA). In Figure 1, light source Q generates a light beam focused by lens L1 on optical axis A1. The diffuse object-reflected light at position x1 is refocused by lens L2 and strikes the PDA. Distance x1 is calculated by position y1 and the light distribution in the beam striking the PDA, in relation to focal length f, basic distance b and the resulting angle α1. The light beam striking pixel no. 3 will result in position y2 and angle α2 out of which the new distance x2 is calculated.
The signal striking the PDA comprises the diffusely reflected light beam generated by light source Q together with the present ambient light. Ambient light includes a content of both DC (constant frequency) and AC (alternating frequency). A very high DC content means direct solar radiation. The AC content may refer to artificial lighting.
LEDs
More and more common LEDs operate on a similar or larger frequency spectrum than optical sensors. This may cause detection errors. In the event of a detection error, the optical sensor is not in a position to identify whether the signal received is emitted by the sensor light or an interfering source.
LED lamps are the most challenging sources of ambient light, since their AC content ranges from 30kHz to 150kHz. Optics, electronics and algorithms are applied to remedy such unwanted interference by ambient light. Known measures are optical and analogue filters as well as pulsed operation of light source Q. However, this would not suffice if for example a sensor with 25kHz sampling rate is exposed to interfering light of 100kHz, as shown in figure 2. The sensor’s considerably lower sampling rate would entail undersampling, the so-called aliasing effect. Mapping of the LED spectral content is in the low frequency range, which may cause measuring errors.
However, thanks to their low power consumption, LED lamps experience an ever-increasing use in new or retrofit lighting installations. Consequently, ambient light conditions respectively interfering factors change.
Tests have proven optical sensors of different brands prone to interference by LED light of various frequencies. Yet, as long as the time- relevant switching behaviour is maintained, troubleshooting is quite simple. But it’s completely different with optical sensors that adapt their internal measuring cycles to the interference frequency. This in turn may entail process cycle times being no longer being adhered to, or system shutdown or, worst case, machine crash. Here, troubleshooting is extremely difficult since the root cause is not immediately evident, nor can the malfunction be remedied automatically. Searching the root cause throughout the entire optical sensor system and interfering light sources is complex and time- consuming.
Switching errors in optical sensors are often due to object properties (material, geometry and surface), to process conditions (detection range, process speed), installation situation (angle dependencies and interfering objects) or incorrect parameterisation/manipulation. Furthermore, lighting situations can change with daylight intensity. This is even worse with sensors featuring the mentioned adaptive response.
Once the cause and hence the interfering light source has been identified, corresponding counteractions can be taken. Depending on process speed and sensor parameterisation options, additional filter settings may already remedy the problem. Aligning the sensor position would be another quite simple solution. Should these actions not be an option due to process conditions and installation situation, the situation gets expensive and time-consuming, since only alternatives may be cover or replace the sensor.
OPTICAL SENSORS
Light barriers/photoelectric sensors are the most common optical sensors.
More than 30 years ago, Baumer decided to develop its own ASICs (application specific integrated circuits) for certain key components. The challenge was exploiting the properties of interfering light and sampling with limited microcontroller processing capacity to extract unwanted interference from the desired signals (see figure 3). Here, interfering light is compensated for by identifying the properties, such as intensity and position, on the photo diode array (PDA), as well as the current frequency.
One example of a Baumer light barrier and photoelectric sensor is excel by Aline, a CMOS sensor (complementary metal-oxide semiconductor). It interacts with optical and analogue filters, with a specially-designed digital filter, a controller concept and ambient light algorithms. By measuring brightness and the measuring light pulse of the sensor’s light source, interference is extracted from measuring respectively desired sensor signals. The filter and control system together with ambient light algorithms will suppress interference frequencies and interfering jumps.
This article is an edited version of a white paper, ‘Ambient light as interference source for optical sensors – Challenges and solutions’ published by Baumer (via www./is.gd/jofemu). It has been republished with permission.