Sensing and motion ICs drive robotic applications

Sensing and motion ICs drive robotic applications

Trends in sensors and sensor control electronics

BY ROGER LEVINSON
Vice President and General Manager
Precision Products Group
Intersil
www.intersil.com

Robotics in various forms has stimulated the imagination of many science fiction writers as well as investment bankers over the last forty years. In applications such as computer-assisted surgery, automobile safety and auto-park systems robotics is becoming commonplace. Industrial automation, the early automation leader, has driven big advances in robotics/factory automation that aided both factory safety and production efficiency. Advances in these segments have been fueled by billions of dollars of investment in four major technology sectors: sensors, transducers, motors, and control electronics.

The IC industry meets the challenge of these complex applications with a broad range of sensors and interface chips. For example, robotic systems can require A/D converters with 8-, 10-, 12-, or 16-bit resolution, and sample rates ranging from 125 ksamples/s to 1 Msample/s.

The ISL29028A ambient and infrared light-to-digital converter comes in an eight-lead optical dual flat plastic package.

The motion control of a multi-axis arm often requires 14- or 16-bit resolution and conversion speeds of 1 Msample/s or more are needed to achieve good position accuracy at high slew rates. Less complex devices such as assembly line actuators may use slower 8-, 10-, or 12 bit converters. SAR-type A/D converters offer numerous solutions for robotics applications, with devices ranging from 8 to 14 bit resolutions and conversion rates of 125 ksample/s to 1 Msample/s.

Other typical measurement needs include current, voltage, and thermal safety monitoring. While these may be fast-changing signals, they do require a high sampling rate in a multiplexed, high-channel-count system.

Simple sensors and smart sensors

Robotic systems view their environment through sensors. The simple act of picking up an object depends upon many sensors. The arm containing the picking element must find the object. Typically image sensors locate objects; however the object may exhibit other material behaviors such as magnetism or heat that can be detected with other sensor types. Motion control for the arm typically involves accelerometers, along with continuous position feedback from another sensor.

Using simple sensors with all processing done in a centralized MCU provides a straightforward structure, but runs the risk of inability of the processor to calculate fast enough or service interrupts quickly enough. The advantages of centralized processing also include centralized power requirements and algorithm development. This type of system may enhance flexibility, but may also limit maximum performance.

The use of smart sensors moves the decision making process to the point of interest, in this case the arm and the “the picker,” and thus provides opportunities for improved capabilities, in addition to multiple events that are easily controlled. However, limited sensor capabilities often restrict possible advanced robotic performance and the system might not be as easy to upgrade/modify. Using distributed control, the system design becomes more complex, but the removal of a central bottleneck may mean the design’s capabilities can be significantly enhanced.

Sensor interfaces

Technology companies continue to develop and deploy devices that assist in moving the field of robotics to the next level. In some cases the sensor and much of the control and processing electronics can be integrated into a monolithic solution.

Proximity sensors are perhaps the most important of all types. They can be found in consumer and communications products ranging from ATM and vending machines, to security systems, and leading-edge personal computers.

Proximity sensors are used in advanced mobile handsets to disable screens when a user is on a call, preventing accidental hangup or muting. They are used in safety systems, intrusion detection, and positioning.

An example of a smart proximity sensor is the ISL29028A ambient and infrared light-to-digital converter with a built-in IR LED driver and I2 C interface. It uses two A/D converters for concurrent measurement of ambient light and proximity and has a flexible interrupt scheme designed for minimal microcontroller loading.

Sensor support ICs

Interface chips for today’s smart sensors include very-low-power and low-noise signal-conditioning elements. High-input-impedance instrumentation amps such as the ISL28274, provide the required rail-to-rail inputs and outputs with extremely low input-bias current and 100-dB CMRR for the strain and pressure sensing used in tactile robotic applications. It costs $1.84 each in lots of 1,000.

Sophisticated multiaxis robotic arms need several kinds of sensors.

Chopper-stabilized amplifiers are another important contributor to sensor interface design. The industry has many versions available from many manufacturers. One example, the ISL28133, is optimized for single-supply operation from 1.65 to 5.5 V, with only 18 μA quiescent current and 8 μV max input offset voltage. It has offset TC of just 0.075 μV/°C max, and rail-to-rail inputs and outputs.

The amp features PSRR of 138 dB and CMRR of 138 dB with operation from –40° to 125°C and a 400-kHz bandwidth. This chip costs less than $1 ea/1,000 and has the good noise specifications needed for this type of application: 1.1 μVp-p typ. (0.01 Hz to 10 Hz).

As robotics systems become more sophisticated, feature sets grow while space and power consumption requirements become more demanding. Physical area constraints are driving significant growth in application-specific feature sets, accuracy specifications, and innovative semiconductor packaging techniques. Overall, the opportunities for sensors and high-performance analog solutions increase in conjunction with advances in robotics solutions. ■

For more on sensing and motion ICs, visit www2.electronicproducts.com/Sensors.aspx