System Solutions for Industrial Sensors/Field TransmittersDecember 1, 2011 By: Deepa Kalyanaraman, Texas Instruments Inc. Sensors
In this article, we review some of the latest trends and system considerations involved when designing sensors and transmitters for industrial automation applications.
Industrial sensors and field transmitters are an integral part of the factory automation and process control industries. Measurement accuracy and precision are critical to tightly controlled process loops, as well as for single-point measurement nodes. This article leads you through the design decisions—concerning the signal chain, power source, and interface—that can help or hinder you from getting the most out of your sensors and your system as a whole. You will see how to translate system demands into device specifications, and clearly understand the design decisions and tradeoffs involved in maximizing the impact of industrial sensor and field transmitter electronics in the process.
First, let's consider a typical factory floor. Most factory floors can be broadly segmented into a hierarchy of three levels: the enterprise level, the control level, and the field level as shown in Figure 1. The enterprise level is, in essence, the office network, the backbone of servers and PCs networked to log all of the data and perform higher-level functions. The control level takes us into nodes and clusters of programmable logic controllers (PLCs), human-machine interface (HMI) panels, and DCS/SCADA and similar systems. This equipment issues commands to, and collects the resultant data from, the lower levels, processing this into actionable information, which can then be used to define and drive the actions of the control loop. The lowest level in our hierarchy is the field level and it is perhaps the most critical level of the three, because the equipment in this level includes the sensors, transmitters, and actuators that perform the actual manufacturing tasks and relay back critical data.
Figure 1. Typical factory floor hierarchy
For example, in a pharmaceutical bottling plant, different constituent chemicals are mixed in the correct proportions to create a bottle of the latest drug. Every step in the process needs to be executed at the right temperature, pressure, and humidity conditions to ensure that the quality and characteristics of the resultant mixture are as desired. This requires the constant measurement of multiple variables, the timely transmission of data, and the instantaneous response of the equipment to the control signals. The various sensors, along with the field transmitters, are what allow any process to become a controlled process, and designing these systems to take measurements with a high degree of precision is critical to the functioning of any control loop.
Depending on the size of the factory and the complexity of the controls needed, the factory floor could have thousands of sensors. As such it is important to understand what a sensor looks like on paper to understand how it will act when put into your control loop. Figure 2 shows a snapshot of typical information from a sensor/field transmitter datasheet. While it can be hard to extract exact operational characteristics from the datasheet, it will give you a ballpark idea of how a device will perform, especially when compared to other sensors that might already be in use in the system. Datasheet information can be broken down into input, output, power, and isolation specifications. This example in Figure 2 is for a popular 8-channel temperature transmitter that can be configured in multiple ways, e.g., for 2- or 3-wire RTD, thermistor, or thermocouple inputs.
Figure 2. Typical sensor/field transmitter datasheet
Understanding Sensor Front Ends
When we take a closer look at the design of the sensor, you find that, irrespective of the type of physical parameter you are measuring, sensor front ends can be broadly classified into a few major categories: resistive, capacitive, magnetic/inductive, and current sense front ends. Sensors basically produce a change in resistance, capacitance, current, or voltage in response to a change in the measured variable. Figure 3 shows some common examples of different sensors and where they fall in these categories, and also lists some key attributes to keep in mind while designing with each of them. In addition, the table in Figure 3 includes amplifiers and data converters from Texas Instruments that are well-suited to each specific implementation.
Figure 3. Common types of sensors and significant attributes
If we take a look at the system block diagrams for different types of sensors, such as those shown in Figure 4, we see that only the input signal chain is unique to the type of sensor element. The remaining elements are often very similar, if not identical, from sensor to sensor.
The interface between the sensor element and the data processing circuitry begins with an amplification stage that may be integrated into the data converter (as a programmable-gain amplifier or PGA) or implemented as a stand-alone discrete device. The amplifier performs several functions, which may include amplification, attenuation, filtering, buffering, offset adjustment, or level shifting. Because the amplifier is a key element of the input signal chain, it is important to select one that is optimized for your sensor element.
The analog-to-digital converter (ADC) converts the conditioned signal to an output that can be fed directly into a processor or microcontroller (MCU). ADC selection involves several tradeoffs, one of the most important of which is resolution versus speed. The processor runs various system routines, calibration routines, and compensation algorithms to further process the collected information, filtering out and/or correcting for known system errors. Finally, the processed data are sent out through an isolation block into either traditional analog lines or a digital bus or wireless interface. The isolation is essential to avoid ground loops, to provide safety from high voltages and surges, and to reduce common-mode noise. Isolation voltage specifies the transient voltage rating across the barrier while operating voltage levels specifies the continuous working voltage across the isolation barriers, which are typically specified as Vrms or Vp-p.
When considering the output or communication stage of the sensor module, there are a plethora of options. While the traditional analog 4–20 mA output continues to dominate the sensor output configurations, newer digital bus standards such as Profibus, IO Link, and others are seeing increased acceptance. These digital fieldbus protocols offer advantages when it comes to maintenance and repair, in addition to providing seamless communication and improved interoperability. It takes only a few seconds to download new parameter settings into the sensor via the digital bus, removing the need for manual calibration by allowing on-the-fly correction and reducing system downtime when replacing failed sensors. The continuous monitoring and diagnostics available when using the digital protocols also support need-based rather than routine maintenance programs.
The power budget of the sensor module is the next most important aspect to consider. Understanding the power source is critical to the decisions you make when designing the sensor electronics. Industrial sensors are typically powered in one of three ways:
- Line-powered—powered by a dedicated VDC line,
- Loop powered—powered off the 4–20 mA loop, or
- Battery powered—powered by batteries for portable transmitters.
The power topology influences the design. For instance, loop-powered transmitters do not have a dedicated voltage supply and therefore all of the power for the sensor needs to be scavenged off the loop current, restricting the design to components that can run on a total of about 3–3.5 mA or less (the sum total power budget for all of the silicon). Although battery-powered topologies may give you more initial headroom, the most important consideration is the battery runtime. Designs for both loop- and battery-powered sensors need amplifiers with low quiescent current, low-power data converters, and ultra low-power processors. Line-powered transmitters typically have significant headroom in the power budget and do not impose such stringent conditions on device selection.
Regardless of the power source, the overall efficiency and the heat generated by power conversion play a key role when making design decisions for the confined environment inside a sensor. The DC/DC step-down converters, buck-boost regulators, and linear regulators selected for such low-power designs also need to have low quiescent current, and must be able to operate in industrial high-voltage realms.
The list of questions below makes a good checklist before you start designing your next industrial sensor module.
|Why are you taking measurements and collecting data?||
|What parameter are you measuring?||
|How many channels will you need?||
|What resolution and sampling rate do you need?||Consider the effective number of bits (ENOB), least-significant bit (LSB) size, and sampling speed. Note that you may oversample to filter noise and to perform error detection|
|What is your noise budget?||
|How much isolation do you need?||Consider the operating voltage, transient voltage, and common-mode transient immunity.|
|Output stage||Analog (4–20 mA) or digital (e.g., Profibus or IO Link)?|
|Power||Line- vs. loop- vs. battery-powered|
|Calibration||It's always a good idea to spend some time understanding how you want to calibrate your system, including the appropriate hardware vs. software routines.|
Most Read Articles