The ubiquitous use of sensors in our smart devices – from cell phones to industrial equipment and even medical devices – has increased the need for more intelligent sensor technologies that are more versatile, lower overall costs, and require fewer resources to develop and maintain.

Most analog sensor systems comprise three key elements: the analog sensor that measures a specific form of energy, the microcontroller (MCU) that processes the digital equivalent of the sensor’s signal, and between them is the Analog Front End (AFE) system (Figure 1). The AFE receives the sensor’s signal and converts/transforms it for the MCU to use, as in most cases the sensor output signals cannot be directly interfaced to an MCU.

The challenge associated with current AFE design approaches is the time-consuming trial-and-error tuning process, and the lack of flexibility and scalability to support multiple sensors from a single AFE. Moreover, many AFEs do not account for sensor drift or adjust for sensor trimming during production, which directly reduces the quality of the sensor. However, new fully configurable AFE technology is enabling designers to overcome these hurdles.

The importance of the AFE

The AFE itself performs multiple functions, depending on the application. One function of the AFE is to amplify signals that are too weak for the MCU to read. The AFE circuitry employs amplifiers to provide output voltages that are hundreds or up to thousands of times larger than the voltage provided by the sensor. This is typically done with op-amps that can vary widely in cost and power based on the required characteristics. Depending on the sensor characteristics, the AFE amplifier structure will vary. For example, if the sensor output is differential and low impedance, a simple differential input can be used. If, on the other hand, the sensor output is differential and high impedance, a more complicated instrumentation amplifier, with matching high-impedance inputs, may be needed.

Another function of the AFE is to filter unwanted frequency ranges from the sensor, for example, to satisfy the Nyquist limit or to remove a DC offset. This noise must be removed before the analog signal is converted to digital. The AFE must employ low-pass filter circuitry to block out high-frequency noise and/or employ high-pass circuits to remove lower-frequency noise.

A third function of the AFE is to convert signals from one signal type to another. For example, typical sensors output a voltage, but some output a current. The MCU ADC circuits do not accept current inputs, so such currents have to be converted to voltages before going to the MCU. This current-to-voltage conversion is performed by the AFE circuit, called a transimpedance (I/V) circuit, which also amplifies the resulting voltage to levels usable by the MCU.

Challenges to AFE designs

Most AFE circuits are custom designed to meet the electrical requirements of a particular system under development. Engineers must design the circuitry, select the appropriate ICs and passive components, then test and tune the resulting circuit and PCB layout. In many cases, this takes a trial-and-error method to calibrate the right analog circuit design. This iterative tuning process is time and resource consuming, adversely affecting development cost and time-to-market. In addition, the AFE is often difficult to simulate and must be adjusted because of specific component behavior, board layout, and nearby noise sources.

There is also limited or no scalability of the AFE circuitry to support multiple sensors, let alone multiple types of sensors (that is, different topologies). The AFE circuit is designed for one particular sensor, making it difficult to swap one sensor for another using the same AFE – even if they employ the same topology.

Finally, sensors need constant tuning either during production – adjusted for sensor trimming – or because they degrade over time and cannot easily be corrected after they are deployed in the field. Fixed-component AFE designs do not correct for sensor drift nor are they easily adjusted for sensor trimming. A software-supported design approach can help.

Let’s examine each of these challenges.

Configurable AFE eases calibration trial and error

Looking at the hundreds of different types of sensors available, one can observe common topologies and signal characteristic ranges and understand that having the ability to simply change the characteristics of the op-amps, or to dynamically change the gain values, will significantly reduce the complexity and reduce development time.

The Renesas Smart Analog technology is an example of a fully configurable AFE technology that allows for such capability. As Figure 2 shows, such technology includes five elements: three separate configurable amplifiers, an additional amplifier with sync detection capability, a general-purpose op-amp, a low-pass filter with variable cutoff frequency, and lastly, a high-pass filter with variable cutoff frequency.

The design engineer can create the desired custom AFE circuitry by simply setting the main parameters for these various circuit blocks, and then selecting the connections between these blocks. Three highly configurable amplifiers can be used to produce a tailored I/V transimpedance converter/amplifier, a noninverting amplifier, an inverting amplifier, a differential amplifier, or a summing amplifier. The chip can be custom configured to implement a range of signal amplification gains, and it provides an adjustable span of signal voltage offsets.

Additionally, the amplifiers in this IC can be configured to implement a single-channel, high-impedance instrumentation amplifier. This type of differential amplifier is essential for interfacing to high-impedance sensors such as piezoelectric types.

As the AFE takes care of amplifying/filtering/converting the signals from the sensor, the MCU (internal or external device) can analyze the AFE signals to dynamically change the gain values (that is, while the system is operating) to compensate for changes in ambient environment. This “closed loop,” self-adjusting AFE structure provides a more robust, intelligent sensor interface.

An integrated AFE+MCU device offers the additional benefit of automating the trimming process as it will read the signals from the AFE and compare that to the known parameters to make the necessary adjustments on the AFE, thereby cutting system production costs. In the same way, the MCU can automatically adjust the AFE gain to counteract the signal-generation deviations expected to occur over time as the sensor degrades.

Configurable AFE provides scalability

While configurability is important to reduce complexity and debugging time, another key design factor is scalability. An AFE with enough connection terminals to accommodate all the sensors typically needed eliminates the traditional requirement to have a separate AFE circuit for each sensor. Handling the entire array of sensors via one AFE helps shrink the circuit board and simultaneously decreases system component counts while reducing power consumption by as much as 20 percent. In fact, because of the simple interface of these AFEs – to just an SPI line and the ADC channels from the MCU – it is possible to connect to as many as 96 sensors using one MCU.

A software-supported design approach

Extreme configurability can come with the burden of tool complexity, so it is important to have a simple software-based design tool that can configure and customize the characteristics of the AFE for that specific application. Designers no longer need to understand the lowest level of the hardware, nor be analog experts when the AFE register values can be simply set, and the topology, gain/offset values, and characteristics can all be done in software.

Such a tool should run on a PC and provide an easy way for selecting typical sensor types, such as pressure, humidity, acceleration, impact, magnetic, and piezoelectric types – supporting multiple topologies and characteristics. The Smart Analog software provides this highly intuitive environment where designers easily set parameters, change topologies, do offset tuning, and have the ability to add filters and, of course, have access to the signal pins.

Because the tool itself already has libraries of different sensor profiles, it is easy for a systems engineer to have a starting point in their design. A graphical representation of the output signals from the AFE can be used to monitor systems with close-to-real-time feedback, which will make it very easy to make the AFE adjustments and tuning. All these features reduce complexity in development and thus reduce resource costs.

Once the configuration is set, the tool outputs a register file that can be used by the software on the MCU. The MCU stores the sensor settings in on-chip flash (nonvolatile) memory within its firmware, and when power is applied to the system, the MCU sends the stored settings to registers in the Smart Analog IC, reconfiguring that chip accordingly.

Simplifying the burden of AFE designs

The AFE is a critical, yet sometimes underappreciated component to a sensor system. The typical discrete approach of adding op-amps and filters, and trial-and-error soldering of resistors is not efficient and the cost of time in debugging and development easily outweighs the cost of adding an intelligent, MCU-based configurable AFE. But not all configurable AFEs are built the same. So, it is important to consider the flexibility and scalability of the AFE to support different types of sensors, and the intelligence to adjust “on the fly” or in the field. Simple, easy-to-use software tools can ease this process and can be used by even the non-analog experts on the team.

Wil Florentino is the Industrial Automation Marketing Manager at Renesas Electronics (North America), and product lead for Smart Analog. He has more than 15 years of experience in semiconductors with previous industrial segment roles at Xilinx and Altera. He has a Mechanical Engineering degree, and an MBA from Virginia Tech.

Renesas http://am.renesas.com/products/promotion/smart_analog/index.jsp

Follow: Twitter Facebook LinkedIn YouTube