Achieving precision analog to digital conversion in test and measurement applications
The control loops of industrial automation equipment rely on a constant stream of data on which to act. Most data comes from our analog world and is collected by different sensor types. Analog sensor data needs to be converted into the digital domain before control systems can use it to adjust the feedback loop.
Instrumentation systems take the raw sensor data, amplify it, condition the signals to remove unwanted artefacts, and then analog to digital converters process the data ready for digital control systems. To achieve conversion accuracy, signal conditioning circuits must keep any system-generated errors to an absolute minimum since it can generate false or inaccurate sensor operations and readings.
This article delves into instrumentation and test applications used for industrial applications. We explain the analog to digital conversion process and some of the sources of noise that can impact the accuracy of conversion, including noise from the power supply. Power protection using a programmable electronic fuse is also discussed.
Closed-loop industrial control systems
Control loops are all around us. Our domestic heating system is an excellent example of a closed control loop. A wall-mounted thermostat senses the ambient temperature and turns the heating on or off according to the desired temperature.
Figure 1 highlights the basic principles of a closed-loop system. The process is the action of heating and maintaining the home at the desired temperature, and the controller is your central heating control unit. The final control device is your boiler or furnace that pumps hot water around radiators or blows hot air through ducts. The measuring device is the temperature sensor.
Figure 1 - The basic concept of a closed-loop control system (source EBV)
The process might be more complex in an industrial context, perhaps involving sensing multiple variables (pressure, flow rate, and movement) and several controlling devices (motors, actuators, safety gates, etc.), but the basic concepts remain the same.
Instrumentation system architecture
A perfect control loop, however, is difficult to achieve. For example, the controller switches off the boiler once the thermostat has reached the desired temperature in our home heating scenario. In reality, there is still latent heat in the radiators, so the ambient temperature continues to rise. If the controller also manages the air conditioning, it might activate a cooling action, resulting in an oscillatory control loop and an unstable system. For many industrial processes, parameters such as temperature need to be finely controlled, so sensor resolution accuracy is crucial. Also, industrial environments tend to be electrically noisy, so mitigating noise from causing poor measurements is imperative.
Figure 2 illustrates the major functional blocks of an industrial instrumentation and test system. The analog to digital conversion process is perhaps one of the most important aspects of any instrumentation system. Measurement of temperature, humidity, air pressure, flow rates, and many more parameters are analog values. Control systems are based around microcontrollers and microprocessors operating in the digital domain, so analog sensor readings require conversion into a digital representation.
Figure 2 - The major functional analog building blocks of an instrumentation and test system (source ST)
Another critical aspect of any instrumentation measurement circuit is the power supply. Maintaining a low-noise stable power supply output to the conversion and signal conditioning functions is essential for maintaining control loop stability.
Analog to digital conversion
An analog to digital converter (ADC) is at the heart of any measurement system. The conversion of the analog input signal into a digital output involves two processes: quantisation and sampling.
Figure 3 - The function of an analog to digital converter (source EBV)
Quantisation: The conversion resolution of the ADC determines the minimum analog value of each digital bit. An 8-bit ADC, for example, provides the ability to represent the analog signal in 256 steps. Consider you are converting an analog signal that constantly varies between zero volts and positive one volts. The least significant bit (LSB) represents 3.9 mV, which is the minimum measurement resolution. However, a 16-bit ADC yields 655,536 discrete level steps and, for our example, an LSB representation of 15.2 µV. Figure 4 highlights the quantisation process and how converter resolution determines the range of analog values represented by the digital output's LSB.
Figure 4 - The process of quantisation and the importance of LSB resolution (source EBV)
Sampling: The sampling rate determines how faithfully the digital output represents the analog input signal. A slow sampling rate maintains a digital output stream that closely represents the input for an analog signal that changes slowly. However, a significantly higher sampling (conversion rate) ADC should be specified for a high bandwidth signal. See Figure 5.
Figure 5 shows how the sampling rate influences the digital output values. In this example, the digital outputs do not fully represent the analog value changes through the timeline (source EBV)
Analog to digital converter methods
There are several different analog to digital conversion methods. Each has its pros and cons; a summary follows:
Flash: This approach uses a comparator for each digital bits signal level; consequently, it offers fast conversion, performing conversion within one clock cycle. Unfortunately, integrating hundreds of comparators into a single IC can be costly, and also it exhibits a high input capacitance.
Pipeline: A pipeline uses multiple conversion stages, offering a reasonable resolution and high dynamic range. However, there is a propagation delay encountered with each stage.
Successive Approximation Architecture (SAR): This method compares the input voltage against a succession of small input voltages. It introduces no pipeline delay, but conversion accuracy may suffer from comparator noise. An example of a SAR ADC is the ST ADC120, an 8-channel, 12-bit analog to digital converter. The eight single-ended inputs are individually selected by the input multiplexer - see Figure 6 - and passed to the track and hold register.
Figure 6 - The internal architecture of the ST ADC120 8-channel 50 ksps to 1 Msps 12-bit ADC (source ST)
The ADC120 has a sampling rate of up to 1 Mega samples per second (Msps) and passes data to the host via SPI. The ADC features low power consumption, consuming only 6.6 mW from a nominal 3.3 Vcc supply rail. The analog input range is from zero volts to Vcc, and the maximum Vcc is 3.6 V.
A recent addition to the ST family of SAR-based analog to digital converters is the ADC1283. This low-power 12-bit ADC has eight multiplex single-ended channels and suits conversion applications from 50 ksps to 200 ksps. Digital output options include binary and SPI.
Sigma-Delta: This conversion approach delivers a precision result for low input bandwidth applications, but latency may limit high bandwidth applications. An example is the ST ISOSD61 a 16-bit sigma-delta-based converter available in either a single-ended TTL/CMOS or LVDS interface variant. The converter offers galvanic isolation of up to 6 kV peak between the analog side and the digital interface and provides a maximum range of +/- 320 mV and up to 25 MHz operation. See Figure 7
Figure 7 - The internal architecture of the TTL/CMOS interface ST ISOSD61 16-bit isolated sigma-delta converter (source ST)
The analog to digital conversion process may introduce noise, resulting in erroneous digital output values. Noise generated by non-linearities are some of the potential internal sources, but input-referred noise, where noise on the analog input causes the converter to jump between digital level steps, may also be encountered.
Low noise power regulation
A popular method of delivering a low-noise power source to measurement circuits is using a low-dropout regulator (LDO). LDOs are linear regulators with very low dropout voltages, this allowing the output voltage to be much closer to the input voltage than with classic linear regulators. LDOs function by operating a pass element (NMOS or PMOS) in the linear mode to reduce the input voltage down to the desired output voltage. An error amplifier senses the output voltage and compares it to an internal bandgap reference voltage, the resulting voltage is used to continuously adjust the resistance of the pass element to ensure the output voltage is correct.
As LDOs do not rely on a switching operation in their power conversion they are significantly less noisy than buck and boost regulators. However, the downside is that using a MOSFET operating in its linear mode for conversion results in comparatively higher power losses.
Figure 8 showcases the internal architecture of a typical LDO, the ST LDLN050.
Figure 8 - The internal functional block diagram of the fixed output version of the ST LDLN050 500 mA low dropout linear regulator (source ST)
The LDLN050 accommodates an input voltage range of 2.7 V to 6.5 V and offers an adjustable output voltage from 1.25 V to 6 V. It can deliver a regulated, low noise supply to a load up to 500 mA. Packaged in a standard industrial DFN8 footprint measuring 3 mm x 3 mm, the LDLN050 is ideal for supplying any circuit that is sensitive to noise, such as analog to digital converters, signal conditioning analog front-ends, and microcontrollers. The LDLN050, with its low power characteristics, typically 48 µA, suits use in battery-powered applications.
Overload protection
Protecting a circuit from experiencing excessive current flows and other overload conditions is essential for any design. Fuses have been utilised for decades but suffer from their inability to be reset manually. Poly fuses initially solved this issue by resetting automatically after some time but suffer from the fact that their resistance increases every time they are triggered, which in turn can cause a significant voltage drop on the bus they are meant to protect.
Programmable electronic fuses are becoming a far more convenient and popular method of integrating protection into a design. An example is the ST STEF01, an 8 V to 48 V fully programmable universal electronic fuse (e-Fuse) - see Figure 9. The STEF01 offers far more capabilities than a conventional fuse. In addition to monitoring output current for overload conditions, it can also detect overvoltage situations, which, if detected, results in the output current being limited to a safe value as configured by the application. If the overload condition continues, the e-Fuse isolates the load from the supply. The e-Fuse can be reset from a microcontroller host or set into an auto-retry mode. In addition to the STEF01's programmable overcurrent and overvoltage monitoring capabilities, it also features a thermal protection function and an adjustable soft-start time, and furthermore its performance is not affected by repeated trigger events.
Figure 9 - The internal architecture of an 8 V to 48 V fully programmable universal electronic fuse (e-Fuse) (source ST)
Selecting essential analog components for industrial instrumentation applications
Industrial instrumentation functions demand accurate and reliable analog circuits and components. The nature of closed-loop control systems, from chemical processing to industrial robots used on automotive production lines, requires loop stability with minimal unwanted noise artifacts.
This article highlights some of the many analog components available from STMicroelectronics suitable for industrial applications, where low noise, low power, and accurate analog to digital conversion are critical.
About Author
Karl Lehnhoff joined EBV Elektronik in 2008 as a regional application manager. From 2011 till 2019 he headed the renewable energies segment, today called city and infrastructure, as a director. He has served in his current role since March 2019. Previously he held multiple different roles in R&D, field application engineering and field sales. He has a degree as Dipl.-Ing. (FH) in electronics from University of Applied Sciences in Dortmund.
About Author
Milan Ivkovic
Milan Ivkovic received his Dipl-ing. diploma in 2000 and MSc diploma in 2007 at the University of Be...
ebv content library/articles/analog-to-digital-conversion-in-test-and-measurement-applications
Analog to digital conversion in test and measurement applications | EBV Elektronik
