Creating a Synthetic Instrument with Virtual Instrumentation Technology

The demand for fast evolution and increased flexibility of electronics systems in many different industry and applications has driven the trend for increasing software content of electronics systems. In measurement and automation, the prevailing trend over the past 20 years has been towards measurement instruments that define their capability through software. Virtual instrumentation, which emerged in the mid-1980s, has been at the forefront of this trend. Recently, the US Department of Defense has articulated their desire for more flexible, software based test systems through an initiative called Synthetic Instrumentation. This paper will define both virtual and synthetic instrumentation, and provide an example of building an RF synthetic instrumentation system with virtual instrumentation technologies.

Virtual Instrumentation Defined

Virtual Instrumentation (VI) is defined as:

A software-defined system, where software based on user requirements defines the functionality of generic measurement hardware.

A virtual instrument shares many of the same functional blocks as its traditional counterpart, the standalone box instrument, but differs primarily in the ability of the end user to define the core functionality of the instrument through software. Where a traditional instrument has vendor-defined embedded firmware, a virtual instrument has open software defined by the user. In this way, the virtual instrument can be reconfigured for a variety of different tasks or completely redefined when an application’s needs change. As will be discussed, a Synthetic Instrument is a type of virtual instrument; currently synthetic instruments are being defined specifically for RF stimulus and measurement within military test systems.

The benefits of software-defined virtual instruments include:

• Increased system flexibility through reconfiguring software
• Increased system longevity by adapting to future needs
• Lower system size by creating multiple software personalities on shared measurement hardware
• Lower system cost through hardware reuse
• Ability to solve unique system requirements not addressed by exiting traditional instruments

  Virtual Instrumentation Software
  

Software is critical to a virtual instrumentation system. It is, after all, the user configurability through software that differentiates a virtual instrument from its traditional counterpart. Creating a virtual instrument requires software with several key attributes:

1. Connectivity to a breadth of I/O hardware
2. A comprehensive set of built-in measurement algorithms
3. The ability to created heterogeneous processing systems to take advantage of technologies such as DSPs and FPGAs.

Because the needs of measurement and automation systems are so diverse, no single bus or I/O standard can meet every need. For example, USB is well suited for applications requiring easy desktop connectivity while internal PC busses like PCI and PCI Express provide the highest performance in latency and throughput. And while many applications are well served with multifunction hardware devices such as a PC-based data acquisition device, others may require higher speed or higher resolution modular instruments. Virtual instrumentation software must be able to integrate this diverse set of I/O seamlessly into a single system. Well-architected software also includes an abstraction layer called measurement and control services which provides portability of software routines across devices and busses. Virtual instrumentation software should also be able to integrate traditional instruments into a hybrid system. This is valuable for two reasons: First, many systems must take advantage of existing measurement equipment to save costs; second, there may be highly specialized requirements that are met by a particular traditional instrument. A comprehensive set of I/O drivers and a well architected measurement and control services layer enables the user to create an integrated hybrid system.

Virtual instrumentation software must also contain a comprehensive set of measurement algorithms for facilitating the creation of user defined measurement and automation systems. At the lowest level, these algorithms should include basic mathematics tailored for measurement systems. For example, frequency domain measurements require transforming time domain data into the frequency domain through the Fast Fourier Transform (FFT). In addition to basic mathematics, users should expect higher-level application-specific libraries. For example, for sound and vibration applications, functions like THD measurements, SINAD, and order tracking are often required for validation and test. While it possible to create any of these with the right mathematical building blocks, it is impractical; the availability of these functions within a software development tool can save significant time in building a system. Similarly, for communication applications, complex digital modulation and demodulation, signal quality measurements, and spectral analysis are all useful tools for creating a virtual instrument.

Finally, virtual instrumentation software should be able to create heterogeneous processing application from a single software development environment. A heterogeneous processing application is one that integrates multiple types of processing elements, for example, a microprocessor, an FPGA, and a DSP. Each of these devices has attractive properties for different types of applications. For example, the inherently parallel architecture of an FPGA makes it ideal for multi-loop control applications, while the shear horsepower of a Pentium-class processor is the highest performance for most vector-based analysis. A software environment that can target multiple processor types and provide a degree of transparency between targets is ideal for maximizing performance while reducing development time. For example, it may be desirable to prototype a measurement algorithm on a host (PC-based) processor, and then deploy parts of it to an FPGA onboard an I/O module for maximum system throughput.

Virtual Instrumentation Hardware

While virtual instrumentation software should be able to integrate hybrid systems comprised of both generic virtual instrumentation hardware as well as traditional instrumentation, there are several hardware attributes that make a compelling virtual or synthetic instrumentation hardware platform. These include:
1. A general-purpose hardware architecture to address the broadest set of applications.
2. A high-speed connection between the hardware and the VI processing element(s).
3. Modularity so that parts of the system can be upgraded as needs evolve.

A primary benefit of virtual instrumentation is the flexibility that comes through reconfiguring a measurement and automation system in software. In order to maximize the degree of software reconfigurability in a system, the hardware should be designed to be as generic as possible. For analog measurement, virtual instrumentation hardware is responsible for digitizing the signal; all other processing for creating a measurement from the digitized signal is accomplished in software. Therefore, the capability of digitization hardware can be plotted according to bits of resolution versus the frequency of sampling, as shown in Figure 2. While the ultimate vision for virtual instrumentation is to use a truly universal measurement device that can do both high resolution measurements, as well as high-speed measurements, in practice, there are tradeoffs in cost and in current semiconductor capability. It is possible, though, to architect systems based on a relatively small number of generic measurement devices and cover a large percentage of requirements through software routines.


Once a signal is digitized in a VI system, it must be transferred over a data bus to a processing element running the appropriate software routine. Because buses vary in their strengths, certain buses offer better performance for particular applications than others. When evaluating bus performance, two important factors to consider are latency and bandwidth. Latency measures the delay of transmission of data, while bandwidth measures the rate at which data is sent across the bus, typically in MB/s. Lower latency improves the performance of applications that require a large number of small commands or data sets to be transferred. Higher bandwidth is important in applications such as waveform generation and acquisition. Figure 3 compares the latency and bandwidth of various instrumentation buses. Note that improving or increasing bandwidth moves up while improving or decreasing latency moves to the right. High-speed or high-channel count applications requires a high-speed data bus with considerable bandwidth to keep up with the system’s requirement. Without the necessary bandwidth, instrumentation must instead resort to embedding the measurement algorithms in the instrument as in traditional instruments, which sacrifices flexibility. Another approach is to add large amounts of on-board memory to the I/O module, which adds cost and yet still provides a typically small limit to the amount of data that can be captured or generated. For this reason, many busses used for control of traditional instrumentation, such as GPIB or LAN, are not ideal for high-performance virtual instrumentation systems. Currently, the highest performance bus used in instrumentation systems is PCI Express, which is capable of transferring data at up to 6 GB/s (for a x16 configuration) with low latency.

Finally, virtual instrumentation hardware should be modular, so that individual components can be added or upgraded as requirements evolve. This makes the system scalable and increases its overall longevity. Modularity can also promote the sharing of common system resources, which lowers the system’s cost. In a PXI (PCI eXtensions for Instrumentation) system, for example, common resources include a shared chassis, system controller, timing engine, and switching modules. Each of these shared resources can be interchanged based on requirements or upgraded as new technology is available. For example, when a system controller is replaced with a higher speed processor, the system’s throughput will increase, and when a timing controller with a more accurate reference is added, the timing accuracy of the entire system is increased.

The Department of Defense Synthetic Instrumentation Initiative

The US Department of Defense, as the largest single purchaser of test equipment in the world, is a key adopter of next-generation instrumentation technology. Maintaining their vast array of disparate, application-specific test equipment has proved to be a significant and costly challenge. Recently, the DoD has begun articulating the need for a more flexible, software-centric approach to building test equipment. A report to congress from the DoD Office of Technology Transition in February 2002 stated, “Recent commercial technology allows for the development of synthetic instruments that can be configured in real time to perform various test functions….A single ‘synthetic’ instrument can replace numerous single function instruments thereby reducing the logistics footprint and solving obsolescence problems.(1)

The DoD has created a standards body called the Synthetic Instrument Working Group (SIWG) who’s role is to define standards for interoperability of synthetic instrument systems. The SIWG defines a synthetic instruments (SI) as:

A reconfigurable system that links a series of elemental hardware and software components with standardized interfaces to generate signals or make measurements using numeric processing
techniques. (2)

The focus of the SIWG has been primarily on the SI concepts as applied to RF stimulus and measurement systems. The group has created a standard block diagram for an RF synthetic instrument, as shown in Figure 4. The functional blocks in this diagram are:

• The frequency translation devices (RF up and down converters),
• The IF (Intermediate Frequency) input and output, and
• The processing engine where the application-specific software is hosted.

To meet the performance of many RF applications, there must be a high-bandwidth connection between the IF devices and the processing engine where real time analysis is performed. For example, to digitize a 50 MHz wide RF signal requires at least 200 Mbytes/s of bandwidth (100 MS/s sampling rate at 2 bytes of resolution per sample). For both an input and output channel, this grows to 400 MB/s. And for increasingly common multi-channel, or MIMO (Multi Input, Multi Output) applications, the bandwidth required can quickly scale to multiple gigabytes per second.

An Example RF Synthetic Instrument

Commercial technologies are currently available for building systems using the synthetic instrument model. A platform that is often used to build these systems is PXI. PXI is a multi-vendor industry standard supported by over 70 companies with over 1200 available products, including modules from several vendors for building RF systems. As shown in Figure 5, PXI includes a shared high-speed backplane combined with shared timing and synchronization resources. The combination of a modular form factor, a high-speed bus, and integrated timing features makes PXI ideal for creating modular, software based systems.


Let’s take a look at an example application: a system for generating and measuring signals up to 2.7 GHz. In this example, we’ll stream the data back to the host for performing software-defined measurements. This is useful because it will provide the flexibility to change the software to generate entirely different type of modulated stimulus signals or different classes of measurements. We could also use this system as a software-defined radio to prototype a real-world communication system. A block diagram of this system is shown in Figure 6 and the actual hardware modules and software screens are shown in Figure 7.




In this system, the RF block downconverter translates the signal, with a real time bandwidth of up to 20 MHz, down to the input range of the IF digitizer. The IF digitizer uses an on-board digital downconverter, implemented in an FPGA, to filter and decimate the data. The data is then streamed over the high-speed PXI backplane to a host controller running a user-defined LabVIEW program. LabVIEW is a graphical development environment that uses a block diagram syntax for programming the system. This block diagram approach is ideally suited for creating RF systems such as this. The LabVIEW program can thus be reconfigured to change the personality of the instrument. For example, using built in spectral analysis functions, the system can operate as a real-time spectrum analyzer. Be adding demodulation functions, measurements such as modulation error ratio and even bit-error rate can be performed. And by changing the type of modulation performed, the system can test any type of standards-based communication signal that is within its frequency and bandwidth capability. The same basic components and capabilities are also available for generation using the IF generator and block upconverter.

This PXI-based synthetic instrument can be combined with other types of instruments to create a hybrid system to extend its capabilities. For example, when paired with VXI or standalone up and down converters, the frequency can be extended to 26.5 GHz and beyond. And because the IF generation and digitization is still done in the PXI modules, the system still can stream and process data at the high data rates needed by most applications. An example of a deployed hybrid PXI/VXI system is shown in Figure 8.

Virtual instrumentation software and hardware technologies go beyond the needs for military synthetic instruments. Because the demand for faster designs and quicker time to market is increasing in many different industries, the demand for the software-oriented approach of virtual instrumentation has expanded into many different domains. In particular, because the challenges to designed many complex systems mirror the challenges in testing these systems, virtual instrumentation technology is well-suited to system design. LabVIEW, for example, now includes the ability to target embedded systems such as FPGAs, DSPs, and embedded microprocessors. Further, it can program heterogeneous embedded systems that included multiple types of processing and intercommunication. These advances are also needed in measurement and automation where technologies like FPGAs can be used to increase system performance while maintaining software-defined flexibility. Using a unified software tool for embedded design, rapid prototyping and system verification and test will help address the challenge to bring new designs to market quicker than ever before.

References

1: Report to Congress on the activities of the DoD Office of Technology Transition, February 2002
2: SIWG Meeting #2 Statements and Definitions, 11 December 2004

Relevant NI Products and Whitepapers

National Instruments, a leader in automated test, is committed to providing the hardware and software products engineers need to create these next generation test systems.

Software:

• NI TestStand Test Management Framework
• LabVIEW Graphical Programming Environment
• Signal Express Interactive Measurement Software
  

Hardware:

• Modular Instruments (Oscilloscopes, Multimeters, RF, Switching, and more)
• Multi-function Data Acquisition
• PXI System Components (Chassis and Controllers)
• Instrument Control (GPIB, USB, and LAN)
  

Whitepapers
NI offers a Designing Next Generation Test Systems Developers Guide. This guide is collection of whitepapers designed to help you develop test systems that lower your cost, increase your test throughput, and can scale with future requirements. To read the entire developers guide, you can: Download the PDF (90+ page) version or view the web-version of the Designing Next Generation Test Systems Developers Guide.

Reader Comments | Submit a comment »