The future of instrumentation – the software-designed revolution

December 9th, 2015, Published in Articles: EngineerIT

The role of software in test and measurement systems has changed dramatically over the past few decades. Today, software is the most critical core technology in modern, high performance measurement systems because it is the only thing that can abstract the fundamentally growing complexity of those systems.

However, simply running software on computer processors is not enough. The most challenging applications aren’t possible without engineers that can use software to specify and design the behavior of their own instrumentation. This ability to create software-designed instruments is at the heart of a revolution that is taking place in RF instrumentation and, more broadly, in the general test & measurement industry.

In the beginning – automating measurements

The role of software in test and measurement systems has advanced steadily since the 1970s when IEEE-488, also known as GPIB, standardised the interface for programmable instrumentation. Up to that time, taking measurements was largely accomplished with benchtop instruments, a pencil and a notebook. There were a variety of different proprietary instrument controllers and interfaces available, but their capabilities were rudimentary and used more by advanced users.

The role of software in test and measurement took a huge leap forward in the 1980s with the introduction of the original IBM PC and the first National Instruments (NI) GPIB interface board. With the use of PC software, engineers could use general-purpose PCs to create automated test systems that reliably and repeatedly acquired measurements, analysed those measurements, and presented results that could be shared widely.

From instrument control to measurement platform

In the later 1980s and into the 90s, a new concept of “virtual instrumentation” began to take hold in the test and measurement world. This concept revolutionised the role of computers – and especially software – in test and measurement systems. Rather than viewing a PC as simply a computer for automating measurements with GPIB, virtual instrumentation used the computer itself as a measurement platform. Moore’s Law ensured that the processing power of a PC quickly outstripped the technology used in stand-alone instruments. Further, computer memory, storage capacity, and graphics display capabilities far outpaced that of traditional instruments. Thus, a general-purpose PC quickly became a better computing platform than a traditional instrument.

Two critical elements were required to enable a virtual instrumentation system to meet, and ultimately exceed, the capabilities of traditional instruments – modular measurement hardware and software. On the hardware side, early computer plug-in boards offered fairly low-quality measurement capabilities compared to benchtop instruments, which relied on proprietary data converters. Demands of broad markets, such as consumer audio and wireless infrastructure, drove the development of off-the-shelf data converters that could perform high quality measurements when used on computer plug-in boards. Computer-based measurements took a big leap forward with the development of measurement-specific computing platforms, especially PXI (PCI eXtensions for Instrumentation), which was developed in the 1990s. PXI combines PCI computer technology with instrumentation-specific timing and synchronisation capabilities. Soon, PXI virtual instruments were solving some of the most challenging measurement challenges – including high-performance RF measurements.

Yet, it was software that really made virtual instrumentation possible. Not only was software required to acquire, analyze, and present measurements in a PC environment, but it had to do so in a sufficiently abstracted manner. In essence, abstraction software was required to enable engineers and scientists to efficiently solve their test and measurement challenges without having to be experts in computer science and architecture. First released in the mid-1980s, NI LabVIEW software set the standard for virtual instrumentation software and started the trend towards software playing a central role in modern test and measurement systems.

Enabling software-designed instrumentation with FPGAs

The next major step in measurement capabilities is being enabled with the inclusion of FPGA-based measurement hardware. Looking forward to the future, it’s important to consider that “instruments” in the traditional sense are no longer single-function measurement devices – but have instead become measurement systems. In addition, engineers are looking to instruments not only to test devices but also to design and prototype much larger systems.

FPGA is a key technology that is bringing new levels of performance to next-generation instrumentation. FPGAs offer substantial processing power (Figure 1). With the inclusion of FPGAs, there is now a software-based capability to push measurement functions deep into the hardware itself.

Fig. 1: FPGA is growing at a rate that even exceeds CPU’s.

Fig. 1: FPGA is growing at a rate that even exceeds CPU’s.

Many of today’s RF instruments already benefit from fixed-functionality FPGAs to execute tasks such as flatness correction, ADC linearisation, IQ calibration, and digital downconversion. Software-designed instruments, such as the NI PXIe-5644R vector signal transceiver (Fig.2), benefit from FPGA technology in an entirely new way because the FPGA is available to the user for customisation. For example, moving the instrument control and decision making from a PC to an FPGA can dramatically reduce measurement time in complex measurement systems. Also, this capability combined with advanced FPGA-based signal processing will enable instruments to function in a wider range of embedded applications as well.

Fig. 2: The new NI PXIe-5644R vector signal transceiver combines a vector signal analyzer, vector signal generator, high speed digital I/O and a user-definable FPGA in one 3-slot PXI module.

Fig. 2: The new NI PXIe-5644R vector signal transceiver combines a vector signal analyzer, vector signal generator, high speed digital I/O and a user-definable FPGA in one 3-slot PXI module.

System design software – the key to software-designed instrumentation

The right system design software tool is essential to pull together the computing and measurement technologies in today’s modular hardware. From its beginning as instrument control software, LabVIEW has evolved to be a comprehensive system design platform that allows engineers to create complex, high performance measurement systems. Engineers can use a common set of tools and languages to target applications on both processors and FPGAs, alleviating the need to know different languages and tools. LabVIEW provides a higher-level abstraction to work at the system level while also enabling engineers to create lower-level optimisations to address very high performance or complex requirements.

Multi-mode RF device characterisation

Faced with testing a new 802.11ac product, Qualcomm Atheros had to test their device under more operating conditions than ever before, resulting in more than an order of magnitude increase in measurement complexity. Using the FPGA-based NI vector signal transceiver and LabVIEW, they were able to design a test system that synchronizes digital DUT control with RF measurements. The resulting test system reduces overall test time by more than order of magnitude and enables engineers to observe device behaviour in multiple operating modes.

As we observe in Fig. 3, the traditional test instrumentation (left) was used to obtain an iterative set of measurements. Because the measurement time was very high, the test engineers had to choose a subset of operating points to characterise, resulting in essentially a “guesstimate” over the device’s operating characteristics.

However, by switching to an FPGA-based instrument approach, they improved measurement performance by 200 times, enabling the engineers to acquire all 300,000 operating modes in a single test sweep. The resulting characteristic curves, shown in the right figure of Fig. 3, showed much more detailed information about the device.

Fig. 3: With traditional instrumentation, approximately 40 points of meaningful WLAN transceiver data were collected per iteration. The speed increase of the NI PXI vector signal transceiver triggered full gain table sweeps to acquire all 300,000 points.

Fig. 3: With traditional instrumentation, approximately 40 points of meaningful WLAN transceiver data were collected per iteration. The speed increase of the NI PXI vector signal transceiver triggered full gain table sweeps to acquire all 300,000 points.

Instrumentation in embedded applications

A second class of applications for software-designed instrumentation is embedded communications and signal processing. While we have traditionally thought of instruments as measurement devices, modular, software-designed instrumentation allows engineers to use RF instrumentation in embedded applications as well. Today, growing numbers of engineers are using modular PXI instruments for embedded applications like spectrum monitoring, passive radar systems (like the one in Fig. 4), and even communications system prototyping and software defined radio. These applications require instrumentation to be increasingly smaller, modular, and have better access to deterministic signal processing targets. Communications system design software must be able to abstract increasingly complex systems, enabling engineers to implement existing and new communications algorithms and deploy those algorithms on processors and FPGAs.

Fig. 4:  Complex systems such as this passive RADAR system are designed and deployed with NI LabVIEW software and NI PXI hardware embedded into the final system.

Fig. 4: Complex systems such as this passive radar system are designed and deployed with NI LabVIEW software and NI PXI hardware embedded into the final system.

Looking towards a future of converged RF design and test

Software-designed instrumentation blurs the line between design and test like never before. One of the more intriguing opportunities will be the ability to share IP between design and test – whether that IP runs on a processor or an embedded FPGA. Using system design software like LabVIEW, engineers will be able to use the same tools to create a new communication protocol, and move that protocol to FPGA-based hardware for prototyping. Today, this is very challenging because of the disparity between the mathematics software used to create the algorithms and the design tools used to implement those algorithms. Advances in higher level synthesis and integration of multiple design models of computation will be required to fully enable engineers to achieve this seamless transition between design, implementation and test.

As a final thought, it is interesting to note that software in test and measurement has come a long way from its beginnings as simply a vehicle to automate a set of stand-alone instruments. Instead, software has become the heart of the instrumentation itself – enabling instruments to solve more difficult problems in both measurement and system design. In effect, automation is now an embedded capability of the system that is required to meet the complex measurement requirements that engineers face. Today’s software-designed measurement systems are just the first generation to deliver what are sure to be game-changing RF design and measurement capabilities for a long time in the future.

Contact  Bianca Powell, National Instruments, 011 805-8197, bianca.powell@ni.com

Related Articles

  • Quantum computing. The next big… wait, what is it really?
  • Accurate device characterisation for high bit rates
  • Advanced driver assist system for mining launched
  • EMI shielding ventilation panels
  • Load sensing by light