Modular instrumentation set to meet new test challenges

4 min read

The world of test and measurement is changing; slowly for some, but at a rapid clip for other users. This change is seeing users moving away from traditional instrumentation in favour of modular devices. It is, in the opinion of some companies, the end of the so called 'rack and stack' era – and it's all being driven by the benefits derived from Moore's Law.

Kyle Voosen, National Instruments' director of marketing for the UK and Ireland, said: "Processor performance doubles every 18months, size decreases and so does cost." Larry Desjardin, general manager of Agilent's modular product operation, is another proponent of modular instrumentation. He said: "Modular instrument adoption is on the upswing. It's a growing market, largely due to the desire to reduce the cost per test and rack footprint, while improving the speed of test systems." More processor power is allowing greater levels of functionality to be integrated, for example, into consumer devices. "One piece of hardware can now be many different things," Voosen continued. "But, along with new design challenges, there are new test challenges. While the device itself may be harder to test, the instrumentation being used is getting smarter." An example given by Voosen is the Android based smartphone. "This has 16 'must test' systems, including the various radios, functions and communications ports. Testing these phones is a big job; it's not just a matter of plugging in software and seeing if it works." And this level of complexity means that, unlike with older mobile phones, one box won't do the testing job. "Where companies used to use one box to test a particular phone," he continued, "that approach doesn't work with software defined electronics." So, if one box won't do the job, how do you test these devices? "The traditional approach was 'rack and stack' – and remember that these boxes aren't going away – but 'rack and stack' isn't appropriate for new test requirements." He gave multinational electronics company Thales as an example. "It wanted to develop communications devices for 'first responders', which required devices with wide frequency use and a range of modulation schemes. Thales recognised that is couldn't have products failing in the field, so it came to NI to design a better way to automate test while not slowing production by testing every different configuration." The solution was a PXI based system. According to Mohammad Admad, engineering test and operations manager with Thales Communications: "NI's PXI hardware and LabVIEW software are essential technologies for test automation productivity and reuse, helping us build a common test architecture and perform 'lights out' testing to meet our growing expectations for increasingly complex product families." BAE Systems is making much the same transistion, moving away from rack and stack test systems towards what it terms synthetic instrumentation. These software defined instruments allow the company to develop systems which are not only smaller, but which are faster, more accurate and less expensive. Many of the systems being developed are based on the PXI standard, where PXI stands for PCI Extensions for Instrumentation. PXI isn't new – the CompactPCI based standard was first announced in 1998 – but it is now gaining momentum, driven by the Moore's Law benefits outlined above by Voosen. He noted: "Around 100,000 systems have been deployed, featuring more than 600,000 PXI modules. This is expected to grow to more than 300,000 systems by 2014," he claimed. "People are moving quickly to the PXI platform; there has been a dramatic shift over the last couple of years." Today, there are more than 55 companies in the PXI Alliance, offering more than 1140 products between them. One of the latest to join is Agilent. In a corporate statement explaining the move, it said 'PXI is currently the dominant standard for modular instrumentation. It's a mature technology and is widely used'. Combining PCI-Express communications with leading edge processors, fgpas and data converters, PXI based devices are faster and smaller than their traditional 'box' equivalents. Voosen pointed to a recent NI announcement, the PXIe-5665. "It's less than 10% of the size of a competitive standalone instrument," he claimed. Analog Devices has developed a PXI based test system to replace a previous ATE approach. According to the company, the system cost less than 10% of the ATE and has a footprint just 7% that of the ATE system. Apart from weighing 30kg, compared to the 2tonnes of the ATE system, it consumes just 600W, rather than 10kW. Its opinion is 'the shipping container for the previous system would cost as much as our entire new PXI test system'. But PXI isn't the only approach being taken. Alongside is AXIe, a standard based on AdvancedTCA, with extensions for instrumentation and test. The AXIe Consortium is looking to provide an open standard that creates what it calls 'a robust ecosystem of components, products and systems for general purpose instrumentation and semiconductor test'. To do this, AXIe builds on existing standards – including PXI, LXI and IVI – to bring scalability and performance to bench top measurements, as well as rack and stack and ATE systems. AXIe is a much younger enterprise than PXI. The consortium was founded at the end of 2009 by Aeroflex, Agilent and Test Evolution. The approach is said by the consortium to have three key attributes: it has been designed for high performance instrumentation with a large board size that provides room for module shielding; it is well suited to high power applications, with single rail power of up to 200W per slot; and it offers greater scalability and rack space efficiency than PXI. An AXIe chassis can contain up to 14 node slots and two hub slots and these can be arranged vertically or horizontally, with multiple chassis connected to create systems with more capacity. ATCA supports a lan connection to each slot, as well as PCI-Express to each slot. As such, AXIe is described by consortium members as PXI's 'big brother'. Desjardin said: "It's 'big', because it is the high performance platform. It's a 'brother' because of its compatibility. And the communications protocols are so similar that an external controller talking to a system can't distinguish if it's PXI or AXIe." ATCA is seen as a suitable platform for a number of reasons. One is its support for large board sizes; useful when developing high performance instruments. It also offers what is described as 'robust' system management. Two versions of AXIe are being developed. Version 1.0 is the general purpose specification, while 3.1 is aimed at semiconductor test. Future developments are likely to include improving integration of ATCA, AXIe 1.0 and AXIe 3.1 based systems, development of the AXIe 2.0 software specification and development of what the consortium calls 'AXIe 3.n'; specifications for additional markets, including network test and custom applications. Desjardin believes traditional box instruments and modular products may well work together in the future, bringing more choice for the user. Concluding, Voosen said: "Things will just get better."