Guest commentary: The technology driving Instrumentation 2.0—FPGAs
Rick Kuhlman, National Instruments -- Test & Measurement World, 7/25/2007 1:09:00 PM
Editor’s note: This article is part 3 in a series on Instrumentation 2.0. Read part 1 (the introduction) and part 2 (on PCI Express).
Instrumentation 2.0 marks a dramatic change in the way engineers create test systems through software-based automation coupled with modular hardware platforms. However, it is important to note that despite best efforts it is sometimes extremely difficult, if not impossible, to get away from custom circuitry all together. In fact, many times test engineers face tradeoffs between custom hardware and commercial off-the-shelf (COTS) hardware on multiple levels of the test design, striving to perfect the balance of throughput and test coverage. Because implementing custom hardware is sometimes necessary, the Instrumentation 2.0 approach has an important enabling technology—based on the FPGA—for engineers to get benefits of custom hardware with the flexibility of software.
Field programmable gate array (FPGA) technology has its roots in simple programmable logic invented in the mid-80s. An FPGA is a reconfigurable hardware chip consisting of logic blocks and programmable interconnects. The interconnects, which essentially take the form of a massive switch-matrix on a micro scale, connect the customizable logic blocks to form custom digital circuits, communicating to the outside world, like any other chip, through I/O pins. The impressive distinction of this technology is that the hardware circuits are defined in software first, then”downloaded” to the FPGA to actually execute in silicon. Furthermore, this hardware circuit is completely reconfigurable by simply downloading a new personality to the FPGA chip (Figure 1). (Go to Figure 1.)
Although available in some form for over 20 years, the return on investment from FPGA technology has not always been evident for test applications because of two key technical difficulties:
• First, an FPGA is just a stand-alone programmable chip. In order to actually use the technology, users must design surrounding circuitry, which can include ADCs, DACs, clocks, power supplies, and components necessary for programming the FPGA as well as components that provide communication with external peripherals the DUT itself.
• Second, hardware description languages (HDLs) for programming FPGAs, like Verilog or VHDL, are not well-known in industry—competence in those languages is often the sole domain of digital design engineers.
Similar to the over-arching concept of Instrumentation 2.0, the first problem is solved by modular COTS hardware incorporating FPGA technology. A number of vendors offer FPGA-based hardware that already has much of the external circuitry built-in to one board or platform. National Instruments, for example, has a family of FPGA offerings, which range from intelligent data-acquisition boards that take advantage of computer bus technologies and on-board ADCs and DACs to modular platforms with many I/O options and built-in signal conditioning.
In addition, current trends show that hardware description languages are becoming more abstracted and available to all engineering disciplines in response to the second difficulty outlined above. To support the next level of FPGA programming abstraction, important work is being done in C-to-VHDL converters and graphical programming techniques. Graphical FPGA programming environments, like the NI LabView FPGA Module (Figure 2), are especially applicable to FPGA design because they can naturally represent parallelism, pipelining, and dataflow, which are important for FPGA considerations. Using these techniques to overcome the resource-intensive barriers of board-layout and FPGA programming affords test designers the luxury of incorporating FPGA technology into next-generation test systems. However, an important question remains. I know I can use FPGAs in my test system, but why would I want to? (Go to Figure 2.)
FPGA benefits for test
One key FPGA benefit for test applications includes off-loading the main processor of your system from data-intensive or inline streaming tasks. FPGAs are perfect for inline filtering, averaging, modulating/demodulating, encrypting/decrypting, serializing/deserializing, decimating, or otherwise processing data. Say your test requires the average of 10,000 data points to determine pass/fail status. Unless you genuinely require every data value, you should avoid clogging the communication busses with lots of raw data. Instead, you can program an FPGA to process and average the sampled values and send only the final single value to higher level software for archiving and alarming. There are many examples of data reduction and complex inline processing functionality well suited for an FPGA.
Another key benefit of FPGAs is their inherent internal parallelism. One chip allows for any number of parallel data-processing paths. The positive relationship between parallel processing and throughput is as true on the micro-level inside an FPGA as it is on the macro-level with multiple entire systems running in parallel. Interestingly, this same “parallelism” trend is evident in the computing industry as manufacturers turn to multi-core processing to increase computer speeds.
Finally, simply being reconfigurable custom hardware is perhaps the most important benefit of the FPGA. From one perspective, this concept is essential for future maintenance of the test system, allowing for DUT changes, software changes, tweaks, and additional tests. However, another perspective shows that reconfigurability does not always have to be for future changes. Imagine your next-generation intelligent test bed reconfiguring its hardware personality based on the type of DUT introduced the system.
Sometimes, custom hardware is necessary for test system design. It might be for reasons of strict hardware timing and synchronization. Perhaps the application requires inline stream processing. Maybe it needs the high speed that custom hardware solutions provide. Whatever the reasons, it is important to know the options before turning to a fixed custom solution that will not grow with your changing needs. FPGA-based software-defined hardware provides many of the important characteristics of custom hardware with the added benefits of complex processing, true parallelism, and reconfigurability. With COTS FPGA hardware and higher level FPGA design tools to mitigate technical barriers, test engineers can efficiently utilize FPGA technology in their next-generation test systems.
Rick Kuhlman is a product marketing engineer for National Instruments. He is responsible for National Instruments LabView FPGA business objectives and works closely with the NI R&D organization and National Instruments customers to determine key product features and develop product strategy. He earned bachelor’s and master’s degrees in electrical engineering as well as an MBA from the University of Tennessee.
More on Instrumentation 2.0:
• Read part 1 of this multipart series, "PCI Express, multicore processors, and FPGAs drive Instrumentation 2.0."
• Read part 2 of this series, "The technology driving Instrumentation 2.0—PCI Express."
• Read Chief Editor Rick Nelson's April Editor's Note on Instrumentation 2.0.
• Read related blog commentary.
Talkback
Related Content
Related Content
There are no other articles related to this article.
Sponsored Links
TMW Resource Center
Browse Resources by Type:
- TEGAM Model 4040A Differential Instrumentation Amplifier
Tegam, Inc. | Product Demo
VIEW NOW
Tegam, Inc. | Application Note
VIEW NOW
DispoDog | Product Demo
VIEW NOW
View More
