When two approaches to the same problem exist in a market, they are bound to compete at some level. For PXI and bench instruments, that competition began in the area of automated testing, where PXI’s PC-based control and modular nature gave it compelling advantages. But improvements in modular instrument technology along with new software interfaces are blurring distinctions between PXI and desktop instruments, making the choice of which approach to use more complex.

The introduction by Aeroflex at Autotestcon 2007 of the 3000A PXI chassis for RF test highlighted this blurring of lines. The chassis includes a built-in touch-sensitive display panel that works with the system controller to turn a populated chassis into a self-contained instrument. Instead of having the separate keyboard and monitor that make most PXI systems seem like computers with an instrumentation peripheral, the Aeroflex chassis seems like a traditional bench instrument.

This type of operation addresses the need for programming that many engineers see as a barrier to using PXI on the R&D engineering bench. Paul Knight, a test development manager for Radio Frame Networks in Redmond, WA, expressed this concern in an e-mail response to T&MW senior technical editor Martin Rowe’s blog posting “Do you use PXI on the bench?”

“We use PXI in our production test equipment, but have not been able to replace traditional instruments on the engineer’s desk,” wrote Knight. “I believe the lack of front panel controls and/or software panel control limits the effectiveness of the [PXI] instruments. Engineers do not want to have to write code to make measurements.”

But that barrier is collapsing. Software packages from companies like ZTEC and National Instruments are now offering graphical user interfaces that mimic the look and operation of bench instruments.

Fig. 1  User interface software is allowing PXI instruments to mimic the look and feel of bench instruments, blurring the distinctions between them. Courtesy of ZTEC.

By allowing users to operate a PXI system like a bench instrument, such interfaces bring the two test methodologies into more direct competition. This might lead a potential instrument purchaser into looking at the relative performance of each approach, but such an evaluation may well prove inconclusive. PXI and bench instruments are evenly matched in many applications and performance specifications (Figure 2).

Fig. 2  In key performance parameters such as frequency and resolution, differences between PXI and bench instruments have virtually vanished.

This does not mean that both PXI and bench methodologies have achieved identical performance. “Benchtop instruments typically lead in high-end (28-bit) resolution or ultra-high frequency ranges,” said McDonell, “but by and large, they are very comparable.”

Further, resolution is only one aspect of performance in test systems. Chris van Woerkom, senior marketing engineer at Agilent Technologies, pointed out that metrics such as measurement throughput can also be important. He noted that many PXI instruments depend on the system’s CPU to turn raw data into meaningful measurements. In a well-populated system, then, the need to share bus bandwidth may limit an instrument’s achievable measurement throughput despite it having a high data rate. Bench instruments do not suffer from such limitations because they have built-in processing.

That kind of built-in processing is making its way into PXI instruments, though. “We’re now seeing PXI modules with processing being performed in the module,” said ZTEC’s director of marketing and product strategy Boyd Shaw. “In the last five years, modular oscilloscopes have gone from being just digitizers to having all the performance in the module. We now have the same kinds of signal conditioning, waveform analysis, and parameter measurement algorithms as benchtop devices.”

The result is that—from a performance standpoint—neither bench nor PXI instruments have a compelling claim to superiority. Nor is that situation likely to change. Individual products of one type or another will temporarily win the top performance slot, but that honor typically is now trading back and forth.

Further, consolidation within the test instrument community is creating vendors that offer both PXI and bench products, ensuring that technical advances in one will inevitably make their way into the other.

Agilent, for example, has acquired PXI companies PXIT and Acqiris and is now offering some PXI and bench instruments made with identical boards, parts, software, and specifications. The only differences are the interface and the footprint. “We are differentiating ourselves with superior metrology,” said van Woerkom, “and letting our customers decide which package to purchase.”

This kind of commonality is also eroding another traditional difference between PXI and bench instruments: the ease of porting R&D test development to the production floor. With modular instrumentation becoming the foundation of production test equipment, using PXI in the lab gave an advantage over bench instruments when it came time to create production test routines. When both types of instruments use the same core elements, either choice yields the benefit.

Instead of performance, then, developers selecting between bench and PXI test instruments might consider how they are going to use the instrument. Each approach has advantages under different scenarios. PXI instruments, for instance, can typically fit more functionality into a smaller package than bench instruments. A PXI system can provide dozens of channels in the same size package as a typical two-channel bench oscilloscope, making PXI better suited for space-constrained installations. On the other hand, bench instruments do not have the bus, size, or power constraints of PXI systems, so they can offer more diverse functionality and built-in processing than individual PXI modules.

Further, most or all of the functionality in bench instruments is hard-wired and does not need to wait for an operating system and software to configure the instrument before making a measurement. This makes bench instruments better suited for taking quick-look measurements. LeCroy points out that its WaveJet portable oscilloscope, for instance, is on and ready to use in less than 3 s. The operating systems used in PC-like PXI controllers take far longer just to boot.

In the end, a user’s choice of platform may simply boil down to preference. Older engineers whose careers pre-date PXI and modular instrumentation may feel more comfortable using buttons and knobs than mouse clicks and keystrokes. Younger engineers, having spent nearly their entire lives using personal computers, may feel precisely the opposite.

None of these factors is compelling enough to make one approach win out over the other across the board. Improvements in interfaces, ease of use, packaging, and function density may eventually allow PXI instruments to overtake bench devices in all uses, but not any time soon. Said ZTEC’s Shaw, “People will need benchtop instruments for years to come.”