Agilent responds to NI VSA comparison
Rick Nelson- August 10, 2011
Taking note of competitive comparisons presented by National Instruments at NIWeek with regard to signal-analyzer performance, Agilent Technologies is presenting a forceful response. The controversy stems in part from claims made at a Tuesday August 2 NIWeek keynote presentation, in which Jin Baines, director of RF R&D at NI, and Raajit Lall, NI product marketing manager for RF, presented a head-to-head comparison of the two companies’ vector signal analyzers: the NI PXIe-5665 and the Agilent PXA.
Baines and Lall made admittedly bold statements regarding the RF performance of the 5665 with respect to TOI (third-order intercept), noise floor, and phase noise. In the Tuesday keynote presentation and in a related video, Lall presented demonstrations using the 5665 and PXA to make ACPR (adjacent-channel power ratio), TOI, and harmonic measurements. The live demonstration I saw on Tuesday dealt with adjacent-channel power measurements on a test WCDMA signal using both instruments. Lall concluded that both instruments achieved about -80-dBc ACPR dynamic range, with the 5665 operating nearly 15 times faster, completing a measurement in 30 ms, vs. 430 ms for the PXA.
In a conference call today, Jim Curran, marketing manager of Agilent’s microwave and communications division, took issue with the ACPR and other demonstrations. With respect to ACPR dynamic range demonstration, he said that the NI presenters, without including a good notch filter, were measuring the adjacent-channel noise of the RF test signal—not the instruments’ performance. He added that with noise correction, the PXA achieves a nominal ACPR dynamic range of -88 dBc. With respect to speed, Curran said that without noise correction, the PXA can acquire a measurement having 0.2-sigma variance in about 14 ms while operating at a dynamic range better than -81 dBc.
Curran had similar issues with the TOI and harmonic-measurement comparisons, noting, for example, that the PXA has a built-in one-button routine that automates and optimizes TOI measurements; that routine, he said, would have yielded a correct TOI of +36 dBm for the PXA used under the conditions of NI’s demo—not the +19 or 20 dBm that NI reported.
One takeaway here is that when choosing something as complex as RF test equipment, seek out demos from multiple vendors, each of whom knows how to optimize measurements using its products. And try to ensure that the demonstrations match as closely as possible the real-world situation in which the instrument you choose will ultimately operate.