'Real' History of Real-Time Spectrum Analyzers: A 50-Year Trip Down Memory Lane, The

Sound and Vibration, Jan 2007 by Deery, Joe

Real-time spectrum analysis is used to predict and analyze mechanical faults and failures in rotating machinery to analyze vibratory motions of components, systems and structures; to analyze the noise constituents in these systems; and for many other purposes. Real-time spectrum analyzer (RTA) development has a 50-year history that involves overcoming numerous technical barriers, challenging measurement techniques and physical obstacles. This article travels down memory lane from the earliest days in the development of these instruments at the Applied Research Laboratories of Columbia University to some of the latest product offerings. The many individuals that have contributed to this history and their accomplishments are chronicled.

In 1957 the US Air Force accepted delivery of a real-time spectrum analyzer with a Coherent Memory Filtur system from resoarchors at Columbia University. This instrument simulated a bank of parallel filter/detectors, providing the constant-bandwidth spectrum of an applied analog input signal. This first RTA was implemented using glass acoustic delay lines and provided real-time processing from less than 1 Hz to 40 kHz. A commercial enterprise. Federal Scientific Corporation, was spun off to develop this exciting new technology. It was founded by Henry Bickel, Reinhold Vogel. Joe Flink, Art Citrin, Mike Laviola and Bill Brookner and located at 615 West 131st Street in New York City. They obtained a patent on the "Time-Compression" methodology that allowed a single filter and heterodyning electronics to produce a spectrum in real time.

This name Coherent Memory Filter gave way to "Simoramic" and eventually to the familiar trademark, Ubiquitous (thank you, Dick Rothschild). Early Federal Scientific employees developed and refined the time-compression or 'deltic' RTA.

1960s Developments

The first commercial RTA from Federal Scientific was the UA-7 Ubiquitous® Spectrum Analyzer. The engineering team had blinders on when it came to mechanical applications and they only projected the sale of about 100 sets to Electronic Warfare and military customers.

Spectral Dynamics, under the direction of Laurie Burrows and Hugh Ness, realized the breakthrough that Federal Scientific had made and entered into a license agreement to produce a product equivalent to the UA-7, which ultimately became the SD 301. These early RTAs processed data up to 20 kHz through a single filter that acted as 500 filters in 50 msec. These systems had street sale prices in the $35,000 range.

While it licensed the UA-7 to Spectral, Federal had an improved model under development. This was introduced (interestingly) as the UA-6 and featured a greater frequency range to 40 kHz. Spectral was furious about this new product offering from Federal, and a legal battle ensued. Spectral placed the highest level of engineering technology on developing RTAs for the environmental testing, rotating machinery, automotive and related markets. Federal/Spectral sales arguments included debating the benefit of 'dithering' the ADC with random noise and a 2 dB difference in dynamic range (48 to 50 dB).

Other scientists who performed work with the group at Columbia University started Saicor, which was ultimately acquired by Honeywell. Another company, EMR Hatboro, introduced an analyzer that held a spot on the stage into the '70s for about four to five years. Multiple window functions were its claim to glory. Princeton Applied Research, which was known for its "boxcar integrators," branched out and developed an RTA as well.

At this time, 85% of the RTA commercial market was divided between Spectral Dynamics and Nicolet Scientific, which purchased Federal Scientific. The lion's share of sales was registered by Spectral. Spectral made a major breakthrough for the U.S. Navy. They integrated an RTA and special waterfall display hardware with passive sonar to produce the AN/BQR-20 system. This sonar analysis system (and its successors) became standard tactical equipment on all U.S. submarines. Spectral sold many hundreds over a ten to fifteen year period. The SD13151 waterfall display was designed by Carl Dubois for sonar applications.

Federal made a half-hearted effort to get into the submarine sonar business with little success. Their 'fish-bowl' advertisement raised the Navy's ire and had to be quickly withdrawn to preserve good working relationships with the many Navy laboratories Federal serviced. However, they did contribute directly to the submarine sonar effort by innovating a new instrument called a Frequency Translator. The Frequency Translator was an accessory for an RTA. It used analog up/down heterodyning to accomplish what is now called Zoom spectral analysis. The translator allowed the RTA to analyze a narrow frequency band with very high resolution.

In 1965, Cooley and Tukey at Princeton University published their historic paper on the computation of the FFT. This development allowed a 1024 point FFT transform (512 lines) to be processed in one one-hundredth of the time previously required. It required 50% less memory to complete the computation.