A recent installment of Found in the Attic looked at the Allison Laboratories 650 Random Noise Generator. Tuning up a multiband audio processor with pink noise was discussed as an application of noise generators. The classic Texar Audio Prizm was used as an example. We mentioned that you could measure the noise output of the device under test with an AC VTVM at the individual band outputs, oscilloscope, low frequency spectrum analyzer, or Real Time Analyzer (RTA). To complete the 1960s vintage noise measurement series, this installment of Found in the Attic looks at the General Radio 1550-A Octave Band Noise Analyzer.
RTAs, or octave-band analyzers, are special sound level meters that divide noise into separate frequency components. Electronic filter circuits are used to divide the sound or noise into individual frequency bands. The 1550-A has filters for eight octaves: 20-75 Hz, 75-150 Hz, 150-300 Hz, 300-600 Hz, 600-1200 Hz, 1200-2400 Hz, 2400-4800 Hz, and 4800 Hz-10 Khz. There is an additional switch position for the full 20 Hz to 10 Khz spectrum.
When compared with the RTAs of today, the 1550-A is quite basic. Nevertheless, the principles of operation are the same. Pink noise is usually the signal source. The analyzer's input is fed to a switch which selects one in a bank of eight filters of equal Q. Each is tuned to a progressively higher octave. The output of the 1550's filter is fed to a detector and dB meter.
More modern RTAs have a detector and indicator for each filter. For better resolution, 1/3- or 1/2-octave bandwidth filters are often used in newer analyzers, giving around 30 or 20 filters respectively. By using pink noise and a constant Q analyzer, the same response graphs will be generated as with swept sinewaves, except for the resolution.
If the device under test has a fairly flat frequency response, then the measurement results will be identical with RTA or sine wave testing. If the device has sharp peaks or notches, then the response will seem much more shallow and rounded when measured with the RTA.
RTAs have several applications. They may be used to help determine the adequacy of various types of frequency-dependent noise controls. Sometimes, soundproof studios aren't. The problem is usually in one part of the spectrum. By using pink noise with calibrated speakers and a calibrated microphone connected to the RTA, the offending band can be easily identified. Knowing that, the source of the leak can be identified, or the right kind of baffles can be designed.
RTAs may be used for developing hearing protection, because they can measure the amount of attenuation offered by the protectors in select octave bands. In addition to documenting the integrity of soundproof studios, architects use them to benchmark the sonic characteristics of auditoriums.
They are also valuable for the calibration of audiometers and to determine the adequacy of various types of noise control. The special signature of any given noise can be obtained by taking sound level meter readings at each of the center frequency bands. The results may indicate octave-bands that contain the majority of the total sound power being radiated.
The GR 1550-A was designed in 1951 by a small team led by Arnold Peterson, a sound and vibration engineer at General Radio. Also on the team was Leo L. Beranek of MIT, founder of the acoustic consulting firm of Bolt, Beranek and Newman.
As competition with rival Hewlett Packard heated up after the war, GenRad set out to develop a family of tools for the acoustic measurements market. Around the same time that the 1550-A was released, so were the 1390-A random noise generator, the 1551-A sound level meter, and an economy version, the 1555-A sound survey meter.
The 1550-A was released after the GR 760-B Sound Analyzer, which was continuously adjustable in frequency, effectively dividing the audio spectrum between 25 and 7500 Hz into 180 separate bands. This was a bit of overkill for many applications, and there was actually customer demand for a device with fewer bands.
In a shrewd bit of marketing, Peterson also wrote A Handbook of Noise Measurement, which was, and still is, a classic text on acoustic noise measurements. Not surprisingly, the book featured illustrations and applications of the GR measurement equipment in action.
As with much of the GR gear from this period, the 1550-A could be operated from either line voltage or batteries. For lowest noise operation, battery mode is recommended. The 1550 runs off a Burgess 6TA60 1.5v A/90v B battery, which mounts inside the analyzer, giving 200-250 hours of use when operated 8 hours per day. Alternately, the amplifier can be powered by a model 1261-A power supply, which is the same size as the battery.
This 1550 came with the Burgess battery inside. Although long defunct, it makes a nice display piece. These large A-B batteries became increasingly hard to purchase as the 1960s progressed. Many were used in suitcase and portable radios of the 1940s and ’50s, as well as test equipment. As everything transitioned to solid state, the equipment itself became obsolete, and there was no reason to manufacture the batteries.
Inside, the 1550 displays the hallmark quality construction that is characteristic of all GR equipment. Precision capacitors and inductors are used for the octave band filters. The amplifier is conventional, with 1-volt filament tubes to facilitate battery operation. The tube lineup consists of a 1T4 pentode and three 1U4 sharp cutoff pentodes.
The 1550-A has been in the attic for so long that the details concerning its origin and acquisition are murky at best. Like the GR 1231 Amplifier and Null Detector described earlier, its stately design and robust construction make it an excellent conversation piece. It is one of several items on rotating display.