Broadband measuring station for the extended testing of the directional response pattern of sound level meters and the calibration of measuring microphones

It is common practice to use sinusoidal sound signals as an excitation signal to reach the low measurement uncertainties required for the testing of the directional response pattern of sound level meters and for the calibration of measuring microphones. By using broadband signals, it has now become possible to reduce the time needed for a measurement, to considerably increase the spectral resolution and to facilitate the identification of distorting reflections. Against all expectations, this has not caused the measurement uncertainty to increase, but rather to decrease.

The free-field secondary calibration of measuring microphones and type approvals of sound level meters (SLMs), among other things, belong to the legal tasks of Working Group 1.63 "Noise-Measuring Technology". One of the numerous tests performed for type approvals is the directional response pattern test in which the tested object has to show that sound coming from all possible directions of incidence is equally evaluated.
Both measuring stations – that for microphone calibration and that for SLM directional response pattern testing – are very similar, as far as the basic arrangement is concerned. In the case of both measuring stations, the prescribed test frequencies are individually excited using sinusoidal sound signals. In times of FFT and real-time analysers, this may sound obsolete, is, however, justified by the fact that for these precision measurements, low measurement uncertainties are required which can only be obtained by means of sinusoidal sound signal excitation. The disadvantage of this method is that it is time-consuming due to the fact that it requires numerous individual measurements. Hence, the corresponding measurement specifications are limited to the frequencies which are absolutely necessary. Thus, a measurement of the directional response pattern can take 12 to 24 hours – depending on the test object – the spectral resolution, however, is so low that distorting influences caused by reflections on the test object or on the stand's fixtures cannot be differentiated or even detected. Much experience is required and it may even be necessary to carry out additional measurements in order to ensure PTB-worthy quality.
Within the scope of a degree dissertation [1], both measuring stations were completed in such a way that in addition to the sinusoidal sound signals required in the standards, also measurements can be carried out on the basis of broadband excitation with sweep signals and a two-channel FFT analyser (Fig. 1).

Figure 1: Comparison of the measured sensitivity level of a typical 1/2" measuring microphone (WS2P), once with the standard method using sinusoidal sound signals, and once using the broadband method.

As a direct consequence, the measurement time is significantly reduced: instead of performing 30 to 40 individual measurements over several hours for just as many frequencies, the determination of a spectrum with 800 or more frequency lines only takes a few minutes. The gain of time brings about several advantages: further measurements can be carried out in order to obtain additional test frequencies, averagings over measurement values and, in the case of the directional response pattern, an improved angular resolution. It makes it possible to identify, e.g., reflections more easily (see Fig. 2). But also the reproducibility of the measurement is improved so that the corresponding contribution to the measurement uncertainty turned out to be much lower than when using the time-consuming sinusoidal sound signal method. Although various hypotheses are necessary in the broadband method which bring about additional contributions to the measurement uncertainty, the low total uncertainty can be maintained in the case of microphone calibration and even be reduced to some extent in the case of SLM directional response pattern testing.

Figure 2: Comparison of the frequency resolution of the deviation from the sphere's directional response pattern used as a basis for the conformity assessment of a sound level meter, once by means of the standard method using sinusoidal sound signal excitation, and once using the broadband method. The DMA curve prescribes a conformity limit. The comb filter effect, which suggests that there is a reflection point close to the microphone, can only be seen in the broadband curve.

The broadband methods are now used as a complement to the standardised sinusoidal sound signal methods at PTB's measuring stations for the free-field secondary calibration of measuring microphones and for the testing of the directional response pattern of sound level meters in order to, for the time being, collect measurement data for a potential modification proposal at the standardisation institutions in charge, but also in order to further increase the quality of the measurements which are already running.

[1] Schuster, Michaela: Entwicklung eines zeitoptimierten Messplatzes zur präzisen Bestimmung von Kalibrierdaten und Richtcharakteristiken von Messmikrofonen mit breitbandigen Prüfsignalen. Diplomarbeit, Technische Universität Ilmenau, 2008.

Contact person:

Christoph Kling, Dept. 1.6, WG 1.63, E-Mail: christoph.kling@ptb.de