Investigation of vibration-based measurements of impact and airborne noise insulation Wayland Dong 1 John LoVerde 2 Veneklasen Associates, Inc. Santa Monica, CA USA Ben Shafer 3 PABCO Gypsum Rancho Cordova, CA USA Sunit Girdhar 4 Michigan Technological University Houghton, MI USA

ABSTRACT Standard impact and airborne noise insulation testing is based on measuring the average sound level in a reverberant field in order to estimate the incident and/or radiated sound power. While this method is straightforward, the variation under reproducibility conditions is significant and attempt to reduce the uncertainty has so far proven unsuccessful. It is possible that a significant portion of the uncertainty is due to unavoidable variation in reverberation rooms, and that measuring the vibration on the surfaces of an assembly provide an alternative and more reliable representation of the impact or airborne isolation of the assembly. Preliminary investigations have been performed by measuring the vibration levels on wall and floor assemblies in situ while excited by airborne and impact sources, and the results are compared to conventional test methods.

1. INTRODUCTION

Scientific inquiry and research over the last century resulted in developments and changes to the measurement of impact sound transmission through vertical assembly partitions. However, the fundamental measurement technique proposed and developed by the German scientist Gastell in the mid-1930s has largely remained common practice [1]. This measurement methodology was the basis for the development of what has become the International Organization for Standardization (ISO) standards for field measurements of impact noise insulation, ISO 16283-2 [2]. North American researchers Lindahl and Sabine, using Gastell’s work as a basis, further developed this measurement method, adapting it for measurements in the Americas [3]. Years of research and experimentation in

1 wdong@veneklasen.com 2 jloverde@veneklasen.com 3 ben.shafer@quietrock.com 4 sgirdhar@mtu.edu

building impact sound measurement resulted in the current ASTM International standard E1007-21, which defines impact insulation measurement in the field [4].

While these measurement methodologies are founded from robust analysis, the fundamental method of building structure excitation with a tapping machine and measuring the average sound pressure level in the presumed reverberant space is a nearly century-old method. The reproducibility of the method is generally considered to be poor, and there are indications that this may be due to inherent uncertainty in the measurement method rather than specific situations [5]. A vibro-acoustic characterization of the assembly can avoid the uncertainties associated with reverberation rooms and sampling the sound field. Some studies suggest that vibro-acoustic measurement and modeling techniques may allow for a more complete analysis of the transmission of impact sound through floors [6].

This research study is a preliminary exploration of the use of vibration measurement techniques in quantifying impact sound radiation of and through vertical (floor-ceiling) building partitions in- situ . This information may be used to develop an alternative analysis of impact sound transmission through vertical assemblies in addition to the standardized measurement methodologies currently in practice. While not exhaustive, this first study provides a proof of concept regarding the feasibility of these vibration measurement techniques for three different sources of impact radiation through floors: a modal impact hammer, a standard impact tapping machine, and an impact ball.

2. EXPERIMENTAL METHODOLOGY

Measurements were performed in a wood-framed multifamily building that was under construction. The floor-ceiling assembly consisted of 19 mm (3/4 inch) OSB subfloor, nominal 2 x 12 (38 x 286 mm) solid wood joists, fiberglass batt insulation in the joist cavity, 1/2-inch (12 mm) resilient furring channels, and one layer of 5/8-inch (16 mm) type X gypsum board. A gypsum concrete topping screed would later be poured, but the measurements were performed before the concrete pour. A pair of vertically adjacent bedrooms was selected for the measurements; the dimensions were 3.4 m x 4.6 m x 2.7 m and identical between floors. The walls and ceiling were gypsum board, the floor was exposed plywood subfloor, and there were no furnishings besides the measurement equipment.

Figure 1: Sketch of assembly under test. The gypsum concrete screed (arrow) had not been poured at the time of the measurements.

Vibration measurements were performed by mounting accelerometers to the floor and ceiling. Five measurement locations were selected on the floor, and three accelerometers were mounted to the ceiling with adhesive. The accelerometer signals were recorded and processed by multichannel analyzers, including Siemens Simcenter Testlab, MATLAB, and Bruel & Kjaer BKConnect.

The floor was impacted with three sources, a modal hammer equipped with a force transducer, the standard tapping machine, and the rubber ball. The same locations were used for all sources. Additional measurements were performed whose results are not described here, including airborne excitation of the floor using a loudspeaker and pink noise source, and sound intensity measurements

of the sound radiating from ceiling when excited by the tapper. Additionally, average sound pressure level measurements were performed in the receiving room per the current standards.

3. RESULTS

The analysis is in process. A sample of results is presented.

3.1. Floor impedance

The drive point impedance of the floor was measured with the impact hammer by placing the accelerometer as close as possible to the impact point. The “input” impedance was also measured as the averaged transfer impedance from the same impact point to three accelerometers at equal distances from the impact point arranged in an equilateral triangle. See Figure 2. This method of measuring floor input impedance was previously described [7].

Figure 2: Diagram and photo showing arrangement of accelerometers to measure input impedance

The impedance measurements are shown in Figure 3. Note that the high frequency data may be unreliable due to poor signal-to-noise ratio. At low frequencies, the two measurement methods vary by about 6 dB. However, at higher frequencies the drive point measurement recorded much lower values than the average transfer impedance. It appears that there were higher vibration levels at the drive point that did not transfer to the more distant accelerometers. We speculate that this may be due to damping in the plywood or non-linear deformation at the impact point. The average transfer impedance values are closer to the theoretical drive point impedance of an infinite plate based on published material values.

Figure 3: Measured floor impedance. Note that the high frequency data may be unreliable due to

poor signal-to-noise ratio.

3.2. Tapping Machine Impact Force

Measurement of the floor impedance allows calculation of the input force of the tapping machine based on vibration measurements with tapping machine excitation. The tapping machine was operated at four different locations on the floor (see Figure 4); the force was calculated for each of the five accelerometers locations based on the previously measured transfer impedance. This is the tapper input force based on the averaged floor response. The average is shown in Figure 5.

Frequency (Hz) e % & 10! 108 (say) aovepeduy

Figure 4: Measuring tapping machine input force

Figure 5: Calculated tapping machine input force

The overall shape of the force spectrum shows a number of expected features. There is a strong 10 Hz component from the impact frequency, with harmonics visible up to about 100 Hz. The spectrum is flat at low frequencies, and then rolls off at a frequency that presumably depends on the local compliance of the floor.

Force (N) rr wo! wo? 10" 10! rr Frequency (Hz) oe

For comparison, Figure 6 shows the same data (as a power spectral density) overlaid with the impact force for a single hit of the tapping machine hammer, measured using a sensor directly on the hammer [7]. The levels in this study are much higher, in part because it is a measurement of the tapping machine running and not a single hit.

Figure 6: Comparison of tapper force spectra

Frequency (Hz)

3.3. Rubber Ball

A similar method was used to calculate the force of the rubber ball drop. Following the testing procedure in [2], the ball was dropped from a height of 1 m five times on each source position. The same source positions were used as for the tapping machine. The calculated force spectrum is shown in Figure 7. Also shown is a direct measurement of the impact ball force using a force plate (labeled “previous work”); that research is described in another paper by the authors during this conference.

Similar to the tapping machine, the force measured using this method has a similar shape as a single impact measured under more controlled conditions, but somewhat higher level.

Figure 7: Rubber ball input force

4. FUTURE WORK

Additional analysis is being performed to examine the vibration transfer from the floor side to the ceiling side of the assembly. Sound intensity measurements were performed, which will be compared to the ceiling vibration levels. Airborne excitation will be examined. The results will be compared to traditional building acoustics measurement methods to evaluate if vibro-acoustic measurements can provide additional information on the assembly. Finally, similar measurements were also performed in a horizontal adjacency to evaluate airborne isolation of a wall.

5. ACKNOWLEDGEMENTS

The authors wish to thank Chris Kezon and the testing personnel at Western Electro-Acoustic Laboratory for their assistance in gathering data, and the management at Veneklasen Associates and PABCO Gypsum for their support.

Portions of this work were funded by the Paul S. Veneklasen Research Foundation.

6. REFERENCES

[1] A. Gastell, “Schalldämmmessungen in der Praxis und Vorschläge zur Normung des

Schallschutzes von Wohnungstrennwänden und Decken” (‘Sound insulation measurements in the practice and suggestions for standardization of sound insulation of walls and floors in housing’),” Akust Z , vol. 1, pp. 24–35, 1936.

fi H Li e a a ‘avg cade Frequency (Hz)

[2] “ISO 16283-2 (2014), Acoustics - Field measurement of sound insulation in buildings and of

building elements Part 2: Impact sound insulation,” International Standards Organization, 2014. [3] R. Lindahl and H. J. Sabine, “Measurement of Impact Sound Transmission Through Floors,” J.

Acoust. Soc. Am. , vol. 11, no. 4, pp. 401–405, Apr. 1940, doi: 10.1121/1.1916052. [4] “ASTM Standard E1007, ‘Standard Test Method for Field Measurement of Tapping Machine

Impact Sound Transmission Through Floor-Ceiling Assemblies and Associated Support Structures,’” ASTM International, West Conshohocken, PA, 2013. [5] E. Reynders, “Parametric uncertainty quantification of sound insulation values,” J. Acoust. Soc.

Am. , vol. 135, no. 4, pp. 1907–1918, Apr. 2014, doi: 10.1121/1.4868394. [6] T.-M. Kim, J.-T. Kim, and J.-S. Kim, “SEA-FEM hybrid analysis for predicting Inter-floor

impact noise,” Appl. Acoust. , vol. 129, pp. 397–407, Jan. 2018, doi: 10.1016/j.apacoust.2017.08.025. [7] S. Girdhar, A. Barnard, J. LoVerde, and W. Dong, “Measuring the force due to standard tapping

machine and floor impedance for ASTM standards,” Washington, D.C., 2021.