Unlocking the myth of low-frequency footstep thumping noise in lightweight wood floor construction Lin Hu 1 FPInnovations 319 rue Franquet, Québec (Québec)G1P 4R4, Canada Anes Omeranovic 2 FPInnovations 319 rue Franquet, Québec (Québec)G1P 4R4, Canada Fabrice Roussiere 3 FPInnovations 319 rue Franquet, Québec (Québec)G1P 4R4, Canada Christian Dagenais 4 Building systems group, FPInnovations 1055, rue du PEPS, Québec (Québec) G1V 4C7, Canada Sylvain Gagnon 5 Building systems group, FPInnovations 1055, rue du PEPS, Québec (Québec) G1V 4C7, Canada

ABSTRACT Major occupants’ complaints received in light-frame wood-joisted floors of Impact Insulation Class (IIC) above 55, are low-frequency footstep thumping noise. Lack of understanding of fundamentals and solutions motived this study. A series of experiments were conducted to answer the following questions: 1) what are the frequency range of the low-frequency footstep thumping noises heard by the occupants in light-frame wood-joisted floors when the floors are impacted by footsteps of a person

1 Lin.hu@fpinnovations.ca 2 Anes.omeranovic@fpinnovations.ca 3 Fabrice.roussiere@fpinnovation s.ca

4 Christian.dagenais@fpinnovatio n .ca 5 Sylvain.gagnon@fpinnovations.ca

walking with shoes, and bare foot? 2) what the impact sound spectrums look like below the low boundary of 50 Hz as required by ISO standard or 100 Hz as required by ASTM standard? 3) Can the tapping machine excite the low-frequency noise down to the range of wood floor natural frequency around 15 Hz? 3) Can the current measurement system measure the sound signals at such low frequency range reliably? 4) can the measured spectrums of impact sound signals reveal the low- frequency impact noise issues? 5) what are the special construction details of the light-frame wood- joisted floors contributing to the low-frequency impact noise problem? 6) what are the other solutions than just adding mass to solve the problem? Our experiments answered the questions. The findings and solutions are provided in the relative sections of the paper.

1. INTRODUCTION

Figure 1 illustrates a typical conventional North American light-frame wood-joisted floor system. It consists of repetitive wood joists, subfloor, optional lateral reinforcement and topping, and ceiling. The joists can be of traditional solid sawn lumber, engineered I-joists or parallel-chord wood trusses and spaced at 406 mm (16”), 488 mm (19.2”), or 610 mm (24”) on centre. The conventional subfloor is a structural wood panels such as plywood or OSB of 15 mm-18 mm thick depending on joist spacing. The subfloor is fastened to the joists with mechanical fasteners such as nails or screws. The lateral reinforcement can be wood cross-bridging, blocking, lumber strapping, strongback and others to be placed between the joists to improve the load distribution across the floor width. For sound insulation, topping can be used and may be either dry or wet. For fire resistance Type X or C gypsum board is used on the ceiling. Based on the Canadian field test database and consumer survey made by FPInnovations in 1990s [1], the majority of the light-frame wood-joisted floors has an area density ranging from 20 to 30 kg/m 2 , depending on the type of topping. The measured field floor fundamental natural frequencies of the acceptable floors were above 15 Hz for bare floors and 10 Hz for the floors with heavy topping such as 38 mm thick concrete topping. The damping ratio of the fundamental vibration mode was around 2%-3% for unoccupied floors. It has been proven that light-frame wood-joisted floor system is the most optimized wood floor system in terms of cost-effectiveness and satisfactory performance for single family houses. However, when promoting “Green” buildings, wood construction has been expended to commercial and multi-family buildings where sound insulation is required. Sound insulation performance of wood construction may then become an issue, especially the low-frequency footstep thumping noise. FPInnovations has received occupants’ complaints about the low-frequency footstep thumping noise in the light-frame wood-joisted floors. Investigation of the complaints revealed that the floors that the occupants complained about had Impact Insulation Class (IIC) above 55 - thus meeting the minimum building code requirement, but the low-frequency footstep thumping noise was annoying. The current building codes for noise control and the single number ratings account for the noise above 100 Hz [2] to [3], or above 50 Hz [4]. Using the current rating methods to calculate the single number impact insulation performance, and using the calculated ratings to verify with the building code requirement, often it has been found those unsatisfied light-frame wood-joisted floors had much higher single number ratings of their impact sound insulation than the building code requirement. However, field inspection and subjective evaluation of those complaints have convinced that the low- frequency footstep thumping noise reported by the occupants were not tolerable, and annoying. This implies that the current building codes and rating methods do not properly address the low-frequency thumping noise issue.

Figure 1: A typical North American light-frame wood-joisted floor

Furthermore, it was found that currently there is a lack of understanding of the phenomena. The lack of understanding led to believe that such phenomena was due to the lightweight nature of wood, so the solution is just limited to add more mass to the wood floors such as concrete topping. But adding concrete topping has other performance problems such as reducing the vibration performance of the light-frame wood-joisted floors and accumulating the moisture in the wood buildings during and after construction, etc. It is important to unlock the myth of the low-frequency foot-step thumping noise in the light-frame wood-joisted floor systems and develop optimized solutions. Now questions arise such as: 1) what are the frequency range of the low-frequency footstep thumping noises heard by the occupants in light-frame wood-joisted floor when the floors are impacted by the ISO tapping machine, footsteps of a person walking with shoes, and with bare foot? 2) how does the impact sound spectrums below 50 Hz look like? 3) Can the tapping machine excite the low-frequency noise down to the range of wood floor natural frequency around 15 Hz? 4) Can the current measurement system measure the sound signals at such low frequency range reliably? 5) can the measured spectrums of impact sound signals reveal the low-frequency impact noise issues? 6) what are the special construction details of the light-frame wood-joisted floors contributing to the low-frequency impact noise problem? 7) what are the other solutions than just adding more mass to the floor such as a concrete topping? To answer these questions, a research program has been established. This paper provides some preliminary answers and some promising other solutions than the concrete topping.

2. OBJECTIVES

The main objective of the entire research program is to develop innovative solutions for controlling the low-frequency footstep thumping noise in wood floors. Specifically, this research aims:

1. to unlock the myth of its fundamentals, which will lead to the development of other solutions

than just adding more mass that has been used currently as the only solution to the problem due to the limited understanding of the fundamentals; 2. to develop subjective evaluation and measurement protocols including the devices of impact

source and measurement of the sound signals down to 6 Hz that has not been found in ASTM [2] and [3], and ISO [4], and in literatures; 3. to develop data analysis protocols and low-frequency impact sound rating method including

the sound from lowest and audible frequency around 15 Hz that has not been found in ASTM [2] and [3], and ISO [4] and literatures; and 4. to derive tentative minimum requirement for the low-frequency noise rating that has not been

found in the National Building Code of Canada (NBCC) [5] and other building codes.

This paper presents the work related to objectives 1 and 2. 3. APPROACHES

To fulfil the objectives nos. 1 and 2, the following approaches were used:

1. Conduct a series of experiments on a wood I-joist floor in FPInnovations’ mock-up of multi-

story five-unit wood construction, see Figure 2; 2. Subjectively evaluate the impact sound generated by a high precision impact system, i.e., the

electromagnetic tapping machine, Look Line model EM 500 “NEW”, a person walking with shoes, and bare foot on the bare wood I-joist floor with ceiling closed; 3. Use the conventional microphone Precision ½-inch microphone, PCB model 377B20, with

pre-amplifier to record the sound of the floor under the three types of impact sources, and setup the measurement frequency range extended from 0 Hz to 4000 Hz; 4. Transfer the time signals of the sound from time domain to frequency domain and to calculate

the impact sound pressure levels using the conventional analyzer, i.e., the Sound Level Meter of Larson Davis model 831; 5. Analyze the spectrums in the 1/3 octave band centre frequency band extended from 6.3 Hz to

4000 Hz; 6. Perform hammer-response modal tests on the floor to obtain the frequency response function

(FRF) from that the natural frequencies were extracted; and 7. Develop a 50 mm thick wood topping and repeat the steps 2-7 to verify whether the thick

wood topping solution eliminates the low-frequency footstep thumping noise in this experiment floor. 4. TEST FACILITY

The experiment was performed in FPInnovations’ mock-up of conventional North American multi- story 5-unit multi-family building made of wood as shown in Figure 2. The open room R1 was used as the source room of the impact sound and room R2 was the receiving room to measure the sound pressure level under the various impact excitations described in the approach section.

Figure 2: Mock-up of multi-story 5-unit wood multi-family building at FPInnovations

5. FLOOR SPECIMEN

The experimental floor was located in the open room R1. The floor was composed of 241 mm deep wood I-joists, spaced at 406 mm o.c. The subfloor was 15.0 mm thick tongue-and-groove OSB panels screwed to the joists. One layer of 15.9 mm thick Type X gypsum board was suspended from 12.5 mm thick resilient channels, which were spaced at 406 mm o.c. and attached to the bottom of the I- joists. 193 mm thick fiberglass insulation was placed in the ceiling cavity. The joints between the gypsum boards were filled with drywall joint compound. The installation followed the common

practice in terms of the size of screws and their spacing. The testing floor had 5.05 m clear span and 3.3 m in width. 6. METHODS

ISO 18432 method [6] was used for the classical impact-response modal testing to determine the frequency-response-function (FRF) of the test floor, and to obtain the floor vibration natural frequencies. To know what the low-frequency footstep thumping noise sounds like, subjective evaluations were conducted on the test floors. The tapping machine, a person’s walking with shoes, and bare foot were used to generate the impact forces. An evaluator was asked to stay quietly in the room below the floor, and carefully listen to the impact sounds generated in the floor under these three types of impact excitations. The evaluator was asked to compare the sound generated in the floor under the three types of impact excitations. Several evaluators were used so that their average opinion was used for the final judgement of the floor sound insulation performance. The ASTM E1007 method [3] for field floor impact sound testing was used for tapping machine impact sound tests. ASTM E989 [2] method was used to calculate the conventional Apparent Impact Insulation Class (AIIC) of the test floor under the three types of excitations. The AIIC then converted to ISO L n,w for the impact sound insulation using a soft-conversion, i.e., L n,w = 110-AIIC. The 1/3 octave central frequency band spectrums from 6.3 Hz to 4000 Hz of the test floor under the three types of impacts were plotted to identify the frequency range where the peak sound pressure levels occurred. Meanwhile the floor natural frequencies were extracted from the floor’s FRF functions. Then the floor vibration natural frequencies were compared with the frequency range of the peak sound pressure levels, and to unlock the myth of the low-frequency impact noise. 7. RESULTS AND DISCUSSIONS

7.1. Subjective Evaluation

The evaluators reported that they did not hear the low-frequency footstep thumping noise when the floor was excited by the tapping machine. But they clearly heard the low-frequency footstep thumping noise when a person was walking on the floor, especially the bare foot walking. The thumping noise excited by the walking forces sounds like a pure tone. Next section will visualize this phenomenon through the spectrums of the sound pressured levels measured under the floor excited with these three types of impact source at each frequency.

7.2. Spectrums of the Measured Sound Pressure Levels of the Bare Floor under the Impact

Figure 3 shows the comparison of the spectrums of the sound pressure levels of the bare test floor under the three types of excitations, e.g., tapping machine, walking with shoes and bare foot. Examining Figure 3 revealed the followings:

• at 10-25 Hz 1/3 octave band center frequency range, the three excitations excited the peak

values of the sound pressure levels of the test I-joist floor. It means that the tapping machine can excite the low frequency sound even if the evaluators did not identify it. • in comparisons of these three spectrums, the highest peak value of the sound pressure level

was found when the floor was excited by the bare footstep force that was the subjective evaluators reported as the most annoying case of the thumping noise when the people was walking on the floor with bare foot. It is noticed that the audible sound is above 15/16 Hz. • the low frequency signals down to 6.3 Hz measured using the conventional microphone

revealed the low-frequency footstep thumping noise occurred at the frequency range of 15- 25 Hz.

Figure 3: Spectrums of sound pressure levels of the test I-joist floor under excitations

7.3. Correlation of Low-Frequency Footstep Thumping Noise Frequencies and the Floor Natural Frequencies

Figure 4 illustrates the frequency-response-function (FRF) measured on the bare test I-joist floor. Observing the FRF, it was found that the floor lowest natural frequencies is in the range of 14-25 Hz that was in the same frequency ranges of the low-frequency footstep thumping noise heard by the evaluators. It implied that the low-frequency footstep thumping noise often occurred at the floor lowest natural frequency range.

Figure 4: FRF of the bare test I-joist floor Further examining the floor construction, it was found that the cluster frequencies in the range of 14- 25 Hz was due to the thin subfloor that had much lower stiffness in the minor strength axis of the

floor than the stiffness of the I-joist in the longitudinal direction, i.e., the major strength axis of the floor. This insight led to believe that an increase in thickness of the floor decking can increase the lateral stiffness, therefore increase the load-sharing capacity of the floor in the lateral direction and finally to better separate the low natural frequencies and to eliminate the low-frequency footstep thumping noise. To prove this assumption, an experimental 50 mm thick oriented strand board (OSB) was added to the bare I-joisted floor floating on a 11 mm wood fiberboard. Next section provides the results.

7.4. Innovative 50 mm Thick OSB Topping

The evaluators reported that the I-joist floor with the 50 mm thick OSB topping of density about 600 kg/m 3 floating on a 11 mm thick wood fiberboard was very quiet when the person was walking on the floor with shoes or with bare foot, and they did not hear the low-frequency footstep thumping noise. Figure 5 presents the comparison of the spectrums of the floor before and after adding the OSB topping when the people was walking on the floor with bare foot.

Figure 5: Comparison of the spectrums of the test I-joist floor under a person’s walking with bare

foot before and after the 50 mm thick OSB topping was added

As can be observed in Figure 5, the 50 mm thick OSB topping reduced the sound pressure level at the 1/3 octave band centre frequency range of around 14 Hz – 25 Hz by about 10 dB. This resulted in the satisfactory performance of the floor without hearing the low-frequency footstep thumping noise.

7.5. Measured AIIC

Using the ASTM E989 [2] method to calculate the AIIC of the bare test floor excited with the walking forces, the AIIC values were 71 and 66 (L n,w were 39 and 44, respectively) for the excitations the force generated by the person walking with shoes and bare foot, respectively. The AIIC values were much higher (L n,w was much lower) than any building codes requirement, but obviously the low- frequency footstep thumping noise was annoying. This because the classification methods only

account for the noise above 100 Hz, or 50 Hz, but the annoying thumping noise occurred at much lower than 50 Hz range. 8. FINDINGS

The experiments confirmed that:

• The low-frequency footstep thumping noise in the light frame wood joisted floor was

corresponding to the floor vibration natural frequencies in a range of 15 Hz to 25 Hz, and audible and perceivable as “Drum effect”. • The conventional measurement device such as the precision ½-inch microphone can measure

such low-frequency sound for the comparison study. • The ISO tapping machine can excite low-frequency noise down to 6.3 Hz. • Current rating in ASTM (IIC/AIIC) and ISO (L n,w ) does not address the noise of lower

frequency than 50 Hz, so the ratings do not catch the annoying noise that the occupants received. • For the light frame wood-joisted floors, the highest sound pressure level occurred at the range

of the floor natural frequency in the range of 15 Hz to 25 Hz or higher depending on the floor construction details. • The thin subfloor of 15-18 mm thick used in the current light-frame wood-joisted floor system

with a closed ceiling was thought responsible for the low-frequency footstep thumping noise because such hollow floor system forms a music-instrument box and amplifies the footstep sound. • Based on the above understanding, three solutions for the light frame wood joisted floor

system are proposed to solve the problem. They are: 1) adding thickness of the subfloor such as wood topping, which needs further verifications; 2) adding mass as the conventional method, but not concrete topping that creates other performance issues; and 3) using insulation materials that can absorb the low frequency sound down to 15 Hz in the floor cavity. • This study proved that the 50 mm thick OSB (30 kg/m 2 ) that is much lighter than the 38 mm

thick conventional concrete topping (78 kg/m 2 ) eliminated the low-frequency footstep thumping noise in the light frame wood I-joist floor system. 8. FURTHER REASEARCH NEEDS

• Develop a new rating method for IIC/AIIC to account for the sound down to 15 Hz. • Develop subjective evaluation protocol for floor impact sound insulation under walking

excitations. • Evaluate if the IIC/AIIC measured using the impact machine and calculated account for the

low-frequency sound down to 15 Hz is correlated to subjective evaluation. If not, • Then develop other impactors than the tapping machine. • Develop new building codes for the impact sound insulation requirement accounting the low-

frequency sound down to 15 Hz. • Develop other toppings, especially toppings other than concrete. • Develop new construction of the light-frame wood-joisted floor system using thicker subfloor

than the conventional subfloor of 15-18 mm thick. • Develop innovative sound insulation materials for absorbing the low frequency sound down

to 15 Hz. Such materials are not available currently on the market. 10. ACKNOWLEDGEMENTS

FPInnovations would like to thank its industry members, Natural Resources Canada (Canadian Forest Service) and Ministère des Forêts, de la Faune et des Parcs du Québec for their financial supports of this work.

12. REFERENCES

1. Hu, L. J. Serviceability design criteria for commercial and multi-family floors. Canadian Forest

Service Report No. for project No. 1092. 4, Forintek Canada Corp., Sainte-Foy, 42 p. (2000). 2. ASTM. Standard Classification for Determination of Impact Insulation Class (IIC). ASTM E989

(2012). 3. ASTM. Standard Test Method for Field Measurement of Tapping Machine Impact Dound

Transmission through Floor-ceiling assemblies and associated support structures. ASTM E1007 (2019). 4. ISO. Acoustics – Field Measurement of Sound Insulation in Buildings and Building Elements -

Part 2: Impact Sound Insulation. International Standard ISO 16283-2 (2015) 5. National Research Council. 2020. National Building code of Canada. Ottawa (2020) 6. ISO. Timber Structures – Test Method – Floor Vibration performance. International Standard

ISO 18324 (2016).