The following two conditions should be simulated for the scale model experiment of room acoustics; (1) the sound absorption by air (medium condition), and (2) the sound absorption characteristics of the interior material (boundary condition). As for the condition (1), the sound absorption by air under normal room conditions can be approximately simulated by use of dry air on N_2 gas as the experimental medium gas in 1/10 scale model experiment. As for the condition (2), however, only a few data of the sound absorption coefficient of the absorbents for the model have been published hitherto, and in many cases, experiments have been carried out at the degree of simulation that only reflective and absorptive surfaces are distinguished. In this study, it has been examined to what an extent the simulation of the sound absorption characteristics of the interior material is possible in 1/10 scale model experiment. For that purpose, by reducing various kinds of sound absorption mechanisms (such as panel vibration, resonator, perforated panel, slit wall and porous material) to 1/10 scale, the reverberant sound absorption coefficient has been measured by use of the model reverberant room. As a result, it turned out that the simulation can be realised at fairly good accuracy over almost all kinds of absorbents. Fig. 1 shows the measurement system employed in this study. It is almost equal to the means for the full scale absorbents except that the frequency range is 10 times higher (800Hz〜40kHz). In these measurements, N_2 substitution method has been adopted to stabilize the sound absorption by experimental medium gas.
According to the works of Barlow et al. on amorphous polymers in molecular weights ranging from 448 to 7. 13×10^4, the shifted curves of shear data follow the Lamb spectrum in transition region from glassy to flow state independently of chemical or molecular structure. Whereas our data on polyvinyl-i-butyl ethers (number average molecular weight M_n=2. 66×10^3 and 1. 2×10^6) and poly-i-butyl methacrylate (M_n=4. 72×10^3) show rather contradictorily that the shifted curves are represented by the Davidson-Cole spectrum which indicates broader distribution of relaxation times than that of the Lamb spectrum. Hence it would be of interest to make a measurment on a series of polybutenes of low molecular weight and a comparison with Barlow's data. The modified rotating plate method was used to determine precisely shear wave velocity at frequencies of 1, 3 and 5 MHz. A reflection method was used to measure the real part of the acoustic impedance in a frequency range from 8. 6 to 102 MHz. Three Kinds of polybutene, LV-50, HV-100 and HV-1900, offered by Furukawa Chemical Co. were used, whose characteristics are listed in Table 1. Viscosity was measured by Cannon-Fenske Viscometer. The real parts of shear impedance in various frequencies are plotted against temperature in Fig. 4. A secondary dispersion on primary dispersion is found in the samples of HV-100 and HV-1900 around the glass transition temperature T_g as shown in Figs. 4 and 5 ; its magnitude in HV-1900 is the highest. It is found that molecular weight dependence of limiting shear modulus G_∞ is similar to those of Barlow et al. As shown in Fig. 7, the differences between the values of G_∞ derived from Litovitz's and Lamb's method reaches 15% at the most. The shifted curves of LV-50, HV-100 and HV-1900 are shown in Figs. 8, 9 and 10, using G_∞ derived from Litovitz's method. The shifted curve of LV-50 agrees well with the theoretical curve derived from Lamb spectrum shown by the solid line in the whole frequency region, while those of HV-100 and HV-1900 deviate from the theory in the region of log ωτ_H>0. 5. The shifted curves of HV-100 and HV-1900 in a higher frequecy fit the Davidson-Cole spectrum shown by dotted lines. It is clear from Fig. 11 that the width of distribution which is expressed by the parameter β become broader as the molecular weight increases. Hirai and Eyring proposed that G_∞=RT/(V_0)exp(E_h/(RT)), where R is the gas constant, V_0=M_0/ρ the volume of monomer unit and E_h the energy necessary to create hole per mole. Value of E_h/(RT_g) for HV-1900 is found to be 4. 25 which is grater than the universal constant (3. 69±0. 12)