Sound insulation of doors - Part 2: Comparison between measurement resultsand predictions

In the first part of this paper the prediction models for both structural a nd leak transmission of doors were presented [1]. In this second part, resu lts are presented for tested models comprising nine steel passage doors and nine timber passage doors. The results are presented in a form of two prac tical case studies. All measurements were made by the two-microphone sound intensity method. The structural SRI of a door was determined when the door was properly tape-sealed. The predicted structural R-w was on an average 1 .0 +/- 1.5 dB higher than measured R-w of tape-sealed doors, the range of v ariation being -1 ... + 3 dB (N = 13). The average difference between the p redicted and the measured SRI increased gradually with frequency from -3 up to +12 dB. The best structural solutions were those where two rigid panels formed a double panel without interpanel connections. The interpanel cavit y was filled with sound-absorbing material which does not form rigid interp anel connections. Structures with previous descriptions were found to give 8-10 dB better values of R-w than structures comprising strong interpanel c onnections with the same mass. Gomperts' model for slit-shaped apertures pr edicted reasonably well the frequency behaviour of slit transmission when n o seals were present (open apertures). When the door seams were sealed with rubber seals, the slits behaved in a more complex way probably because of the irregular shape of the slit. The total SRI of the door was calculated b y the area-weighted sum of the predicted structural transmission and the pr edicted slit transmission. Approximating the slit transmission coefficient by Gomperts' model and predicting the structural transmission by Sharp's mo del produced a good overall prediction accuracy for the total SRI of the do ors throughout the frequency range of interest. (C) 2000 Academic Press.