Sound insulation of a single-frame light steel framing wall: Rock wool versus glass wool and the limits of double-leaf prediction models

Light steel framing walls lined with gypsum plasterboard are usually predicted as double-leaf drywall partitions, although their profiles are structural cold-formed sections several times thicker, and both boards are often screwed to a single frame. One such wall, with 12.5 mm plasterboard on each side of 90 × 40 mm studs of 0.95 mm steel at 600 mm centers, was measured with the cavity alternately filled with rock wool at 32 kg/m3 and glass wool at 13 kg/m3. Both fillings gave the same rating, R w ( C ; C tr ) = 28 ( _ 1 ; _ 1 ) dB. Rock wool gave 1.2 dB more on average from 100 to 1000 Hz across all bands, and nothing above 1250 Hz. The results are compared with Davy’s double-leaf model, which treats the studs as parallel-line connections whose equivalent translational compliance is derived from the section geometry using Castigliano’s theorem and scales with the cube of the steel thickness. Published empirical compliances, calibrated on thin non-structural studs, overestimate R w by 9–10 dB. From 100 to 630 Hz, the measured curve follows the field-incidence mass law of the combined surface mass to within 1.3 dB root mean square, so the single frame suppresses the double-leaf mechanism. Above 1250 Hz, the insulation is flat at 25 dB, 13 dB below the lowest value the line-connection model can yield for any compliance, indicating a residual path outside the model. The choice of mineral wool is therefore immaterial, and design effort belongs in decoupling the leaves and in airtightness.