In assembling at least one metal frame in abutting supporting relation to a diaphragm to form a diaphragm assembly for an electro-acoustic transducer, for example, as in a microphone or headphone, the diaphragm, which may be an electret, is formed at least in part of a thermoplastic resin so as to provide the latter at each surface of the diaphragm which, in the abutting supporting relation, is engaged by a surface of a metal frame, and the frame or frames and the diaphragm are pressed together in the desired abutting supporting relation while each metal frame is heated so that heat is transmitted through the latter to melt the thermoplastic resin at the engaged surfaces of the diaphragm and frame or frames, whereupon the heating of the frame or frames is halted so that the thermoplastic resin hardens to bond together the diaphragm and frame or frames in the desired abutting supporting relation.