Conception acoustique de systèmes microtechniques

Considered as one of the most difficult complications in horology, the mechanical chiming watch, known as the striking watch, has today become one of the symbols of high craftsmanship. When it chimes the hour the sound of a striking watch must not only be harmonious to the ear of the hearer but also be identical from one watch to another. Today only a few talented watchmakers produce these watches. The resulting empiricism sometimes translates into significant differences between the sounds of these watches. One particularly complex aspect in the application of sound to a striking watch is the appreciation of the sound quality as well as the adequacy to the sound reference. This thesis aims to identify the principal influencing factors at the origin of the variances in sound from one striking watch to another and to identify those that allow the sound of one watch to conform to a recognised ideal sound. The problem is approached from two different angles. The first angle of physical reality, attempts to describe and characterise the sound produced by a watch according to physical parameters. The second angle of perceptive reality models our perception of sound. A virtual spectrum can be established by taking into account psychoacoustic aspects. This is derived from the real spectrum and allows not only a quantified measure of the perceived pitch but also an appreciation of the tone quality. The experimental work relies on the implementation of designed experiments using a functional sample. This reproduces the striking system at the same time allowing the modification of parameters necessary for the planned experiments. The physical spectrum combined with the virtual spectrum allows the identification of the principal influencing factors. Amongst these are hammer-gong pre-penetration, the contact geometry as well as the position of the hitting point and the contact time. By means of the generalised mass, it is possible to modify the perceived pitch within a significant range of frequencies which makes this factor an important element in the adequacy to the recognised ideal of sound. Finally, this work illustrates the importance that should be accorded to the implementation of measuring tools and appropriate characterisation of sound if the processes in the mechanical striking watch industry are to be mastered. Also included is a description of the factors that the watchmaker must control to ensure minimal variation from one watch to another in a production series.