It has been shown that local compressive stress (or anisotropic stress effect, called ASE) gives large effect upon p-n junction characteristics. Especially when local compressive stress is applied to the emitter surface of a Si planar transistor, large change of current gain is observed, as shown in Fig. 2. This phenomenon is so sensitive to pressure change that some investigators have tried to apply it to mechano-electrical transducers such as acoustic devices, pressure gauges and other transducers. However there has been very little discussion from the practical point of view. The purpose of this article is to describe the sufficiently stable mechano-electrical transducer unit. As a transducer element, a new type of piezo-transistor is developed. It is gold-doped Si transistor which originaly has high signal-to-noise ratio, and has stressing area shown in Fig. 1. This electrode construction enables us to make transducer unit easily. However there remain some problems concerning stability. First of all, the fatigue of the stressed area was examined. Because of extremely small area of contact point formed by stressing needle and transistor, stress over the contact point is estimated to be extremely large. According to the experimental result, however, which was obtained from the cyclic load test(Fig, 3), it is proved to be tolerable for practical use. It is desirable to use a stressing needle having a large radius of curvature. On the other hand, the larger the radius of curvature is, the less becomes the sensitivity (Fig. 4). Therefore the selection of proper radius of curvature is needed to satisfy better signal-to-noise ratio and other practical demands. The circuit of piezo-transistor can be thought of as just the same as conventional transistor circuit. In order to obtain high sensitivity, it is desirable to have large base source impedance. For this reason it would be desirable to adopt fixed bias circuit. However, this circuit, as well known, is unstable. More stable operation against bias load change can be obtained by using current feedback bias circuit with a small sensitivity loss, shown in Fig. 6. In order to realize a mechanically stable transducer unit, stiffness of supporting structure of stressing needle must be small to axial direction and large to lateral direction. The reason why these conditions should be satisfied are explained as follows. Because of extremely large contact stiffness between the needle and the transistor surface, a slight axial displacement of stressing needle causes large stress change. This axial displacement mainly occurs from the difference of thermal expansion between materials constructed. Then the materials having the same thermal expansion coefficient should be chosen. These are Kovar for transistor header and frame construction, sapphire for stressing needle, and special elastic material developed in our laboratory(Fe-Ni-Ti alloy)for elastic plates. From the above consideration for realizing mechanical stability and for easy construction of units, the structure shown in Fig. 8 is adopted. With this structure, the following performances are obtained. 1) Sensitivity and noise level : shown in Table 2. 2) Humidity test : It was found that bias load changed during first 100 hr, at the condition of 45℃;, 95% humidity. After that, almost no changes were observed up to 1000 hr (Fig. 9). This means that 100 hr aging and readjustment of bias load are needed. 3) Mechanical strength : This test was carried by applying static force or shock from three directions shown in Fig. 10. To the direction (1) and (3), static forces were applied. It endured up to 5g to direction (1) and 40g to direction (3). To the direction (2) repetitive impulse forces were applied, and it endured 2. 5g peak amplitude of force at more than 500 repetitions. 4) Temperature dependence : Units were tested with regard to temperature dependence of bias load from 0 to 40℃;. The result is shown in Fig. 11. From this curve
Funneling mechanism proposed by Bekesy is thought to be a fundamental process in sensation. Single unit analyses in the mammalian auditory nerve and the cochlear nucleus have shown that the spike discharges responding to one tone can be reduced ot abolished characteristically by the addition of a second tone if appropriate stimulus parameters are chosen. Although this phenomenon is widely recognized, the mechanism involved is still obscure. In cochlear nucleus units the response to one tone is almost completely abolished by the presence of a second tone and this effect is more easily observable here than in the auditory nerve as shown in Fig. 7. This abolition of the discharge is phenomenologically called two-tone inhibition. Two-tone inhibition in the cochlear nucleus was studied by means of sweep-frequency method and simultaneous two-tone stimulation. Fig. 2B, C, and D represent two-tone inhibition just reflecting the pattern of Fig. 1. This inhibition was released by intermittent microelectrophoretic administration of picrotoxin, but not by strychnine as seen in Figs. 4, 5, 6. Two-tone inhibition in the cochlear nucleus is intensified further under the influence of some synaptic action for inhibiting incoming two-tone suppression in the auditory nerve. And, this inhibitory action may therefore play an important role for the neural sharpening of the response area. GABA may be a principal candidate among the possible inhibitory transmitters which might produce two-tone inhibition in the cochlear nucleus.
In previous works, Y. Ando, Electron. Commun. (Japan)51, 10-18(1968) and Y. Ando, J. Acoust. Soc. Amer. , 45, 1563(L) (1969), the complex reflection coefficients from even surfaces were accurately measured by this method, considering the directivity and acoustic center of the transducers and excluding sphericity errors and the effect of diffraction by finite dimensions of the test materials. The reflection of sound from both even-and uneven walls has been a problem of interest in architectural acoustics. For example, to simulate the acoustic field of a rooms by computer before construction. It is evident that the method is also useful in the scatterd field near the boundary under certain conditions (see Fig. 2, and Eqs. (4) and (5)). Measurements were taken on a semicylindrical boss on a rigid plane (Fig. 3 and Fig. 5), a periodically uneven surface of rectangular profile (Fig. 6 and Fig. 8), and a louver array obliquely pleced on an absorbing material (Fig. 9 and Fig. 10). Tests were made considering a function of distance from reflecting point and angle of incidence (Fig. 1). The experimental values of the first two cases were compared with the calculated values. They agreed satisfactory in the lower frequency range. In the last case, it was found that experimental values were greatly dependent on the positive or negative sign of the incident angle, although no theoretical analysis has been attempted.