In the middle-frequency range many approximations usually can be made to simplify the analogous circuit of Fig. 9.3. Because the drive unit is very small, the mass of the diaphragm and the voice coil MMD is very small. This in turn usually means that the compliance of the suspension CMS is large to keep the resonance frequency low. Also, the conductance of the suspension GMS usually is large, and the reactance ωCM1 is small. Hence, in this frequency range, the circuit reduces essentially to that of Fig. 9.4a, where the conductance behind the diaphragm is
GMB≡1ρ0cSDm⋅N/s.
(9.2)
With the area-changing and electromechanical transformers removed, we get Fig. 9.4b, where the radiation conductance at the throat is
GMT≡STρ0cS2Dm⋅N/s.
(9.3)
As before, ST is the area of the throat and SD is the area of the diaphragm in m2. We have assumed here that the cavity behind the diaphragm in this frequency range is nearly perfectly absorbing, which may not always be true. Usually, however, this circuit is valid over a considerable frequency range because of the heavy damping provided by the conductance of the horn GMT. Also, GMT usually is smaller than GMB so that most of the power supplied by the diaphragm goes into the horn.
Assuming that the output resistance Rg of the generator is small compared with the coil resistance RE of the drive unit, the total electrical power supplied from the generator is
where we have used Eqs. (6.27) and (6.30) for the Bl factor. We note that the value of GMT, and hence the ratio ST/SD would seem to need to be large for high efficiency. However, if ST/SD becomes too large, reference to Fig. 9.4b shows that too much power will be dissipated in GMB and the efficiency will be low. To optimize the efficiency, let us
now differentiate the above with respect to (ST/SD) and equate the result to zero. Hence maximum efficiency occurs when
which can also be given in terms of Thiele–Small parameters:
Eff(max)=100SDc(SDc+ωSQESVAS√+ωSQESVAS√)2.
(9.11)
To increase the efficiency further, it is seen from Eq. (9.10) that the length l of the wire on the voice coil should be increased as much as possible without altering electrical resistance RE. Within given space limitations, this can be done by winding the voice coil from wire with a rectangular cross section rather than with a circular cross section. This means that the voice-coil mass will be increased. Increasing l further will demand a wire of larger cross section, which will require a larger air gap, with a corresponding reduction in B or increase in magnet size. Also, the voice coil must not become too large as its mass will limit the high-frequency response.