Figure 1 depicts a switching circuit that drives a piezoelectric pump at a high voltage with a polarity that alternates at the desired mechanical pump frequency. This circuit offers advantages of (1) high energy efficiency relative to conventional direct-drive circuits and (2) compactness relative to conventional resonant drive circuits.
Conventional direct drive circuits are inefficient because the piezoelectric actuators in the pump behave electrically as large capacitors and consume only a small part of the energy supplied to them during each pump cycle. Conventional resonant drive circuits are more efficient, but they must include inductive coils, which can be unacceptably large.
The present switching drive circuit includes inductive coils, but they are small and not used to resonate the actuator capacitance; instead, the charging of the actuator capacitances and the exchanges of energy among inductive and capacitive circuit elements are accomplished (by design) in times much shorter than the mechanical pump cycle. In other words, this drive circuit rapidly charges the actuator capacitance and turns itself off until the actuators reach their maximum mechanical expansion or contraction. Some time after the actuators reach their maximum mechanical expansion or contraction, the circuit turns itself back on to charge the actuator capacitance in the opposite polarity; as it does so, it replenishes the energy lost since the previous charge.
When power is first turned on, the capacitance of the piezoelectric actuators (C2 in Figure 1) is first charged to a potential of +450 V via transistor Q2. The charging current and, hence, charging time are determined by the characteristics of the dc-to-dc converter and the characteristics of the piezoelectric films. Inductor L1 is used to reduce the initial current spike during each recharge cycle, and C1 is used to stabilize the output voltage of the converter as well as reduce the peak output current demand on the dc-to-dc converter. Inductor L2 is the main energy storage inductor; it is used to exchange energy with C2 for switching polarity, as described in more detail below.
The timing signals that govern the operation of this circuit are transistor/ transistor-logic (TTL)-level pulses with a duration of 5 ms and a repetition frequency equal to the desired pump frequency (20 Hz in the original design). The short pulse duration is necessary in order to enable the triac (Q2) to turn itself off when the current in inductor L2 reaches zero. The timing signal is generated by timing/control oscillator U6, which supplies timing drive current to opto-isolator U4, which, in turn, provides the switch-on signal for the triac gate.