1 Semi-Bipolar Piezo Actuator Driven by a Single Op Amp

A piezo transducer or actuator can make use of the inverse piezoelectric effect to take an electrical signal and turn it into a mechanical signal, or actuation. The voltage across the piezo actuator determines how much it expands or contracts, often on the order of micrometers (µm). This facilitates extremely precise displacement of the actuator via a drive or control apparatus. These actuators are commonly used in test and measurement applications, for precise control of probes or for sample positioning; in aerospace and defense, for active vibration dampening; and even in optical applications, for minute tuning and adjustments to lasers. While commercial off-the-shelf piezo drive units are available, a piezo actuator may also be controlled by a precision operational amplifier (op amp) drive circuit. The specific circuit design is highly application-dependent, with the capacitance of the actuator, its resonant frequency, and its supply voltage range all playing a role; however, this document explores a general case of a low-to-medium supply semi-bipolar actuator that is operating below its resonant frequency, allowing it to be modeled as a capacitive load.

Consider first the case of a semi-bipolar piezo actuator driven by a single op amp, with the non-driven terminal of the actuator tied to the system ground. The input of the op amp can be controlled via a digital-to-analog converter (DAC); the amplifier will then gain up this control signal and drive the piezo, as shown in Figure 1-1A.

Figure 1-1 Driving a Piezoelectric Load With a Single Amplifier (A) and Bridge Arrangement (B)

Another way to drive a piezo actuator is to treat it as a floating load, “bridging” the outputs of two different op amps as shown in Figure 1-1B. The difference in the outputs of the amplifier is the effective voltage across the actuator. This means that by carefully selecting the supply voltages of the two amplifiers, a much greater effective voltage can be achieved. The “high” side of the load, driven by U1, can swing approximately as high as VCC1 and as low as VEE1, depending on the output swing limitations of U1 and the required current. In a similar fashion, the “low” side of the load can swing approximately as high as VCC2 and as low as VEE2, although some headroom is necessary due to the output swing limitations of U2. The greatest effective voltage across the load occurs when U1 swings high and U2 (which is 180° out of phase) swings low – the effective voltage across the load is thus:

Vpiezo = (VCC1 – U1 output swing to positive rail) – (VEE2 – U2 output swing to negative rail)

For example, if U1 and U2 can each swing within 5 V of either rail, and the supplies are VCC1 = VCC2 = 90 V and VEE1 = VEE2 = –90 V, then the greatest voltage across the load is (90 V – 5 V) – (–90 V – [–5 V]) = 170 V. Likewise, the lowest effective voltage across the load occurs when U1 swings low and U2 swings high. The effective voltage across the load is thus:

Vpiezo = (VEE1 – U1 output swing to negative rail) – (VCC2 – U2 output swing to positive rail)

For the supplies given in our prior example, this means a voltage of (–90 V – [–5 V]) – (90 V – 5 V) = –170 V is possible. Thus, the piezo load can swing from +170 V to –170 V, or 340 V pk-pk.

Most piezo actuators, however, are not intended to be driven in this fashion – rather, the “high” voltage across the load is intended to be higher than the “low” voltage, such as 150 V to 0 V, or 120 V to –30 V rather than 75 V to –75 V. We can achieve an offset, non-zero-centered output by shifting the supplies of the U1 and U2 op amps. For example, assume VCC1 = 175 V, VEE1 = –5 V, VCC2 = 5 V, and VEE2 = –175 V. The voltage across the load is now (175 V – 5 V) – (–175 V – [–5 V]) = 340 V when “high”, and (5 V – 5 V) – (–5 V – [–5 V]) = 0 V when “low”. The peak-to-peak voltage is unchanged, but the output is now centered on 170 V instead of 0 V. Thus, achieving a desired output swing range for a given piezoelectric load is simply a matter of shifting the supplies accordingly.

When driving a large capacitive load such as a piezoelectric actuator, two major concerns for the designer are output current requirements and circuit stability. The piezo load is constrained by the same i = C dv/dt equation that covers typical capacitive loads (there is a small intrinsic resistance due to electromechanical losses, but the actual value will depend on the actuator used). Thus, an actuator with a high capacitance or one that is excited at a high frequency will have correspondingly high current demands. In some cases, an external output stage after the op amp may be required to achieve high output current. This output stage, and any related compensation adjustments to meet stability requirements will be highly application specific. The stability of the circuit will depend on the load, the AC characteristics of the op amps employed, and when applicable the presence of an added output stage. Even if the U1 and U2 op amps are from the same fabrication lot, their gain-bandwidth product (GBW) may differ slightly, resulting in somewhat different gain and phase roll-offs. Therefore, it is suggested that a feedback capacitor (CF) be added across each op amp feedback resistor (RF) to reduce any excessive system bandwidth. Reducing the bandwidth reduces the circuit noise, which is especially important if the noise falls near the resonant frequency of the actuator. Additionally, the capacitors can help force the frequency roll off of the two amplifier circuits to be more closely matched.

Piezo actuators are comprised of specially formulated ceramics that exhibit piezoelectric properties similar to natural quartz. The actuators are designed to produce a particular mechanical response where they are able to extend or contract in response to the applied voltage. An electric field is developed across the internal dielectric of the actuator as charge collects on the actuator plates. The actuator reacts in the physical realm, in a mechanical manner as a response to the electric field developed. Conversely, a mechanical shock to the actuator can result in a momentary generation of a high-voltage output spike. This response is attributed to the forward piezoelectric effect that produces a very strong field manifested as a high voltage at the terminals of the actuator. This is commonly observed in the piezo strikers found on many gas stoves, which deliberately utilize this principle to generate an electrical arc that ignites the gas burners.

If a shock to an actuator that is driven by an amplifier circuit occurs, this high voltage spike will be applied directly to the output pin of the op amp, and to the internal output stage circuitry connected to the pin. The voltage levels of the spike can exceed the safe operating voltages of the output transistors and other components, possibly leading to voltage breakdown and circuit damage. In applications where a mechanical shock to the actuator might occur, the addition of electrical over-stress (EOS) protection is required to protect the op amp from being damaged. One common protective method to prevent such overvoltage damage is to include a transient voltage suppressor diode (TVS) directly across the piezo actuator. A TVS diode is similar to a Zener diode in that it turns on at a specified voltage, and then clamps the voltage to that level. They differ from a Zener in that they are designed specifically for clamping an EOS event, are very fast, and are made to handle repetitive surges without degradation. The TVS diode may be specified for either unidirectional single polarity clamping, or bidirectional dual polarity clamping. The simple TVS diode connected across the piezo actuator is sufficient for a single op amp driver actuator circuit. However, in a bridged output a different TVS clamp scheme is suggested, as will be shown.