Piezoelectric sensors and signal-conditioning front-end solutions

Piezoelectric sensors and signal-conditioning front-end solutions

Part 1 of a multi-part series of sensor articles

BY STEVE TARANOVICH

Designing with piezoelectric sensors might seem, at first glance, a bit trivial but make no mistake, there are certain pitfalls of which to be aware as well as diverse configurations and internal structure differences that can make or break a design. In this article, I will show the designer how to properly interface these diverse “beasts” to an analog front-end design for optimum performance.

In my quest to help guide customers to the most optimum solutions for their products, many times during my career I have had to research multiple areas of expertise from many diverse authors and publications. These authors typically have an in-depth explanation of multiple parts of the problem and their proposed solution in varied media across the globe.

I find that in some cases the information is scattered so widely that I must compile this information in an orderly, concise fashion in a single document to provide a good overall understanding of the problem and a well thought out plan emerges as a solution.

The following is one such example that treats many parts of the system and multiple solutions with their pros and cons laid out so the designer can properly choose the right one for the product they are about to design.

Piezoelectric transducers

Piezoelectric sensors (For many sensing techniques, the sensor can act as both a sensor and an actuator — often the term transducer is preferred when the device acts in this dual capacity, but most piezo devices have this property of reversability whether it is used or not 1 ) measure dynamic phenomena such as force, pressure and acceleration (including shock and vibration).

In industry there are high- and low-impedance piezoelectric transducers.

High-impedance transducers have a charge output that requires a charge amplifier or external impedance converter for charge-to-voltage conversion.

Low-impedance transducers have a miniature charge-to-voltage converter internally, typically a field effect transistor (FET). Frequently these devices require an external power supply to energize the electronics and decouple the dc bias voltage from the output signal. Measurement Specialties ACH series of accelerometers is one such example of this type of transducer.

The signal conditioning task as part of the measurement system is to couple information, contained within the small amount of electrical charge generated by the crystals, to the outside world without dissipating it or otherwise changing it. (The quantity of charge generated by the piezo element is measured in units of picocoulombs (pC), which is 1 x 10-12 coulombs.)

The two types of signal-conditioning systems most often used are the charge-Mode system using charge amplifiers and the Low Impedance Voltage Mode (LIVM) System typically using a junction field effect transistor (JFET) or metal-oxide semiconductor field-effect transistor (MOSFET) op amp.

In the following section I will try to help designers make the best choice between these systems a little easier by pointing out the advantages and limitations of each type.

Charge amplifiers2

Charge amplifiers are typically made up of a high-gain inverting voltage amplifier with a MOSFET or JFET high-impedance input to achieve high insulation resistance. The input stage of the charge amplifier uses a capacitive feedback circuit to balance or “null” the effect of the applied input charge signal.

The feedback signal is then a measure of input charge. This amplifier presents essentially infinite input impedance to the sensor and thus measures its output without changing it – the goal of all measurement processes. The gain (transfer function) of the basic charge amplifier is dependent only upon the value of the feedback capacitor Cf (see Fig. 1 ) and is independent of input capacitance, an important feature of the charge amplifier. Following stages may add voltage gain and attenuation, filtering, and other functions to further process and refine the data.

Fig. 1: Simplified charge amplifier model. (Courtesy of Kistler Instruments.)

Ct = transducer capacitance

Cc = cable capacitance

Cr = range (or feedback) capacitor

Rt = time constant resistor (or insulation of range capacitor)

Ri = insulation resistance of input circuit (cable and transducer)

q = charge generated by the transducer

Vo = output voltage

A = open-loop gain

Charge-mode advantages3

Since there are no electronic components contained within the sensor housing, the upper temperature limit of charge mode sensors is much higher than the +250°F (121°C) limit imposed by the internal electronics of LIVM sensors. Rather, the high-temperature limit is set by the Curie temperature of the piezoelectric material or by the properties of insulating materials employed in the specific design.

• Laboratory type charge amplifiers currently available offer a wide range of signal augmentation choices such as filtering, ranging, standardization, integrating for velocity and displacement, peak hold and more — all conveniently contained in one package.

• Charge amplifier gain is independent of input capacitance, therefore system sensitivity is unaffected by changes in input cable length or type, an important point when interchanging cables.

• A special type of charge amplifier, the very long time constant “Electrostatic” type, used in conjunction with certain quartz element charge mode force and pressure sensors can, with certain precautions, be used to make near static (quasi-static) measurements of events lasting up to several minutes duration.

The effects of Rt and Ri will be discussed below2. Neglecting their effects, the resulting output voltage becomes:

Equation 1 (Courtesy of Kistler Instrument.)

For sufficiently high open-loop gain, the cable and transducer capacitance can be neglected and the output voltage depends only on the input charge and the range capacitance.

Equation 2 (Courtesy of Kistler Instrument.)

In summary, the amplifier acts as a charge integrator which compensates the transducers electrical charge with a charge of equal magnitude and opposite polarity and ultimately produces a voltage across the range capacitor. In effect, the purpose of the charge amplifier is to convert the high-impedance charge input (q) into a usable output voltage

Time constant and drift

Two of the more important considerations in the practical use of charge amplifiers are time constant and drift. The time constant is defined as the discharge time of an AC coupled circuit. In a period of time equivalent to one time constant, a step input will decay to 37% of its original value.

Time constant (TC ) of a charge amplifier is determined by the product of the range capacitor (Cr ) and the time constant resistor (Rt ):

TC = Rt Cr

Equation 3 (Courtesy of Kistler Instrument.)

Drift is defined as an undesirable change in output signal over time which is not a function of the measured variable. Drift in a charge amplifier can be caused by low insulation resistance at the input (Ri ) or by leakage current of the input MOSFET or J-FET.

Drift and time constant simultaneously affect a charge amplifier's output. One or the other will be dominant. Either the charge amplifier output will drift towards saturation (power supply) at the drift rate OR it will decay towards zero at the time constant rate.

The charge amplifier can have selectable time constants which are altered by changing the time constant resistor (Rt). These charge amplifiers can have a “Short”, “Medium” or “Long” time constant. In the “Long” time constant design, drift dominates any time constant effect. As long as the input insulation resistance (Rj) is maintained at greater than 1013 ohms, the charge amplifier (with MOSFET input) will drift at an approximate rate of 0.03 pC/s. Charge amplifiers with J-FET inputs are available for industrial applications but have an increased drift rate of about 0.3 pC/s.

For the “Short” and “Medium” time constants, the time constant effect dominates normal leakage drift.

Frequency and time domain considerations2

When considering the effects of time constant, the user must think in terms of either frequency or time domain. The longer the time constant, the better the low-end frequency response and the longer the usable measuring time. When measuring vibration, time constant has the same effect as a single-pole, high-pass (HP) filter whose amplitude and phase are:

phase lead (deg) = arc tan

Equation 4 (Courtesy of Kistler Instrument.)

For example, the output voltage has declined approximately 5% when f x (TC) equals 0.5 and the phase lead is 18 deg.

When measuring events with wide (or multiple) pulse widths, the time constant should be at least 100 times longer than the total event duration. Otherwise, the dc component of the output signal will decay towards zero before the event is completed.

Other design features can be incorporated into charge amplifiers including range normalization for whole number output, low-pass filters for attenuating transducer resonant effects and electrical isolation for minimizing ground loops.

Because of the very high input impedance of the charge amplifier, the sensor must be connected to the amplifier input with low-noise coaxial cable. The cable is specially treated to minimize triboelectric noise, e.g., noise generated within the cable due to physical movement of the cable. Coaxial cable is necessary to affect an electrostatic shield around the high-impedance input lead, precluding extraneous noise pickup.

Low-Impedance Piezoelectric Transducers2

Piezoelectric transducers with miniature, built-in charge-to-voltage converters (JFET or MOSFET) are usually low-impedance buffered units. These units utilize the same types of piezoelectric sensing element(s) as their high-impedance counterparts.

This concept has become known as low impedance or voltage mode. PIEZOTRON is a Kistler Instrument patented design for a low-impedance transducer with internal circuitry.

Fig. 2: PIEZOTRON circuit and coupler. (Courtesy of Kistler Instrument.)

q = charge generated by piezoelectric element

Vi = input signal at gate

V0 = output voltage (usually bias decoupled)

Cq = transducer capacitance

Cr = range capacitance

CG = MOSFET gate capacitance

Rt = time constant resistor

Time constant

The time constant of a low-impedance or voltage-mode transducer is:

TC = Rt (Cq + Cr + CG )

Equation 5 (Courtesy of Kistler Instrument.)

So the time constant is the product of its hybrid charge amplifier's range capacitor and time constant resistor.

Time constant effects in low-impedance transducers and in charge amplifiers are the same. That is, both act as a single-pole high-pass filter.

Low-impedance Power Supply

All of the low-impedance types mentioned above require excitation for their built-in electronics. A single two-wire coaxial cable and a Power Supply is all that is needed. Both the power into and the signal out from the transducer are transmitted over this two-wire cable. The coupler provides the constant current excitation required for linear operation over a wide voltage range and also decouples the bias voltage from the output.

Time constant

Bias decoupling methods can be categorized as ac or dc. Dc methods of bias decoupling will not affect a low-impedance transducer's time constant and therefore permit optimum low-frequency response. An offset voltage adjust is used to “zero” the bias. Ac decoupling methods, however, can shorten the low-impedance transducer's time constant and degrade low-frequency response. In low-impedance systems, with ac bias decoupling the system time constant can be approximated by taking the product of the transducer and coupler time constants and dividing by their sum. The resulting frequency response can be computed as before.

Proper Interface Circuitry for Low-Impedance Transducers

An Operational Amplifier (Op Amp) is used to provide amplification as well as frequency shaping of the transducer output signal. A great in-depth analysis of the interface circuitry is provided in application note by Texas Instruments.4

For those who would like a highly integrated solution, Analog Devices has the ADIS16227, a full 3-axis accelerometer which has sense electronics, A to D converter, digital filter, power management and digital I/O bus5.

High- And Low-impedance System Comparison3

Similarities

Both systems use the same type of piezoelectric sensing element(s) and therefore are AC coupled systems with limited low-frequency response or quasi-static measuring capability. Their respective time constants determine the usable frequency range.

High-impedance systems

Usually high-impedance systems are more versatile than low impedance. Time constant, gain, normalization and reset are all controlled via an external charge amplifier. In addition, the time constants are usually longer with high-impedance systems allowing easy short-term static calibration. Because they contain no built-in electronics, they have a wider operating temperature range.

Low-impedance systems

Generally, low-impedance systems are tailored to a particular application. Since the low-impedance transducer has an internally fixed range and time constant, it may limit use to their intended application. High-impedance systems, with control of range and time constant via an external charge amplifier, have no such restriction.

However, for applications with well-defined measuring frequency and temperature ranges, low-impedance systems offer a potentially lower cost (i.e., charge amplifier vs. coupler cost) alternative to high-impedance systems. In addition, low-impedance transducers can be used with general-purpose cables in environments where high humidity/contamination could be detrimental to the high insulation resistance required for high-impedance transducers. Also, longer cable lengths, between transducer and signal conditioner and compatibility with a wide range of signal display devices are further advantages of low-impedance transducers.

So as a designer, you should now be able to design the optimum front end solution for the many types of piezoelectric transducers out there on the market. Be sure to take advantage of suppliers’ websites with amplifier Evaluation Boards, Gerber files and modeling software such as Spice for your front-end designs. These tools will help ensure a robust and stable design over temperature. ■

References

1—Wikipedia

2— “The Piezoelectric Effect, Theory, Design and usage” by Kistler Instrument Corporation

3—“Piezoelectric Measurement System Comparison: Charge Mode vs. Low Impedance Voltage Mode (LIVM)” by Dytran Instruments

4—Jim Karki, “3-V Accelerometer featuring TLV2272”, Texas Instruments, 1998, #SLVA040

5—Richard Comerford, “Use ‘Good Vibrations’ to Extend System Integrity”, Electronic Products, 2011

About the Author

Steve Taranovich is a 39 year veteran analog electronics designer, applications engineer and global account manager with a passion for technical writing and education, IEEE Senior member, Chairman of the Educational Activities Committee for the IEEE Long Island and member of Eta Kappa Nu Electrical Engineering Honor Society. Steve is a graduate of NYU Engineering with a BEEE and Polytechnic University with an MSEE.