Interface Electronic Circuit

A system designer is rarely able to connect a sensor directly to processing, monitoring, or

recording instruments unless a sensor has a built-in electronic circuit with an appropriate output
format. When a sensor generates an electric signal, that signal often is either too weak, or too
noisy, or it contains undesirable components. Besides, the sensor output may be not compatible
with the input parameters of a data acquisition system, that is, it may have a wrong output
format. To mate a sensor and a processing device, they either must share a “common value” or
use in-between some kind of a “mating” device. In other words, signals from a sensor usually
have to be conditioned and modified before they are fed into a processing device (a load). Such
a load usually requires voltage or current as its analog input signal, or a digital code.
Nowadays, it is preferable if the sensor’s output is preprocessed and presented at the output in
a ready to-use form. An example is an accelerometer that outputs a digital signal with an
encoded number of measured g. Thus, a great majority of sensors that produce analog signals
require the interface circuits. Trend in modern sensor designs focuses on integration of sensing
components with signal conditioning, converting, and communication circuits. Such a combina
tion is called a sensing module. As an illustration, consider Fig. below that shows an integrated
sensing module having two sensing elements that selectively respond to two input stimuli.
For operation, a sensing element may require some supporting parts

Since a typical sensing element produces low-level analog signals, its output signals need
amplification, filtering, impedance matching, and perhaps a level shifting, before it can be
digitized. All these functions are performed by signal conditioners. Since a module may
comprise more than one channel, outputs of all signal conditioners should be converted into a
common digital format. One option is to have one analog-to-digital converter (ADC) per
channel, but in most cases it is more convenient and economical to have a single high-quality
ADC common for all channels. Thus, the outputs of all signal conditioners shall be connected
one at a time to the common ADC. This function is performed by a switching analog gate or
multiplexer (MUX).
ADC generates a digital code that is fed into a processor for an on-board computation of the
input stimuli.
Signal Conditioners A signal conditioning circuit has a specific purpose—to bring a signal from
the sensing element up to the format that is compatible with the load device—typically an ADC.
To do its job effectively, a signal conditioner must be a faithful slave of two masters: the sensing
element (sensor) and the load device. The signal conditioner input characteristics shall be
compatible to the output characteristics of the sensor while its output should generate voltage
for ease of interfacing with an ADC or other load.
The front end of a signal conditioner depends on type of the sensor’s output electrical
characteristics. Table 6.1 lists five basic types of the sensor output properties: voltage, current,
resistive, capacitive, and inductive. Selecting the appropriate input stage of a signal conditioner
is essential for the optimal data collection. A difference between a voltage-generating and
current-generating sensors should be clearly understood. The former has a relatively low-
output impedance. Its output voltage is little dependent on the load but the current is function
of the load. This sensor resembles a battery whose voltage is driven by a stimulus.

Sensor types and corresponding inputs of signal conditioners

Amplifiers
Many sensing elements produce weak output signals. Magnitudes of these signals may be on
the order of microvolts (μV) for the voltage-generating sensors or picoamperes (pA) for the
current-generating sensors. On the other hand, standard electronic data processors, such as
analog-to-digital converters (ADC), frequency modulators, data recorders, etc. require input
signals of sizable magnitudes—in the order of volts (V). Therefore, amplification of the sensor
output signals has to be made with a voltage gain up to 10,000 and a current gain up to one
million. Amplification is part of a signal conditioning. There are several standard
configurations of the amplifiers that might be useful for amplifying low-level signals. These
amplifiers may be built of discrete components, such as semiconductors, resistors, capacitors,
and inductors. Nowadays, amplifiers are frequently composed of standard building blocks,
such as operational amplifiers (OP-AMPs) that are augmented with various discrete
components.
The purpose of an amplifier is much broader than just increasing the signal magnitude. An
amplifier also is an impedance matching device, an enhancer of a signal-to-noise ratio, a
frequency filter, and an isolator between the sensor and rest of the circuit .
Operational Amplifiers
One of the principle building blocks for amplifiers is the so-called operational amplifier or OP-
AMP, which is either an integrated (monolithic) or hybrid (a combination of monolithic and
discrete parts) circuit. An integrated OP-AMP may contain hundreds of transistors, diodes, as
well as resistors and capacitors. An analog circuit designer, by arranging around the OP-AMP
discrete components (resistors, capacitors, inductors, etc.), may create endless number of useful
circuits—not only amplifiers, but also many other circuits as well. OP-AMPs are also used as
cells in custom-made integrated circuits—analog or mixed technology. A custom circuit is
called application-specific integrated circuit or ASIC, for short. Below, we will describe some
typical circuits with OP-AMP, which are often used as front ends of various signal-conditioning
circuits.
As a building block, a good operational amplifier has the following properties (a symbol
representation of OP-AMP is shown in Fig. below

Two inputs: one is inverting ( ) and the other is noninverting (+).– High-input resistance (on
the order of GΩ).– Low-output resistance (a fraction of Ω), mostly independent of a load.–
Ability to drive capacitive loads without becoming unstable.– low input offset voltage e0 (few
mV or even few μV).– low input bias current i0 (few pA or even less).– Very high open-loop
gain AOL (at least 104 and preferably over 106). That is, the OP-AMP must be able to
magnify(amplify) a voltage difference Vin, between its two inputs by a factor of AOL.
An example of a feedback with a resistive divider R1 and R2 is shown in Fig. below. The input
voltage is applied to a noninverting (+) input of the OP-AMP. This input has a very high-input
impedance. The feedback resistors convert the Op-AMP to a noninverting amplifier, where the
resulting closed-loop gain.

Considering the AOL being very large, the closed-loop gain A depends only on the feedback
components and is nearly constant over a broad frequency range. However, f1 is still the
frequency limiting factor, regardless of the feedback. Linearity, gain stability, and output
impedance—all are improved by the amount of a feedback. The feedback may have various
linear components, including resistors, capacitors, inductors, as well as nonlinear components,
such as diodes.

Voltage Follower:
A voltage follower shown in Fig. below is an electronic circuit that provides impedance
conversion from a high to low level. It is a particular case of the amplifier , where R1 is removed
(“infinite” value) and R2=0. A typical follower has high-input impedance (very high-input
resistance and low-input capacitance) and very low-output resistance (the output capacitance
makes no difference). A good follower has a voltage gain being very close to unity (typically,
0.999 over a broad frequency range). The buffering properties—high-input and low-output
impedances make it indispensable for interfacing between many sensors and signal processing
devices.

Charge- and Current-to-Voltage Converters:

Charge-to-voltage converters (CVC) are employed to convert signals from the charge-
generating sensors. Like a voltage follower, a CVC is a buffer between the charge sensor and
ADC that requires voltage from a low-impedance source. A basic circuit of a CVC is shown in
Fig. below.

Capacitor, C, is connected into a negative feedback network of an OP-AMP. Its leakage

resistance R must be substantially larger than impedance of the capacitor at the lowest operating
frequency. A good film capacitor is usually recommended along with a good quality printed
circuit board where the components are coated with conformal coating. A transfer function of
the converter is:

If the sensor is of a capacitive nature (many charge-generating sensors are),

connecting it to an inverting input of OP-AMP may cause instability at high frequencies. In
other words, the amplifier may oscillate—a highly undesirable behavior. To prevent
oscillations, a small resistor r should be added in series with the capacitive sensor.

Sensor Connections
A sensor may be directly connected to a signal conditioner, but often it is desirable to reduce
errors and noise caused by various interfering sources even before the signal is conditioned.
Some errors can be reduced or even eliminated by use of special sensor connections in front of
the signal conditioner. Below are descriptions of the most popular techniques.

1. Ratiometric Circuits
A powerful method of improving accuracy of a sensor is a ratiometric technique. It should be
emphasized, however, that the method is useful only if a source of error has a multiplicative
nature but not additive. That is, the technique is useless for reduction, for instance, thermal
noise. On the other hand, it is quite potent to solve problems as dependence of sensor’s
sensitivity to such factors as power supply instability, ambient temperature, humidity, pressure,
effects of aging, etc. The technique essentially requires the use of two sensors where one is the
acting sensor that responds to an external stimulus and the other is a compensating sensor,
which is either shielded from that stimulus or is insensitive to it. Both sensors must be exposed
to all other external effects, which may multiplicatively change their performance. The second
sensor, which is often called reference, must be subjected to a reference stimulus, which is
ultimately stable during the lifetime of the product. In many practical systems, the reference
sensor is not necessarily exactly similar to the acting sensor, however its physical properties,
which are subject to instabilities, should be the same. A ratiometric technique essentially
requires the use of a division. It can be performed by two standard methods: digital and analog.
In a digital form, output signals from both the acting and the reference sensors are multiplexed
and converted into binary codes in an analog-to-digital converter (ADC). Subsequently, a
computer or a microprocessor performs the operation of a division. In an analog form, a divider
may be part of a signal conditioner. A “divider”, Fig. below, produces an output voltage or
current proportional to a ratio of two input voltages or currents or numbers:

where the numerator is denoted as VN, the denominator VD, and k is equal
to the output voltage, when VN=VD.

2. Differential Circuits
Beside multiplicative interferences, the additive interferences are very common and pose
serious problems for low-level output signals. Consider for example a pyro electric sensor,
where a heat flow sensitive ceramic plate supports inside a metal can. Since a pyroelectric
is also a piezoelectric, besides heat flow the sensor is susceptible to mechanical stress
interferences. Even a slight vibration will generate a spurious piezoelectric signal that may
be several orders of magnitude higher than a pyroelectric current. The solution is to
fabricate the sensor with dual electrodes deposited on the same ceramic substrate. This
essentially creates two identical sensors on the same ceramic plate. Both sensors respond
to all stimuli nearly identically. Since they are oppositely connected and assuming that
Vpyro and Vpiezo from one sensor are respectively equal to those of the other sensor, the
resulting output voltage is essentially zero:

If one of the sensors is blocked from receiving thermal radiation (Vpyro2=0), then Vout
=Vpyro1. In other words, thanks to subtraction (Vpiezo1=Vpiezo2 are subtracted), the
combined sensor becomes insensitive to piezoelectric spurious signals.