Sunteți pe pagina 1din 60

Contents

3.1 Analog Signal Processing 3.5 Amplifiers


3.1.1 Linearity 3.5.1 Amplifier Configurations
3.1.2 Linear Operations 3.5.2 Transistor Amplifiers
3.1.3 Nonlinear Operations 3.5.3 Bipolar Transistor Amplifiers
3.2 Analog Devices 3.5.4 FET Amplifiers
3.2.1 Terminology 3.5.5 Buffer Amplifiers
3.2.2 Gain and Transconductance 3.5.6 Cascaded Buffers
3.2.3 Characteristic Curves 3.5.7 Using the Transistor as a Switch
3.2.4 Manufacturers Data Sheets 3.5.8 Choosing a Transistor
3.2.5 Physical Electronics of Semiconductors 3.6 Operational Amplifiers
3.2.6 The PN Semiconductor Junction 3.6.1 Characteristics of Practical Op-Amps
3.2.7 Junction Semiconductors 3.6.2 Basic Op Amp Circuits
3.2.8 Field-Effect Transistors (FET) 3.7 Analog-Digital Conversion
3.2.9 Semiconductor Temperature Effects 3.7.1 Basic Conversion Properties
3.2.10 Safe Operating Area (SOA) 3.7.2 Analog-to-Digital Converters (ADC)
3.3 Practical Semiconductors 3.7.3 Digital-to-Analog Converters (DAC)
3.3.1 Semiconductor Diodes 3.7.4 Choosing a Converter
3.3.2 Bipolar Junction Transistors (BJT) 3.8 Miscellaneous Analog ICs
3.3.3 Field-Effect Transistors (FET) 3.8.1 Transistor and Driver Arrays
3.3.4 Optical Semiconductors 3.8.2 Voltage Regulators and References
3.3.5 Linear Integrated Circuits 3.8.3 Timers (Multivibrators)
3.3.6 Comparison of Semiconductor Devices for 3.8.4 Analog Switches and Multiplexers
Analog Applications 3.8.5 Audio Output Amplifiers
3.4 Analog Systems 3.8.6 Temperature Sensors
3.4.1 Transfer Functions 3.8.7 Electronic Subsystems
3.4.2 Cascading Stages 3.9 Analog Glossary
3.4.3 Amplifier Frequency Response 3.10 References and Bibliography
3.4.4 Interstage Loading and Impedance Matching
3.4.5 Noise
3.4.6 Buffering

Chapter 3 CD-ROM Content


Supplemental Articles
Hands-On Radio: The Common Emitter Amplifier Hands-On Radio: Load Lines by Ward Silver, NAX
by Ward Silver, NAX Hands-On Radio: The Effects of Gain-Bandwidth
Hands-On Radio: The Emitter-Follower Amplifier Product by Ward Silver, NAX
by Ward Silver, NAX Cathode Ray Tubes
Hands-On Radio: The Common Base Amplifier Large Signal Transistor Operation
by Ward Silver, NAX
Hands-On Radio: Field Effect Transistors Tools and Data
by Ward Silver, NAX LTSpice Simulation Files for Chapter 3
Hands-On Radio: Basic Operational Amplifiers Frequency Response Spreadsheet
by Ward Silver, NAX
Chapter 3
Analog Basics

The first section of this chapter 3.1 Analog Signal Processing


discusses a variety of techniques for
working with analog signals, called The term analog signal refers to voltages, currents and waves that make up ac radio and
signal processing. These basic func- audio signals, dc measurements, even power. The essential characteristic of an analog signal
tions are combined to produce useful is that the information or energy it carries is continuously variable. Even small variations of
systems that become radios, instru- an analog signal affect its value or the information it carries. This stands in contrast to digital
ments, audio recorders and so on. signals (described in the Digital Basics chapter) that have values only within well-defined
Subsequent sections present sev- and separate ranges called states. To be sure, at the fundamental level all circuits and signals
eral types of active components and are analog: Digital signals are created by designing circuits that restrict the values of analog
circuits that can be used to manipulate signals to those discrete states.
analog signals. (An active electronic Analog signal processing involves various electronic stages to perform functions on analog
component is one that requires a signals such as amplifying, filtering, modulation and demodulation. A piece of electronic
power source to function; passive equipment, such as a radio, is constructed by combining a number of these circuits. How these
components such as resistors, capaci- stages interact with each other and how they affect the signal individually and in tandem is
tors and inductors are described in the the subject of sections later in the chapter.
Electrical Fundamentals chapter.)
The chapter concludes with a discus-
sion of analog-digital conversion by 3.1.1 Linearity
which analog signals are converted
into digital signals and vice versa. The premier properties of analog signals are superposition and scaling. Superposition is
Material in this chapter was updated the property by which signals are combined, whether in a circuit, in a piece of wire, or even
by Ward Silver, NAX, and Courtney in air, as the sum of the individual signals. This is to say that at any one point in time, the
Duncan, N5BF, building on material voltage of the combined signal is the sum of the voltages of the original signals at the same
from previous editions by Greg Lapin, time. In a linear system any number of signals will add in this way to give a single combined
N9GL, and Leonard Kay, K1NU. signal. (Mathematically, this is a linear combination.) For this reason, analog signals and
The section Poles and Zeroes was components are often referred to as linear signals or linear components. A linear system whose
adapted from material contributed by characteristics do not change, such as a resistive voltage divider, is called time-invariant. If
David Stockton, GM4ZNX and Fred the system changes with time, it is time-varying. The variations may be random, intermittent
Telewski, WA7TZY. (such as being adjusted by an operator) or periodic.
One of the more important features of superposition, for the purposes of signal process-
ing, is that signals that have been combined by superposition can be separated back into the
original signals. This is what allows multiple signals that have been received by an antenna
to be separated back into individual signals by a receiver.

3.1.2 Linear Operations


Any operation that modifies a signal and obeys the rules of superposition and scaling is
a linear operation. The following sections explain the basic linear operations from which
linear systems are made.

AMPLIFICATION AND ATTENUATION


Amplification and attenuation scale signals to be larger and smaller, respectively. The
operation of scaling is the same as multiplying the signal at each point in time by a constant
value; if the constant is greater than one then the signal is amplified, if less than one then the
signal is attenuated.
An amplifier is a circuit that increases the amplitude of a signal. Schematically, a generic

Analog Basics 3.1


amplifier is signified by a triangular symbol,
its input along the left face and its output
at the point on the right (see Fig 3.1). The Obtaining a Frequency Response
linear amplifier multiplies every value of a With the computer tools such as spreadsheets, its easy to do the calculations
signal by a constant value. Amplifier gain and make a graph of frequency response. If you dont have a spreadsheet program,
is often expressed as a multiplication factor then use semi-log graph paper with the linear axis used for dB or phase and the
logarithmic axis for frequency. An Excel spreadsheet set up to calculate and display
( 5, for example).
frequency response is available on the CD that comes with this book. You can modify
V it to meet your specific needs.
=I IS e Vt 1 (1) Follow these rules whether using a spreadsheet or graph paper:


Measure input and output in the same units, such as volts, and use the same
where Vo is the output voltage from an am- conventions, such as RMS or peak-to-peak.
Measure phase shift from the input to the output. (The Test Equipment and
plifier when an input voltage, Vi, is applied.
Measurements chapter discusses how to make measurements of amplitude and
(Gain is often expressed in decibels (dB) phase.)
as explained in the chapter on Electrical Use 10 log (PO/PI) for power ratios and 20 log (VO/VI) for voltage or current.
Fundamentals.) To make measurements that are roughly equally spaced along the logarithmic fre-
Certain types of amplifiers for which cur- quency axis, follow the 1-2-5 rule. Dividing a range this way, for example 1-2-5-10-
rents are the input and output signals have 20-50-100-200-500 Hz, creates steps in approximately equal ratios that then appear
current gain. Most amplifiers have both volt- equally spaced on a logarithmic axis.
age and current gain. Power gain is the ratio
of output power to input power.
Ideal linear amplifiers have the same gain
for all parts of a signal. Thus, a gain of 10 put voltages (and currents) that are within earity is called slew rate. Applied to an ampli-
changes 10 V to 100 V, 1 V to 10 V and 1 V the range of its power supply. (Power-supply fier, this term describes the maximum rate at
to 10 V. (Gain can also be less than one.) The voltages are also called the rails of a circuit.) which a signal can change levels and still be
ability of an amplifier to change a signals level As the amplified output approaches one of the accurately amplified in a particular device. In-
is limited by the amplifiers dynamic range, rails, the output can not exceed a given voltage put slew rate is the maximum rate of change to
however. An amplifiers dynamic range is the near the rail and the operation of the amplifier which the amplifier can react linearly. Output
range of signal levels over which the amplifier becomes nonlinear as described below in the slew rate refers to the maximum rate at which
produces the required gain without distortion. section on Clipping and Rectification. the amplifiers output circuit can change. Slew
Dynamic range is limited for small signals Another similar limitation on amplifier lin- rate is an important concept, because there is
by noise, distortion and other nonlinearities.
Dynamic range is limited for large signals
because an amplifier can only produce out-

Fig 3.1 Generic amplifier. (A) Symbol.


For the linear amplifier, gain is the
constant value, G, and the output voltage
is equal to the input voltage times G; (B)
Transfer function, input voltage along the
x-axis is converted to the output voltage
along the y-axis. The linear portion of the
response is where the plot is diagonal; its
slope is equal to the gain, G. Above and
below this range are the clipping limits,
where the response is not linear and the Fig 3.2 Bode plot of (A) band-pass filter magnitude response and (B) an RC low-
output signal is clipped. pass filter phase response.

3.2 Chapter 3
a direct correlation between a signal levels COMPLEX FREQUENCY wave with which we are familiar and that has
rate of change and the frequency content of We are accustomed to thinking of frequency frequency f, where f = /2. The first term,
that signal. The amplifiers ability to react to as a real number so many cycles per second est, represents the rate at which the signal
or reproduce that rate of change affects its but frequency can also be a complex increases or decreases. If is negative, the
frequency response and dynamic range. number, s, with both a real part, designated exponential term decreases with time and the
An attenuator is a circuit that reduces the by , and an imaginary part, designated by signal gets smaller and smaller. If is positive,
amplitude of a signal. Attenuators can be con- j. ( is also equal to 2pf.) The resulting the signal gets larger and larger. If = 0, the
structed from passive circuits, such as the at- complex frequency is written as s = + j. exponential term equals 1, a constant, and the
tenuators built using resistors, described in the At the lower left of Fig 3.3 a pair of real and signal amplitude stays the same.
chapter on Test Equipment and Measure- imaginary axes are used to plot values of s. Complex frequency is very useful in
ments. Active attenuator circuits include am- This is called the s-plane. Complex frequency describing a circuits stability. If the response
plifiers whose gain is less than one or circuits is used in Laplace transforms, a mathematical to an input signal is at a frequency on the
with adjustable resistance in the signal path, technique used for circuit and signal analysis. right-hand side of the s-plane for which
such as a PIN diode attenuator or amplifier (Thorough treatments of the application of > 0, the system is unstable and the output
with gain is controlled by an external voltage. complex frequency can be found in college- signal will get larger until it is limited by
level textbooks on circuit and signal analysis.) the circuits power supply or some other
FREQUENCY RESPONSE AND When complex frequency is used, a mechanism. If the response is on the left-hand
BODE PLOTS sinusoidal signal is described by Aest, where side of the s-plane, the system is stable and
Another important characteristic of a cir- A is the amplitude of the signal and t is the response to the input signal will eventually
cuit is its frequency response, a description time. Because s is complex, Aest = Ae(+j)t die out. The larger the absolute value of , the
of how it modifies a signal of any frequency. = A(est)(ejwt). The two exponential terms faster the response changes. If the response
Frequency response can be stated in the form describe independent characteristics of the is precisely on the j axis where = 0, the
of a mathematical equation called a transfer signal. The second term, ejwt, is the sine response will persist indefinitely.
function, but it is more conveniently presented
as a graph of gain vs frequency. The ratio of
output amplitude to input amplitude is often
called the circuits magnitude or amplitude
response. Plotting the circuits magnitude
response in dB versus frequency on a loga-
rithmic scale, such as in Fig 3.2A, is called
a Bode plot (after Henrik Wade Bode). The
combination of decibel and log-frequency
scales is used because the behavior of most
circuits depends on ratios of amplitude and
frequency and thus appears linear on a graph
in which both the vertical and horizontal
scales are logarithmic.
Most circuits also affect a signals phase
along with its amplitude. This is called phase
shift. A plot of phase shift from the circuits
input to its output is called the phase response,
seen in Fig 3.2B. Positive phase greater than
0 indicates that the output signal leads the
input signal, while lagging phase shift has a
negative phase. The combination of an am-
plitude and phase response plot gives a good
picture of what effect the circuit has on signals
passing through it.

TRANSFER CHARACTERISTICS
Transfer characteristics are the ratio of an
output parameter to an input parameter, such
as output current divided by input current, hFE.
There are different families of transfer char-
acteristics, designated by letters such as h, s,
y or z. Each family compares parameters in
specific ways that are useful in certain design
or analysis methods. The most common
transfer characteristics used in radio are the
h-parameter family (used in transistor models)
and the s-parameter family (used in RF design,
particularly at VHF and above). See the RF Fig 3.3 The transfer function for a circuit describes both the magnitude and phase
Techniques chapter for more discussion of response of a circuit. The RC circuit shown at the upper left has a pole at f = 2RC, the
transfer characteristics. filters 3 dB or cutoff frequency, at which the phase response is a 45 lagging phase
shift. Poles cause an infinite response on the imaginary frequency axis.

Analog Basics 3.3


In Fig 3.3 the equation for the simple RC- frequencies, even though we cant experience ally not found by themselves, making the
circuits transfer function is shown at the imaginary frequencies directly. A pole is as- magnitude response go up, but rather paired
left of the figure. It describes the circuits sociated with a bend in a magnitude response with a pole of a different frequency, result-
behavior at real-world frequencies as well as plot that changes the slope of the response ing in the magnitude response having a slope
imaginary frequencies whose values contain downward with increasing frequency by 6 dB between two frequencies but flat above and
j. Because complex numbers are used for f, per octave (20 dB per decade; an octave is a below them.
the transfer function describes the circuits 2:1 frequency ratio, a decade is a 10:1 fre- Real-world circuit zeroes are only found
phase response, as well as amplitude. At one quency ratio). accompanied by a greater or equal number
such frequency, f = j/2pRC, the denominator There are four ways to identify the exis- of poles. Consider a classic RC high-pass
of the transfer function is zero, and the gain tence and frequency of a pole as shown in filter, such as if the resistor and capacitor in
is infinite! Infinite gain is a pretty amazing Fig 3.3: Fig 3.3 were exchanged. The response of such
thing to achieve with a passive circuit but 1. For a downward bend in the magni- a circuit increases at 6 dB per octave from
because this can only happen at an imaginary tude versus frequency plot, the pole is at the 0 Hz (so there must be a zero at 0 Hz) and
frequency, it does not happen in the real world. 3 dB frequency for a single pole. If the bend then levels off at 1/2RC Hz. This leveling
The practical effects of complex frequency causes a change in slope of more than 6 dB/ off is due to the presence of a pole adding its
can be experienced in a narrow CW crystal octave, there must be multiple poles at this 6 dB-per-octave roll-off to cancel the 6 dB-
or LC filter. The poles of such a filter are frequency. per-octave roll-up of the zero. The transfer
just to the left of the j axis, so the input 2. A 90 lagging change on a phase versus function for such as circuit would equal zero
signal causes the filters to ring, or output frequency plot, where the lag increases with at zero frequency and infinity at the imaginary
a damped sine wave along with the desired frequency. The pole is at the point of 45 pole frequency.
signal. Similarly, the complex frequency of added lag on the S-shaped transition. Multiple
an oscillators output at power-up must have poles will add their phase lags, as above. FEEDBACK AND OSCILLATION
> 0 or the oscillation would never start! 3. On a circuit diagram, a single pole looks The stability of an amplifier refers to its
The output amplitude continues to grow un- like a simple RC low-pass filter. The pole is ability to provide gain to a signal without
til limiting takes place, reducing gain until at the 3 dB frequency (f = 1/2RC Hz). Any tending to oscillate. For example, an amplifier
= 1 for a steady output. other circuit with the same response has a pole just on the verge of oscillating is not gener-
at the same frequency. ally considered to be stable. If the output
POLES AND ZEROES 4. In an equation for the transfer function of an amplifier is fed back to the input, the
Frequencies that cause the transfer function of a circuit, a pole is a theoretical value of feedback can affect the amplifier stability. If
to become infinite are called poles. This is frequency that would result in infinite gain. the amplified output is added to the input,
shown at the bottom right of Fig 3.3 in the This is clearly impossible, but as the value of the output of the sum will be larger. This
graph of the circuits amplitude response frequency will either be absolute zero, or will larger output, in turn, is also fed back. As this
for imaginary frequencies shown on the have an imaginary component, it is impos- process continues, the amplifier output will
horizontal axis. (The pole causes the graph sible to make an actual real-world signal at a continue to rise until the amplifier cannot go
to extend up as a pole under a tent, thus pole frequency. any higher (clamps). Such positive feedback
the name.) Similarly, circuits can have zeroes For example, comparing the amplitude re- increases the amplifier gain, and is called
which occur at imaginary frequencies that sponses at top and bottom of Fig 3.3 shows regeneration. (The chapter on Oscillators
cause the transfer function to be zero, a less that the frequency of the pole is equal to the and Synthesizers includes a discussion of
imaginative name, but quite descriptive. circuits 3 dB cutoff frequency (1/2fC) positive feedback.)
A circuit can also have poles and zeroes multiplied by j, which is also the frequency Most practical amplifiers have some intrin-
at frequencies of zero and infinity. For at which the circuit causes a 45 (lagging) sic and unavoidable feedback either as part
example, the circuit in Fig 3.3 has a zero at phase shift from input to output. of the circuit or in the amplifying device(s)
infinity because the capacitors reactance is itself. To improve the stability of an amplifier,
zero at infinity and the transfer function is What Is a Zero? negative feedback can be added to counteract
zero, as well. If the resistor and capacitor A zero is the complement of a pole. In any unwanted positive feedback. Negative
were exchanged, so that the capacitor was in math, it is a frequency at which the transfer feedback is often combined with a phase-
series with the output, then at zero frequency function equation of a circuit is zero. This is shift compensation network to improve the
(dc), the output would be zero because the not impossible in the real world (unlike the amplifier stability.
capacitors reactance was infinite, creating pole), so zeroes can be found at real-number Although negative feedback reduces am-
a zero. frequencies as well as complex-number fre- plifier or stage gain, the advantages of stable
Complex circuits can have multiple poles quencies. gain, freedom from unwanted oscillations and
or multiple zeroes at the same frequency. Each zero is associated with an upward the reduction of distortion are often key design
Poles and zeroes can also occur at frequencies bend of 6 dB per octave in a magnitude re- objectives and advantages of using negative
that are combinations of real and imaginary sponse. Similarly to a pole, the frequency of feedback.
numbers. The poles and zeroes of a circuit the zero is at the +3 dB point. Each zero is The design of feedback networks depends
form a pattern in the complex plane that associated with a transition on a phase-versus- on the desired result. For amplifiers, which
corresponds to certain types of circuit frequency plot that reduces the lag by 90. The should not oscillate, the feedback network
behavior. (The relationships between the zero is at the 45 leading phase point. Multiple is customized to give the desired frequency
pole-zero pattern and circuit behavior is zeroes add their phase shifts just as poles do. response without loss of stability. For oscil-
beyond the scope of this book, but are covered In a circuit, a zero creates gain that in- lators, the feedback network is designed to
in textbooks on circuit theory.) creases with frequency forever above the zero create a steady oscillation at the desired fre-
frequency. This requires active circuitry that quency.
What is a Pole? would inevitably run out of gain at some fre-
Poles cause a specific change in the circuits quency, which implies one or more poles up SUMMING
amplitude and phase response for real-world there. In real-world circuits, zeroes are usu- In a linear system, nature does most of

3.4 Chapter 3
the work for us when it comes to adding sig- half of (3 dB) the mid-band response. These The carrier is considered to be part of a time-
nals; placing two signals together naturally points are called the circuits cutoff or corner varying linear system and not a second signal.
causes them to add according to the principle or half-power frequencies. The range between A curious trait of amplitude modulation
of superposition. When processing signals, the cutoff frequencies is the filters passband. (and demodulation) is that it can be performed
we would like to control the summing op- Outside the filters passband, the amplitude nonlinearly, as well. Accurate analog multi-
eration so the signals do not distort or com- response drops to 1/200th of (23 dB) mid- pliers (and dividers) are difficult and expen-
bine in a nonlinear way. If two signals come band response at 1 Hz and only 1/1000th sive to fabricate, so less-expensive nonlinear
from separate stages and they are connected (30 dB) at 500 kHz. The steepness of the methods are often employed. Each nonlinear
together directly, the circuitry of the stages change in response with frequency is the fil- form of amplitude modulation generates the
may interact, causing distortion of either or ters roll-off and it is usually specified in dB/ desired linear combination of signals, called
both signals. octave (an octave is a doubling or halving of products, in addition to other unwanted prod-
Summing amplifiers generally use a resistor frequency) or dB/decade (a decade is a change ucts that must then be removed. Both linear
in series with each stage, so the resistors con to 10 times or 1/10th frequency). and nonlinear modulators and demodula-
nect to the common input of the following Fig 3.2B represents the phase response of tors are discussed in the chapter on Mixers,
stage. This provides some isolation between a different filter the simple RC low-pass Modulators, and Demodulators.
the output circuits of each stage. Fig 3.4 illus- filter shown at the upper right. As frequen-
trates the resistors connecting to a summing cy increases, the reactance of the capacitor
amplifier. Ideally, any time we wanted to com- becomes smaller, causing most of the input 3.1.3 Nonlinear Operations
bine signals (for example, combining an audio signal to appear across the fixed-value resis- All signal processing doesnt have to be
signal with a sub-audible tone in a 2 meter tor instead. At low frequencies, the capacitor linear. Any time that we treat various sig-
FM transmitter prior to modulating the RF has little effect on phase shift. As the signal nal levels differently, the operation is called
signal) we could use a summing amplifier. frequency rises, however, there is more and nonlinear. This is not to say that all signals
more phase shift until at the cutoff frequency, must be treated the same for a circuit to be
FILTERING there is 45 of lagging phase shift, plotted as linear. High-frequency signals are attenuated
A filter is a common linear stage in ra- a negative number. Phase shift then gradually in a low-pass filter while low-frequency sig-
dio equipment. Filters are characterized by approaches 90. nals are not, yet the filter can be linear. The
their ability to selectively attenuate certain Practical passive and active filters are de- distinction is that the amount of attenuation
frequencies in the filters stop band, while scribed in the RF and AF Filters chapter. at different frequencies is always the same,
passing or amplifying other frequencies in the Filters implemented by digital computation regardless of the amplitude of the signals pass-
passband. If the filters passband extends to or (digital filters) are discussed in the chapter on ing through the filter.
near dc, it is a low-pass filter, and if to infinity DSP and Software Radio Design. Filters at What if we do not want to treat all volt-
(or at least very high frequencies for the cir- RF may also be created by using transmission age levels the same way? This is commonly
cuitry involved), it is a high-pass filter. Filters lines as described in the Transmission Lines desired in analog signal processing for clip-
that pass a range of frequencies are band-pass chapter. All practical amplifiers are in effect ping, rectification, compression, modulation
filters. All-pass filters are designed to affect either low-pass filters or band-pass filters, and switching.
only the phase of a signal without changing because their magnitude response decreases
the signal amplitude. The range of frequencies as the frequency increases beyond their gain- CLIPPING AND RECTIFICATION
between a band-pass circuits low-pass and bandwidth products. Clipping is the process of limiting the range
high-pass regions is its mid-band. of signal voltages passing through a circuit (in
Fig 3.2A is the amplitude response for a AMPLITUDE MODULATION/ other words, clipping those voltages outside
typical band-pass audio filter. It shows that the DEMODULATION the desired range from the signals). There are
input signal is passed to the output with no Voice and data signals can be transmitted a number of reasons why we would like to
loss (0 dB) between 200 Hz and 5 kHz. This over the air by using amplitude modulation do this. As shown in Fig 3.1, clipping is the
is the filters mid-band response. Above and (AM) to combine them with higher frequency process of limiting the positive and negative
below those frequencies the response of the carrier signals (see the Modulation chapter). peaks of a signal. (Clipping is also called
filter begins to drop. By 20 Hz and 20 kHz, the The process of amplitude modulation can be clamping.)
amplitude response has been reduced to one- mathematically described as the multiplica- Clamping might be used to prevent a large
tion (product) of the voice signal and the car- audio signal from causing excessive deviation
rier signal. Multiplication is a linear process in an FM transmitter that would interfere with
amplitude modulation by the sum of two communications on adjacent channels. Clip-
audio signals produces the same signal as ping circuits are also used to protect sensitive
the sum of amplitude modulation by each inputs from excessive voltages. Clipping dis-
audio signal individually. Another aspect of torts the signal, changing it so that the original
the linear behavior of amplitude modulation signal waveform is lost.
is that amplitude-modulated signals can be Another kind of clipping results in recti-
demodulated exactly to their original form. fication. A rectifier circuit clips off all volt-
Amplitude demodulation is the converse of ages of one polarity (positive or negative)
amplitude modulation, and is represented as and passes only voltages of the other po-
division, also a linear operation. larity, thus changing ac to pulsating dc (see
Fig 3.4 Summing amplifier. The output In the linear model of amplitude modula- the Power Sources chapter). Another use
voltage is equal to the sum of the input of rectification is in a peak detection circuit
voltages times the amplifier gain, G. As
tion, the signal that performs the modulation
long as the resistance values, R, are (such as the audio signal in an AM transmit- that measures the peak value of a waveform.
equal and the amplifier input impedance ter) is shifted in frequency by multiplying it Only one polarity of the ac voltage needs
is much higher, the actual value of R does with the carrier. The modulated waveform is to be measured and so a rectifier clips the
not affect the output signal. considered to be a linear function of the signal. unwanted polarity.

Analog Basics 3.5


LIMITING used in FM receivers to amplify the signal levels more than high levels. This type of
Another type of clipping occurs when an until all amplitude variations in the signal amplification is often called signal compres-
amplifier is intentionally operated with so are removed and the only characteristic of the sion. Speech compression is sometimes used
much gain that the input signals result in an original signal that remains is the frequency. in audio amplifiers that feed modulators. The
output that is clipped at the limits of its power voice signal is compressed into a small range
supply voltages (or some other designated LOGARITHMIC AMPLIFICATION of amplitudes, allowing more voice energy to
voltages) . The amplifier is said to be driven It is sometimes desirable to amplify a signal be transmitted without overmodulation (see
into limiting and an amplifier designed for logarithmically, which means amplifying low the Modulation chapter).
this behavior is called a limiter. Limiters are

3.2 Analog Devices


There are several different kinds of com- gain. (See the section on transistor ampli- 3.2.3 Characteristic Curves
ponents that can be used to build circuits fiers later in this chapter for a discussion of Analog devices are described most com-
for analog signal processing. Bipolar semi- the common-emitter circuit.) Qualifiers are pletely with their characteristic curves. The
conductors, field-effect semiconductors and sometimes added to the subscripts to indicate characteristic curve is a plot of the interrela-
integrated circuits comprise a wide spec- certain operating modes of the device. SS for tionships between two or three variables. The
trum of active devices used in analog signal saturation, BR for breakdown, ON and OFF are vertical (y) axis parameter is the output, or
processing. (Vacuum tubes are discussed in all commonly used.
the chapter on RF Power Amplifiers, their Power supply voltages have two subscripts
primary application in Amateur Radio.) Sev- that are the same, indicating the terminal to
eral different devices can perform the same which the voltage is applied. VDD would rep-
function, each with its own advantages and resent the power supply voltage applied to the
disadvantages based on the physical charac- drain of a field-effect transistor.
teristics of each type of device. Since integrated circuits are collections of
Understanding the specific characteristics semiconductor components, the abbrevia-
of each device allows you to make educated tions for the type of semiconductor used also
decisions about which device would be best apply to the integrated circuit. For example,
for a particular purpose when designing ana- VCC is a power supply voltage for an inte-
log circuitry, or understanding why an exist- grated circuit made with bipolar transistor
ing circuit was designed in a particular way. technology in which voltage is applied to
transistor collectors.
3.2.1 Terminology
A similar terminology is used when de- 3.2.2 Gain and
scribing active electronic devices. The letter Transconductance
V or v stands for voltages and I or i for cur-
The operation of an amplifier is specified
rents. Capital letters are often used to denote
by its gain. Gain in this sense is defined as the
dc or bias values (bias is discussed later in
change () in the output parameter divided by
this chapter). Lower-case often denotes in-
the corresponding change in the input param-
stantaneous or ac values.
eter. If a particular device measures its input
Voltages generally have two subscripts
and output as currents, the gain is called a
indicating the terminals between which the
current gain. If the input and output are volt-
voltage is measured (VBE is the dc voltage
ages, the amplifier is defined by its voltage
between the base and the emitter of a bipolar
gain. Power gain is often used, as well. Gain
transistor). Currents have a single subscript
is technically unit-less, but is often given in
indicating the terminal into which the current
flows (IC is the dc current into the collector V/V. Decibels are often used to specify gain,
of a bipolar transistor). If the current flows particularly power gain.
out of the device, it is generally treated as a If an amplifiers input is a voltage and the
negative value. output is a current, the ratio of the change in
Resistance is designated with the letter R output current to the change in input voltage
or r, and impedance with the letter Z or z. For is called transconductance, gm.
example, rDS is resistance between drain and I C Fig 3.5 Characteristic curves. A
source of an FET and Zi is input impedance. 1 (2) forward voltage vs forward current
I E
For some parameters, values differ for dc and characteristic curve for a semiconductor
ac signals. This is indicated by using capital diode is shown at (A). (B) shows a set
Transconductance has the same units as of characteristic curves for a bipolar
letters in the subscripts for dc and lower-case transistor in which the collector current
subscripts for ac. For example, the common- conductance and admittance, Siemens (S), but
is only used to describe the operation of active vs collector-to-emitter voltage curve is
emitter dc current gain for a bipolar transistor plotted for five different values of base
is designated as hFE, and hfe is the ac current devices, such as transistors or vacuum tubes. current.

3.6 Chapter 3
result of the device being operated with an in- BIASING value of the output current is subtracted, a
put parameter on the horizontal (x) axis. Often The operation of an analog signal-process- reproduction of the input signal is the result.
the output is the result of two input values. ing device is greatly affected by which portion If Bias Point 2 is chosen, we can see that
The first input parameter is represented along of the characteristic curve is used to do the the input voltage is reproduced as a chang-
the x-axis and the second input parameter processing. The devices bias point is its set ing output current with the same shape. In
by several curves, each for a different value. of operating parameters when no input signal this case, the device is operating linearly. If
Almost all devices of concern are nonlinear is applied. The bias point is also known as either Bias Point 1 or Bias Point 3 is chosen,
over a wide range of operating parameters. the quiescent point or Q-point. By changing however, the shape of the output signal is dis-
We are often interested in using a device the bias point, the circuit designer can affect torted because the characteristic curve of the
only in the region that approximates a linear the relationship between the input and output device is nonlinear in this region. Either the
response. Characteristic curves are used to signal. The bias point can also be considered increasing portion of the input signal results
graphically describe a devices operation in as a dc offset of the input signal. Devices that in more variation than the decreasing por-
both its linear and nonlinear regions. perform analog signal processing require ap- tion (Bias Point 1) or vice versa (Bias Point
Fig 3.5A shows the characteristic curve propriate input signal biasing. 3). Proper biasing is crucial to ensure that a
for a semiconductor diode with the y-axis As an example, consider the characteris- device operates linearly.
showing the forward current, IF, flowing tic curve shown in Fig 3.6. (The exact types
through the diode and the x-axis showing of device and circuit are unimportant.) The
forward voltage, VF, across the diode. This
3.2.4 Manufacturers
characteristic curve shows the relationship
curve shows the relationship between current Data Sheets
between an input voltage and an output cur-
and voltage in the diode when it is conduct- rent. Increasing input voltage results in an Manufacturers data sheets list device char-
ing current. Characteristic curves showing increase in output current so that an input acteristics, along with the specifics of the part
voltage and current in two-terminal devices signal is reproduced at the output. The char- type (polarity, semiconductor type), identity
such as diodes are often called I-V curves. acteristic curve is linear in the middle, but is of the pins and leads (pinouts), and the typi-
Characteristic curves may include all four quite nonlinear in its upper and lower regions. cal use (such as small signal, RF, switching
quadrants of operation in which both axes In the circuit described by the figure, bias or power amplifier). The pin identification
include positive and negative values. It is also points are established by adding one of the is important because, although common
common for different scales to be used in the three voltages, V1, V2 or V3 to the input signal. package pinouts are normally used, there
different quadrants, so inspect the legend for Bias voltage V1 results in an output current are exceptions. Manufacturers may differ
the curves carefully. of I1 when no input signal is present. This is slightly in the values reported, but certain
The parameters plotted in a characteristic shown as Bias Point 1 on the characteristic basic parameters are listed. Different batches
curve depend on how the device will be used curve. When an input signal is applied, the of the same devices are rarely identical, so
so that the applicable design values can be input voltage varies around V1 and the output manufacturers specify the guaranteed limits
obtained from the characteristic curve. The current varies around I1 as shown. If the dc for the parameters of their device. There are
slope of the curve is often important because
it relates changes in output to changes in in-
put. To determine the slope of the curve, two
closely-spaced points along that portion of
the curve are selected, each defined by its
location along the x and y axes. If the two
points are defined by (x1,y1) and (x2,y2), the
slope, m, of the curve (which can be a gain,
a resistance or a conductance, for example)
is calculated as:
kT 26 (3)
re
=
qIc Ie
It is important to pick points that are close
together or the slope will not reflect the
actual behavior of the device. A device whose
characteristic curve is not a straight line will
not have a linear response to inputs because
the slope changes with the value of the input
parameter.
For a device in which three parameters in-
teract, such as a transistor, sets of characteris-
tic curves can be drawn. Fig 3.5B shows a set
of characteristic curves for a bipolar transistor
where collector current, IC, is shown on the
y axis and collector-to-emitter voltage, VCE,
is shown on the x axis. Because the amount
of collector current also depends on base cur-
rent, IB, the curve is repeated several times for
different values of IB. From this set of curves,
an amplifier circuit using this transistor can Fig 3.6 Effect of biasing. An input signal may be reproduced linearly or nonlinearly
be designed to have specific values of gain. depending on the choice of bias points.

Analog Basics 3.7


usually three values listed in the data sheet for of 1022 electrons/cm3. In an insulator, nearly electrons donates free electrons to the crystal-
each parameter: guaranteed minimum value, all the electrons are tightly held by their atoms line structure; this is called an N-type impu-
the guaranteed maximum value, and/or the and the concentration of free electrons is very rity, for the negative charge of the majority
typical value. small on the order of 10 electrons/cm3. carriers. Some examples of donor impurities
Another section of the data sheet lists AB- Between the classes of materials consid- are antimony (Sb), phosphorus (P) and arsenic
SOLUTE MAXIMUM RATINGS, beyond which ered to be conductors and insulators is a class (As). N-type extrinsic semiconductors have
device damage may result. For example, the of elements called semiconductors, materials more electrons and fewer holes than intrin-
parameters listed in the ABSOLUTE MAXIMUM with conductivity much poorer than metals sic semiconductors. Acceptor impurities with
RATINGS section for a solid-state device are and much better than insulators. (In electron- three valence electrons accept free electrons
typically voltages, continuous currents, total ics, semiconductor means a device made from the lattice, adding holes to the overall
device power dissipation (PD) and operating- from semiconductor elements that have been structure. These are called P-type impurities,
and storage-temperature ranges. chemically manipulated as described below, for the positive charge of the majority carriers;
Rather than plotting the characteristic leading to interesting properties that create some examples are boron (B), gallium (Ga)
curves for each device, the manufacturer often useful applications.) and indium (In).
selects key operating parameters that describe Semiconductor atoms (silicon, Si, is the It is important to note that even though
the device operation for the configurations most widely used) share their valence elec- N-type and P-type material have different
and parameter ranges that are most commonly trons in a chemical bond that holds adjacent numbers of holes and free electrons than
used. For example, a bipolar transistor data atoms together, forming a three-dimensional intrinsic material, they are still electrically
sheet might include an OPERATING PARAM- lattice that gives the material its physical char- neutral. When an electron leaves an atom,
ETERS section. Parameters are listed in an
acteristics. A lattice of pure semiconductor the positively-charged atom that remains in
OFF CHARACTERISTICS subsection and an ON
material (one type of atom or molecule) can place in the crystal lattice electrically balances
CHARACTERISTICS subsection that describe
form a crystal, in which the lattice structure the roaming free electron. Similarly, an atom
the conduction properties of the device for and orientation is preserved throughout the gaining an electron acquires a negative charge
dc voltages. The SMALL-SIGNAL CHARACTERIS- material. Monocrystalline or single-crystal that balances the positively-charged atom it
is the type of material used in electronic semi- left. At no time does the material acquire a net
TICS section might contain a minimum Gain-
conductor devices. Polycrystalline material is electrical charge, positive or negative.
Bandwidth Product (fT or GBW), maximum
made of up many smaller crystals with their Compound semiconductor material can
output capacitance, maximum input capaci-
own individual lattice orientations. be formed by combining equal amounts of
tance, and the range of the transfer param-
Crystals of pure semiconductor material N-type and P-type impurity materials. Some
eters applicable to a given device. Finally, the
are called intrinsic semiconductors. When examples of this include gallium-arsenide
SWITCHING CHARACTERISTICS section might
energy, generally in the form of heat, is added (GaAs), gallium-phosphate (GaP) and in-
list absolute maximum ratings for Delay to a semiconductor crystal lattice, some elec- dium-phosphide (InP). To make an N-type
Time (td), Rise Time (tr), Storage Time (ts), trons are liberated from their bonds and move compound semiconductor, a slightly higher
and Fall Time (tf). Other types of devices list freely throughout the lattice. The bond that amount of N-type material is used in the mix-
characteristics important to operation of that loses an electron is then unbalanced and the ture. A P-type compound semiconductor has
specific device. space that the electron came from is referred a little more P-type material in the mixture.
When selecting equivalent parts for re- to as a hole. In these materials the number of Impurities are introduced into intrinsic
placement of specified devices, the data sheet free electrons is equal to the number of holes. semiconductors by diffusion, the same physi-
provides the necessary information to tell if Electrons from adjacent bonds can leave cal process that lets you smell cookies baking
a given part will perform the functions of their positions to fill the holes, thus leaving from several rooms away. (Molecules diffuse
another. Lists of cross-references and substi- behind a hole in their old location. As a conse- through air much faster than through solids.)
tution guides generally only specify devices quence of the electron moving, two opposite Rates of diffusion are proportional to tem-
that have nearly identical parameters. There movements can be said to occur: negatively perature, so semiconductors are doped with
are usually a large number of additional de- charged electrons move from bond to bond impurities at high temperature to save time.
vices that can be chosen as replacements. in one direction and positively charged holes Once the doped semiconductor material is
Knowledge of the circuit requirements adds move from bond to bond in the opposite di- cooled, the rate of diffusion of the impurities
even more to the list of possible replacements. rection. Both of these movements represent is so low that they are essentially immobile
The device parameters should be compared forms of electrical current, but this is very dif- for many years to come. If an electronic de-
individually to make sure that the replacement ferent from the current in a conductor. While vice made from a structure of N- and P-type
part meets or exceeds the parameter values of a conductor has free electrons that flow inde- materials is raised to a high temperature, such
the original part required by the circuit. Be pendently from the bonds of the crystalline as by excessive current, the impurities can
aware that in some applications a far superior lattice, the current in a pure semiconductor again migrate and the internal structure of
part may fail as a replacement, however. A is constrained to move from bond to bond. the device may be destroyed. The maximum
transistor with too much gain could easily Impurities can be added to intrinsic semi- operating temperature for semiconductor de-
oscillate if there were insufficient negative conductors (by a process called doping) to vices is specified at a level low enough to limit
feedback to ensure stability. enhance the formation of electrons or holes additional impurity diffusion.
and thus improve conductivity. These mate- The conductivity of an extrinsic semicon-
rials are extrinsic semiconductors. Since the ductor depends on the charge density (in other
3.2.5 Physical Electronics of additional electrons and holes can move, their words, the concentration of free electrons in
Semiconductors movement is current and they are called car- N-type, and holes in P-type, semiconductor
In a conductor, such as a metal, some of the riers. The type of carrier that predominates material). As the energy in the semiconductor
outer, or valence, electrons of each atom are in the material is called the majority carrier. increases, the charge density also increases.
free to move about between atoms. These free In N-type material the majority carriers are This is the basis of how all semiconductor
electrons are the constituents of electrical cur- electrons and in P-type material, holes. devices operate: the major difference is the
rent. In a good conductor, the concentration of There are two types of impurities that can way in which the energy level is increased.
these free electrons is very high, on the order be added: a donor impurity with five valence Variations include: The transistor, where con-

3.8 Chapter 3
ductivity is altered by injecting current into semiconductor diode to act as rectifier.
the device via a wire; the thermistor, where
the level of heat in the device is detected by its
conductivity, and the photoconductor, where 3.2.7 Junction
light energy that is absorbed by the semicon- Semiconductors
ductor material increases the conductivity. Semiconductor devices that operate using
the principles of a PN junction are called junc-
3.2.6 The PN Semiconductor tion semiconductors. These devices can have
one or several junctions. The properties of
Junction junction semiconductors can be tightly con-
If a piece of N-type semiconductor material trolled by the characteristics of the materials
is placed against a piece of P-type semicon- used and the size and shape of the junctions.
ductor material, the location at which they
join is called a PN junction. The junction SEMICONDUCTOR DIODES
has characteristics that make it possible to Diodes are commonly made of silicon
develop diodes and transistors. The action and occasionally germanium. Although they
of the junction is best described by a diode act similarly, they have slightly different
operating as a rectifier. characteristics. The junction threshold
Initially, when the two types of semicon- voltage, or junction barrier voltage, is the
ductor material are placed in contact, each type forward bias voltage (VF) at which current
of material will have only its majority carriers: begins to pass through the device. This volt-
P-type will have only holes and N-type will age is different for the two kinds of diodes. In
have only free electrons. The presence of the the diode response curve of Fig 3.7, VF cor-
positive charges (holes) in the P-type material responds to the voltage at which the positive
attracts free electrons from the N-type material portion of the curve begins to rise sharply from
immediately across the junction. The opposite the x-axis. Most silicon diodes have a junc-
is true in the N-type material. tion threshold voltage of about 0.7 V, while
These attractions lead to diffusion of some the voltage for germanium diodes is typically
of the majority carriers across the junction, 0.3 V. Reverse leakage current is much lower
which combine with and neutralize the major- for silicon diodes than for germanium diodes.
Fig 3.7 Semiconductor diode (PN
ity carriers immediately on the other side (a The characteristic curve for a semicon- junction) characteristic curve. (A)
process called recombination). As distance ductor diode junction is given by the fol- Forward- biased (anode voltage higher
from the junction increases, the attraction lowing equation (slightly simplified) called than cathode) response for Germanium
quickly becomes too small to cause the car- the Fundamental Diode Equation because it (Ge) and Silicon (Si) devices. Each curve
riers to move. The region close to the junction describes the behavior of all semiconductor breaks away from the x-axis at its junction
threshold voltage. The slope of each curve
is then depleted of carriers, and so is named PN junctions. is its forward resistance. (B) Reverse-
the depletion region (also the space-charge biased response. Very small reverse
region or the transition region). The width V current increases until it reaches the
the I IS e t 1
V
of the depletion region is very small, on= (4) reverse saturation current (I0). The reverse

order of 0.5 m. current increases suddenly and drastically
If the N-type material (the cathode) is where
when the reverse voltage reaches the
reverse breakdown voltage, VBR.
placed at a more negative voltage than the I = diode current
P-type material (the anode), current will V = diode voltage
pass through the junction because electrons Is = reverse-bias saturation current
are attracted from the lower potential to the The reverse current flow rapidly reaches a
Vt = kT/q, the thermal equivalent of voltage
higher potential and holes are attracted in level that varies little with the reverse bias
(about 25 mV at room temperature)
the opposite direction. This forward bias voltage. This is the reverse-bias saturation
= emission coefficient.
forces the majority carriers toward the junc- current, Is.
tion where recombination occurs with the The value of Is varies with the type of semi For bias (dc) circuit calculations, a useful
opposite type of majority carrier. The source conductor material, with the value of 10-12 model for the diode that takes these two ef-
of voltage supplies replacement electrons to used for silicon. also varies from 1 to 2 with fects into account is shown by the artificial
the N-type material and removes electrons the type of material and method of fabrication. I-V curve in Fig 3.8C. This model neglects
from the P-type material so that the majority ( is close to 1 for silicon at normal current the negligible reverse bias current Is.
carriers are continually replenished. Thus, values, increasing to 2 at high currents.) This When converted into an equivalent circuit,
the net effect is a forward current flowing curve is shown in Fig 3.8B. the model in Fig 3.8C yields the circuit in Fig
through the semiconductor, across the PN The obvious differences between Fig 3.8A 3.8D. The ideal voltage source Va represents
junction. The forward resistance of a diode and B are that the semiconductor diode has a the turn-on voltage and Rf represents the ef-
conducting current is typically very low and finite turn-on voltage it requires a small but fective resistance caused by the small increase
varies with the amount of forward current. nonzero forward bias voltage before it begins in diode voltage as the diode current increases.
When the polarity is reversed, majority conducting. Furthermore, once conducting, The turn-on voltage is material-dependent:
carriers are attracted away from the junc- the diode voltage continues to increase very approximately 0.3 V for germanium diodes
tion, not toward it. Very little current flows slowly with increasing current, unlike a true and 0.7 for silicon. Rf is typically on the or-
across the PN junction called reverse leak- short circuit. Finally, when the applied volt- der of 10 , but it can vary according to the
age current in this case. Allowing only age is negative, the reverse current is not specific component. Rf can often be com-
unidirectional current flow is what allows a exactly zero but very small (microamperes). pletely neglected in comparison to the other

Analog Basics 3.9


Fig 3.9 Bipolar transistors. (A)
A layer of N-type semiconductor
sandwiched between two layers of P-type
semiconductor makes a PNP device.
The schematic symbol has three leads:
collector (C), base (B) and emitter (E),
with the arrow pointing in toward the
base. (B) A layer of P-type semiconductor
sandwiched between two layers of N-type
semiconductor makes an NPN device.
The schematic symbol has three leads:
collector (C), base (B) and emitter (E), with
the arrow pointing out away from the base.
Figure 3.8 Circuit models for rectifying switches (diodes). A: I-V curve of the ideal
rectifier. B: I-V curve of a typical semiconductor diode showing the typical small
leakage current in the reverse direction. Note the different scales for forward and
reverse current. C shows a simplified diode I-V curve for dc-circuit calculations (at a
much larger scale than B). D is an equivalent circuit for C. saturation. Conversely, when base-emitter
current is reduced to the point at which col-
lector current ceases to flow, that is the situ-
ation of cutoff.

resistances in the circuit. This very common come minority carriers because the materials THYRISTORS
simplification leaves only a pure voltage drop of the emitter and base regions have opposite Thyristors are semiconductors made with
for the diode model. polarity. The excess minority carriers in the four or more alternating layers of P- and N-type
base are then attracted across the very thin semiconductor material. In a four-layer thyris-
BIPOLAR TRANSISTOR base to the base-collector junction, where they tor, when the anode is at a higher potential than
A bipolar transistor is formed when two are collected and are once again considered the cathode, the first and third junctions are
PN junctions are placed next to each other. majority carriers before they can flow to the forward biased and the center junction reverse
If N-type material is surrounded by P-type base terminal. biased. In this state, there is little current, just
material, the result is a PNP transistor. Al- The flow of majority carriers from emitter as in the reverse-biased diode. The different
ternatively, if P-type material is in the middle to collector can be modified by the applica- types of thyristor have different ways in which
of two layers of N-type material, the NPN tion of a bias current to the base terminal. they turn on to conduct current and in how they
transistor is formed (Fig 3.9). If the bias current causes majority carriers turn off to interrupt current flow.
Physically, we can think of the transistor as to be injected into the base material (elec-
two PN junctions back-to-back, such as two trons flowing into an N-type base or out of PNPN Diode
diodes connected at their anodes (the positive a P-type base) the emitter-collector current The simplest thyristor is a PNPN (usually
terminal) for an NPN transistor or two diodes increases. In this way, a transistor allows a pronounced like pinpin) diode with three
connected at their cathodes (the negative ter- small base current to control a much larger junctions (see Fig 3.10). As the forward bias
minal) for a PNP transistor. The connection collector current. voltage is increased, the current through the
point is the base of the transistor. (You cant As in a semiconductor diode, the forward device increases slowly until the breakover
actually make a transistor this way this is biased base-emitter junction has a threshold (or firing) voltage, VBO, is reached and the
a representation for illustration only.) voltage (VBE) that must be exceeded before flow of current abruptly increases. The PNPN
A transistor conducts when the base- the emitter current increases. As the base- diode is often considered to be a switch that
emitter junction is forward biased and the emitter current continues to increase, the point is off below VBO and on above it.
base-collector is reverse biased. Under these is reached at which further increases in base-
conditions, the emitter region emits majority emitter current cause no additional change Bilateral Diode Switch (Diac)
carriers into the base region, where they be- in collector current. This is the condition of A semiconductor device similar to two

3.10 Chapter 3
Fig 3.10 PNPN diode. (A) Alternating Fig 3.11 Bilateral switch. (A) Fig 3.12 SCR. (A) Alternating layers
layers of P-type and N-type semiconduc- Alternating layers of P-type and N-type of P-type and N-type semiconductor.
tor. (B) Schematic symbol with cathode semiconductor. (B) Schematic symbol. This is similar to a PNPN diode with gate
(C) and anode (A) leads. (C) I-V curve. (C) I-V curve. The right-hand side of the terminals attached to the interior layers.
Reverse-biased response is the same curve is identical to the PNPN diode (B) Schematic symbol with anode (A),
as normal PN junction diodes. Forward response in Fig 3.10. The device responds cathode (C), anode gate (GA) and cathode
biased response acts as a hysteresis identically for both forward and reverse gate (GC). Many devices are constructed
switch. Resistance is very high until the bias so the left-hand side of the curve is without GA. (C) I-V curve with different
bias voltage reaches VBO (where the center symmetrical with the right-hand side. responses for various gate currents.
junction breaks over) and exceeds the IG = 0 has a similar response to the
cutoff current, IBO. The device exhibits PNPN diode.
a negative resistance when the current
increases as the bias voltage decreases
until a voltage of VH and saturation current
of IH is reached. After this, the resistance
is very low, with large increases in current
for small voltage increases.
the anode is called the anode gate. In nearly 100 A and voltage differentials of greater than
all commercially available SCRs, only the 1000 V, yet can be switched with gate currents
cathode gate is connected (Fig 3.12). of less than 50 mA. Because of their high
Like the PNPN diode switch, the SCR is current-handling capability, SCRs are used as
used to abruptly start conducting when the crowbars in power supply circuits, to short
PNPN diodes facing in opposite directions voltage exceeds a given level. By biasing the the output to ground and blow a fuse when an
and attached in parallel is the bilateral diode gate terminal appropriately, the breakover overvoltage condition exists.
switch or diac. This device has the characteris- voltage can be adjusted. SCRs and triacs are often used to control
tic curve of the PNPN diode for both positive ac power sources. A sine wave with a given
and negative bias voltages. Its construction, Triac RMS value can be switched on and off at
schematic symbol and characteristic curve A five-layered semiconductor whose op- preset points during the cycle to decrease the
are shown in Fig 3.11. eration is similar to a bidirectional SCR is RMS voltage. When conduction is delayed
the triac (Fig 3.13). This is also similar to a until after the peak (as Fig 3.14 shows) the
Silicon Controlled Rectifier (SCR) bidirectional diode switch with a bias control peak-to-peak voltage is reduced. If conduc-
Another device with four alternate layers of gate. The gate terminal of the triac can control tion starts before the peak, the RMS voltage
P-type and N-type semiconductor is the sili- both positive and negative breakover voltages is reduced, but the peak-to-peak value remains
con controlled rectifier (SCR). (Some sources and the devices can pass both polarities of the same. This method is used to operate light
refer to an SCR as a thyristor, as well.) In voltage. dimmers and 240 V ac to 120 V ac converters.
addition to the connections to the outer two The sharp switching transients created when
layers, two other terminals can be brought out Thyristor Applications these devices turn on are common sources
for the inner two layers. The connection to the The SCR is highly efficient and is used in of RF interference. (See the chapter on RF
P-type material near the cathode is called the power control applications. SCRs are avail- Interference for information on dealing with
cathode gate and the N-type material near able that can handle currents of greater than interference from thyristors.)

Analog Basics 3.11


ing (rDS(ON)) may be anywhere from a few
hundred ohms to much less than an ohm. The
output impedance of devices made with FETs
is generally quite low. If a gate bias voltage is
added to operate the transistor near cutoff, the
circuit output impedance may be much higher.
In order to achieve a higher gain-bandwidth
product, other materials have been used. Gal-
lium-arsenide (GaAs) has electron mobility
and drift velocity (both are measures of how
easily electrons are able to move through the
crystal lattice) far higher than the standard
doped silicon. Amplifiers designed with
GaAsFET devices operate at much higher
frequencies and with a lower noise factor at
Fig 3.13 Triac. (A) Alternating layers VHF and UHF than those made with silicon
of P-type and N-type semiconductor. FETs (although silicon FETs have improved
This behaves as two SCR devices facing Fig 3.15 JFET devices with terminals dramatically in recent years).
in opposite directions with the anode labeled: source (S), gate (G) and drain (D).
of one connected to the cathode of the A) Pictorial of N-type channel embedded JFET
other and the cathode gates connected in P-type substrate and schematic
together. (B) Schematic symbol. symbol. B) P-channel embedded in N-type One of two basic types of FET, the junc-
substrate and schematic symbol. tion FET (JFET) gate material is made of
the opposite polarity semiconductor to the
channel material (for a P-channel FET the
gate is made of N-type semiconductor mate-
rial). The gate-channel junction is similar to
opposite polarity (a P-channel FET has N- a diodes PN junction with the gate material
type substrate). Most FETs are constructed in direct contact with the channel. JFETs are
with silicon. used with the junction reverse-biased, since
Within the FET, current moves in a channel any current in the gate is undesirable. The
as shown in Fig 3.15. The channel is made of reverse bias of the junction creates an electric
either N-type or P-type semiconductor mate- field that pinches the channel. Since the
rial; an FET is specified as either an N-channel magnitude of the electric field is proportional
Fig 3.14 Triac operation on sine wave. or P-channel device. Current flows from the to the reverse-bias voltage, the current in the
The dashed line is the original sine wave source terminal (where majority carriers are channel is reduced for higher reverse gate
and the solid line is the portion that
conducts through the triac. The relative injected) to the drain terminal (where major- bias voltages. When current in the channel is
delay and conduction period times are ity carriers are removed). A gate terminal completely halted by the electric field, this is
controlled by the amount or timing of gate generates an electric field that controls the called pinch-off and it is analogous to cutoff
current, IG. The response of an SCR is the current in the channel. in a bipolar transistor. The channel in a JFET
same as this for positive voltages (above In N-channel devices, the drain potential is at its maximum conductivity when the gate
the x-axis) and with no conduction for
negative voltages.
must be higher than that of the source (VDS > and source voltages are equal (VGS = 0).
0) for electrons (the majority carriers) to flow Because the gate-channel junction in a
in channel. In P-channel devices, the flow of JFET is similar to a bipolar junction diode,
holes requires that VDS < 0. The polarity of the this junction must never be forward biased;
3.2.8 Field-Effect Transistors electric field that controls current in the chan- otherwise large currents will pass through the
nel is determined by the majority carriers of gate and into the channel. For an N-channel
(FET) the channel, ordinarily positive for P-channel JFET, the gate must always be at a lower
The field-effect transistor (FET) controls FETs and negative for N-channel FETs. potential than the source (VGS < 0). The pro-
the current between two points but does so Variations of FET technology are based hibited condition is for VGS > 0. For P-channel
differently than the bipolar transistor. The on different ways of generating the electric JFETs these conditions are reversed (in nor-
FET operates by the effects of an electric field field. In all of these, however, electrons at the mal operation VGS > 0 and the prohibited
on the flow of electrons through a single type gate are used only for their charge in order condition is for VGS < 0).
of semiconductor material. This is why the to create an electric field around the channel.
FET is sometimes called a unipolar transistor. There is a minimal flow of electrons through MOSFET
Unlike bipolar semiconductors that can be the gate. This leads to a very high dc input Placing an insulating layer between the
arranged in many configurations to provide resistance in devices that use FETs for their gate and the channel allows for a wider
diodes, transistors, photoelectric devices, input circuitry. There may be quite a bit of range of control (gate) voltages and further
temperature sensitive devices and so on, the capacitance between the gate and the other decreases the gate current (and thus increases
field effect technique is usually only used FET terminals, however, causing the input the device input resistance). The insulator is
to make transistors, although FETs are also impedance to be quite low at high frequencies. typically made of an oxide (such as silicon
available as special-purpose diodes, for use The current through an FET only has to dioxide, SiO2). This type of device is called a
as constant current sources. pass through a single type of semiconductor metal-oxide-semiconductor FET (MOSFET)
FET devices are constructed on a substrate material. Depending on the type of mate- or insulated-gate FET (IGFET).
of doped semiconductor material. The chan- rial and the construction of the FET, drain- The substrate is often connected to the
nel is formed within the substrate and has the source resistance when the FET is conduct- source internally. The insulated gate is on the

3.12 Chapter 3
opposite side of the channel from the substrate
(see Fig 3.16). The bias voltage on the gate
terminal either attracts or repels the majority
carriers of the substrate across its PN-junction
with the channel. This narrows (depletes) or
widens (enhances) the channel, respectively,
as VGS changes polarity. For example, in the
N-channel enhancement-mode MOSFET,
positive gate voltages with respect to the sub-
strate and the source (VGS > 0) repel holes
from the channel into the substrate, thereby
widening the channel and decreasing channel
resistance. Conversely, VGS < 0 causes holes
to be attracted from the substrate, narrowing
the channel and increasing the channel resis-
tance. Once again, the polarities discussed
in this example are reversed for P-channel
devices. The common abbreviation for an
N-channel MOSFET is NMOS, and for a
P-channel MOSFET, PMOS.
Because of the insulating layer next to
the gate, input resistance of a MOSFET is
usually greater than 1012 (a million meg-
ohms). Since MOSFETs can both deplete the
channel, like the JFET, and also enhance it,
the construction of MOSFET devices differs
based on the channel size in the quiescent
state, VGS = 0.
A depletion mode device (also called a
normally-on MOSFET) has a channel in the Fig 3.16 MOSFET devices with terminals labeled: source (S), gate (G) and drain
quiescent state that gets smaller as a reverse (D). N-channel devices are pictured. P-channel devices have the arrows reversed in
the schematic symbols and the opposite type semiconductor material for each of the
bias is applied; this device conducts current layers. (A) N-channel depletion mode device schematic symbol and (B) pictorial of
with no bias applied (see Fig 3.16A and B). P-type substrate, diffused N-type channel, SiO2 insulating layer and aluminum gate
An enhancement mode device (also called region and source and drain connections. The substrate is connected to the source
a normally-off MOSFET) is built without a internally. A negative gate potential narrows the channel. (C) N-channel enhancement
channel and does not conduct current when mode device schematic and (D) pictorial of P-type substrate, N-type source and
VGS = 0; increasing forward bias forms a drain wells, SiO2 insulating layer and aluminum gate region and source and drain
connections. Positive gate potential forms a channel between the two N-type wells by
temporary channel that conducts current (see repelling the P-carriers away from the channel region in the substrate.
Fig 3.16C and D).

Complementary Metal Oxide


Semiconductors (CMOS)
Power dissipation in a circuit can be re-
duced to very small levels (on the order of a
few nanowatts) by using MOSFET devices in
complementary pairs (CMOS). Each ampli-
fier is constructed of a series circuit of MOS-
FET devices, as in Fig 3.17. The gates are tied
together for the input signal, as are the drains
for the output signal. In saturation and cutoff,
only one of the devices conducts. The current
drawn by the circuit under no load is equal
to the off leakage current of either device
and the voltage drop across the pair is equal
to VDD, so the steady-state power used by
the circuit is always equal to VDD ID(OFF).
Power is only consumed during the switching
process, so for ac signals, power consumption
is proportional to frequency.
CMOS circuitry could be built with discrete
components, but the number of extra parts and Fig 3.17 Complementary metal oxide semiconductor (CMOS). (A) CMOS device is
the need for the complementary components made from a pair of enhancement mode MOS transistors. The upper is an N-channel
to be matched has made that an unusual design device, and the lower is a P-channel device. When one transistor is biased on, the other
technique. The low power consumption and is biased off; therefore, there is minimal current from VDD to ground. (B) Implementation
ease of fabrication has made CMOS the most of a CMOS pair as an integrated circuit.

Analog Basics 3.13


common of all IC technologies. Although for semiconductors that are related to heat. circuits analog and digital dissipating
CMOS is most commonly used in digital in- The semiconductor material is connected to significant amounts of heat under normal op-
tegrated circuitry, its low power consumption the outside world through metallic bonding erating conditions. Computer microprocessors
has also been put to work by manufacturers leads. The point at which the lead and the are a good example, often requiring their own
of analog ICs, as well as digital ICs. semiconductor are connected is a common cooling systems. Once the device cools, nor-
place for the semiconductor device to fail. mal operation is usually restored.
As the device heats up and cools down, the To reduce the risk of thermal failures, the
3.2.9 Semiconductor materials expand and contract. The rate of designer must comply with the limits stated
Temperature Effects expansion and contraction of semiconductor in the manufacturers data sheet, devising an
The number of excess holes and electrons material is different from that of metal. Over adequate heat removal system. (Thermal is-
in semiconductor material is increased as the many cycles of heating and cooling the bond sues are discussed in the Electrical Funda-
temperature of a semiconductor increases. between the semiconductor and the metal mentals chapter.)
Since the conductivity of a semiconductor can break. Some experts have suggested that
is related to the number of excess carriers, the lifetime of semiconductor equipment can 3.2.10 Safe Operating Area
this also increases with temperature. With be extended by leaving the devices powered
(SOA)
respect to resistance, semiconductors have a on all the time, but this requires removal of
the heat generated during normal operation. Devices intended for use in circuits han-
negative temperature coefficient. The resis-
tance of silicon decreases by about 8% per A common failure mode of semiconduc- dling high currents or voltages are speci-
C and by about 6% per C for germanium. tors is caused by the heat generated during fied to have a safe operating area (SOA).
Semiconductor temperature properties are the semiconductor use. If the temperatures of the This refers to the area drawn on the devices
opposite of most metals, which increase their PN junctions remain at high enough levels for characteristic curve containing combina-
resistance by about 0.4% per C. These op- long enough periods of time, the impurities tions of voltage and current that the device
posing temperature characteristics permit the resume their diffusion across the PN junc- can be expected to control without dam-
design of circuits with opposite temperature tions. When enough of the impurity atoms age under specific conditions. The SOA
coefficients that cancel each other out, making cross the depletion region, majority carrier combines a number of limits voltage,
a temperature insensitive circuit. recombination stops functioning properly and current, power, temperature and various break-
Semiconductor devices can experience an the semiconductor device fails permanently. down mechanisms in order to simplify
effect called thermal runaway as the current Excessive temperature can also cause fail- the design of protective circuitry. The SOA
causes an increase in temperature. (This is ure anywhere in the semiconductor from heat is also specified to apply to specific dura-
primarily an issue with bipolar transistors.) generation within any current-carrying con- tions of use steady-state, long pulses, short
The increased temperature decreases resis- ductor, such as an FET channel or the bonding pulses and so forth. The device may have
tance and may lead to a further increase in leads. Integrated circuits with more than one separate SOAs for resistive and inductive
current (depending on the circuit) that leads output may have power dissipation limits that loads.
to an additional temperature increase. This depend on how many of the outputs are active You may also encounter two specialized
sequence of events can continue until the at one time. The high temperature can cause types of SOA for turning the device on and
semiconductor destroys itself, so circuit de- localized melting or cracking of the semicon- off. Reverse bias safe operating area (RB-
sign must include measures that compensate ductor material, causing a permanent failure. SOA) applies when the device is turning off.
for the effects of temperature. Another heat-driven failure mode, usually Forward bias safe operating area (FBSOA)
not fatal to the semiconductor, is excessive applies when turning the device on. These
Semiconductor Failure leakage current or a shift in operating point that SOAs are used because the high rate-of-
Caused by Heat causes the circuit to operate improperly. This change of current and voltage places addi-
There are several common failure modes is a particular problem in complex integrated tional stresses on the semiconductor.

3.3 Practical Semiconductors


3.3.1 Semiconductor Diodes material, as indicated in Fig 3.18. Most diodes in the reverse direction. This is not the case
Although many types of semiconductor di- are marked with a band on the cathode end for actual devices, which behave as shown in
odes are available, they share many common (Fig 3.18). the plot of a diode response in Fig 3.7. Note
characteristics. The different types of diodes that the scales of the two graphs are drastically
DIODE RATINGS different. The inverse of the slope of the line
have been developed to optimize particular
characteristics for one type of application. Five major characteristics distinguish (the change in voltage between two points
You will find many examples of diode ap- standard junction diodes from one another: on a straight portion of the line divided by
plications throughout this book. current handling capacity, maximum volt- the corresponding change in current) on the
The diode symbol is shown in Fig 3.18. age rating, response speed, reverse leakage upper right is the resistance of the diode in
Forward current flows in the direction from current and junction forward voltage. Each the forward direction, RF.
anode to cathode, in the direction of the ar- of these characteristics can be manipulated The range of voltages is small and the range
row. Reverse current flows from cathode to during manufacture to produce special pur- of currents is large since the forward resis-
anode. (Current is considered to be conven- pose diodes. tance is very small (in this example, about
tional current as described in the Electrical 2 ). Nevertheless, this resistance causes heat
Current Capacity dissipation according to P = IF2 RF.
Fundamentals chapter.) The anode of a semi-
conductor junction diode is made of P-type The ideal diode would have zero resistance In addition, there is a forward voltage, VF,
material and the cathode is made of N-type in the forward direction and infinite resistance whenever the forward current is flowing. This

3.14 Chapter 3
may not be able to turn current on and off as While component derating does reduce
fast as the changing polarity of the signal. self-heating effects, circuits must be designed
Diode response speed mainly depends for the expected operating environment. For
on charge storage in the depletion region. example, mobile radios may face tempera-
When forward current is flowing, electrons tures from 20 to +140 F (29 to 60 C).
and holes fill the region near the junction
to recombine. When the applied voltage re- Forward Voltage
verses, these excess charges move away from The amount of voltage required to cause
the junction so that no recombination can majority carriers to enter the depletion region
take place. As reverse bias empties the deple- and recombine, creating full current flow, is
tion region of excess charge, it begins to act called a diodes forward voltage, VF. It de-
like a small capacitor formed by the regions pends on the type of material used to create
containing majority carriers on either side of the junction and the amount of current. For
Fig 3.18 Practical semiconductor the junction and the depletion region acting silicon diodes at normal currents, VF = 0.7 V,
diodes. All devices are aligned with as the dielectric. This junction capacitance is and for germanium diodes, VF = 0.3 V. As you
anode on the left and cathode on the inversely proportional to the width of the de- saw earlier, VF also affects power dissipation
right. (A) Standard PN junction diode. (B) pletion region and directly proportional to the in the diode.
Point-contact or cats whisker diode.
(C) PIN diode formed with heavily doped cross-sectional surface area of the junction.
P-type (P+), undoped (intrinsic) and The effect of junction capacitance is to POINT-CONTACT DIODES
heavily doped N-type (N+) semiconductor allow current to flow for a short period after One way to decrease charge storage time
material. (D) Diode schematic symbol. (E) the applied voltage changes from positive to in the depletion region is to form a metal-
Diode package with marking stripe on the negative. To halt current flow requires that the semiconductor junction for which the deple-
cathode end.
junction capacitance be charged. Charging tion is very thin. This can be accomplished
this capacitance takes some time; a few s with a point-contact diode, where a thin piece
for regular rectifier diodes and a few hundred of aluminum wire, often called a whisker, is
nanoseconds for fast-recovery diodes. This is placed in contact with one face of a piece of
also results in heat dissipation as P = I VF. the diodes charge-storage time. The amount lightly doped N-type material. In fact, the
In power applications where the average for- of time required for current flow to cease is original diodes used for detecting radio sig-
ward current is high, heating from forward the diodes recovery time. nals (cats whisker diodes) were made with
resistance and the forward voltage drop can be a steel wire in contact with a crystal of impure
significant. Since forward current determines Reverse Leakage Current lead (galena). Point-contact diodes have high
the amount of heat dissipation, the diodes Because the depletion region is very thin, response speed, but poor PIV and current-
power rating is stated as a maximum aver- reverse bias causes a small amount of re- handling ratings. The 1N34 germanium point-
age current. Exceeding the current rating in verse leakage or reverse saturation current to contact diode is the best-known example of
a diode will cause excessive heating that leads flow from cathode to anode. This is typically point-contact diode still in common use.
to PN junction failure as described earlier. 1 A or less until reverse breakdown voltage
is reached. Silicon diodes have lower reverse SCHOTTKY DIODES
Peak Inverse Voltage (PIV) leakage currents than diodes made from other An improvement to point-contact diodes,
In Fig 3.7, the lower left portion of the materials with higher carrier mobility, such as the hot-carrier diode is similar to a point-
curve illustrates a much higher resistance that germanium. contact diode, but with more ideal character-
increases from tens of kilohms to thousands The reverse saturation current Is is not istics attained by using more efficient metals,
of megohms as the reverse voltage gets larg- constant but is affected by temperature, with such as platinum and gold, that act to lower
er, and then decreases to near zero (a nearly higher temperatures increasing the mobil- forward resistance and increase PIV. This type
vertical line) very suddenly. This sudden ity of the majority carriers so that more of of contact is known as a Schottky barrier,
change occurs because the diode enters re- them cross the depletion region for a given and diodes made this way are called Schottky
verse breakdown or when the reverse voltage amount of reverse bias. For silicon diodes diodes. The junctions of Schottky diodes, be-
becomes high enough to push current across (and transistors) near room temperature, Is ing smaller, store less charge and as a result,
the junction. The voltage at which this occurs increases by a factor of 2 every 4.8 C. This have shorter switching times and junction
is the reverse breakdown voltage. Unless the means that for every 4.8 C rise in tempera- capacitances than standard PN-junction di-
current is so large that the diode fails from ture, either the current doubles (if the voltage odes. Their forward voltage is also lower,
overheating, breakdown is not destructive and across it is constant), or if the current is held typically 0.3 to 0.4 V. In most other respects
the diode will again behave normally when the constant by other resistances in the circuit, they behave similarly to PN diodes.
bias is removed. The maximum reverse volt- the diode voltage will decrease by VT ln 2
age that the diode can withstand under normal = 18 mV. For germanium, the current doubles PIN DIODES
use is the peak inverse voltage (PIV) rating. every 8 C and for gallium-arsenide (GaAs), The PIN diode, shown in Fig 3.18C is a
A related effect is avalanche breakdown in 3.7 C. This dependence is highly reproduc- slow response diode that is capable of pass-
which the voltage across a device is greater ible and may actually be exploited to produce ing RF and microwave signals when it is for-
than its ability to control or block current flow. temperature-measuring circuits. ward biased. This device is constructed with
While the change resulting from a rise of a layer of intrinsic (undoped) semiconductor
Response Speed several degrees may be tolerable in a circuit placed between very highly doped P-type and
The speed of a diodes response to a change design, that from 20 or 30 degrees may not. N-type material (called P+-type and N+-type
in voltage polarity limits the frequency of ac Therefore its a good idea with diodes, just material to indicate the extra amount of dop-
current that the diode can rectify. The diode as with other components, to specify power ing), creating a PIN junction. These devices
response in Fig 3.7 shows how that diode will ratings conservatively (2 to 4 times margin) provide very effective switches for RF signals
act at dc. As the frequency increases, the diode to prevent self-heating. and are often used in transmit-receive switches

Analog Basics 3.15


Fig 3.19
Varactor diode. (A)
Schematic symbol.
(B) Equivalent
circuit of the
reverse biased
varactor diode.
RS is the junction
resistance, RJ
is the leakage
resistance and Fig 3.20 Zener diode. (A) Schematic
CJ is the junction symbol. (B) Basic voltage regulating
capacitance, which circuit. VZ is the Zener reverse breakdown
is a function of voltage. Above VZ, the diode draws cur-
the magnitude rent until VI IIR = VZ. The circuit design
of the reverse should select R so that when the maxi-
bias voltage. (C) mum current is drawn, R < (VI VZ) / IO.
Plot of junction The diode should be capable of passing
capacitance, CJ, the same current when there is no output
as a function of current drawn.
reverse voltage,
VR, for three
different varactor
devices. Both axes
are plotted on a as it does when it is forward biased. This cur-
logarithmic scale. rent will not destroy the diode if it is limited
to less than the devices maximum allow-
able value. By using heavy levels of doping
during manufacture, a diodes PIV can be
precisely controlled to be at a specific level,
called the Zener voltage, creating a type of
voltage reference. These diodes are called
Zener diodes after their inventor, American
in transceivers and amplifiers. The majority VP I DQ
=RS 1


physicist Clarence Zener.
carriers in PIN diodes have longer than nor- I DQ I DSS (5)
When the Zener voltage is reached, the
mal lifetimes before recombination, resulting
reverse voltage across the Zener diode re-
in a slow switching process that causes them where Cj0 = measured capacitance with zero mains constant even as the current through it
to act more like resistors than diodes at high applied voltage. changes. With an appropriate series current-
radio frequencies. The amount of resistance Note that the quantity under the radical is limiting resistor, the Zener diode provides
can be controlled by the amount of forward a large positive quantity for reverse bias. As an accurate voltage reference (see Fig 3.20).
bias applied to the PIN diode and this allows seen from the equation, for large reverse bi- Zener diodes are rated by their reverse-
them to act as current-controlled attenuators. ases Cj is inversely proportional to the square breakdown voltage and their power-handling
(For additional discussion of PIN diodes and root of the voltage. capacity, where P = VZ IZ. Since the same
projects in which they are used, see the chap- Although special forms of varactors are current must always pass though the resistor to
ters on Transmitters and Transceivers, RF available from manufacturers, other types of drop the source voltage down to the reference
Power Amplifiers, and Test Equipment and diodes may be used as inexpensive varactor voltage, with that current divided between the
Measurements.) diodes, but the relationship between reverse Zener diode and the load, this type of power
voltage and capacitance is not always reliable. source is very wasteful of current.
VARACTOR DIODES When designing with varactor diodes, the The Zener diode does make an excellent
Junction capacitance can be used as a circuit reverse bias voltage must be absolutely free of and efficient voltage reference in a larger volt-
element by controlling the reverse bias voltage noise since any variations in the bias voltage age regulating circuit where the load current
across the junction, creating a small variable will cause changes in capacitance. For ex- is provided from another device whose volt-
capacitor. Junction capacitances are small, on ample, if the varactor is used to tune an oscil- age is set by the reference. (See the Power
the order of pF. As the reverse bias voltage on lator, unwanted frequency shifts or instability Sources chapter for more information about
a diode increases, the width of the depletion will result if the reverse bias voltage is noisy. using Zener diodes as voltage regulators.)
region increases, decreasing its capacitance. It is possible to frequency modulate a signal When operating in the breakdown region,
A varactor (also known by the trade name by adding the audio signal to the reverse bias Zener diodes can be modeled as a simple
Varicap diode) is a diode with a junction spe- on a varactor diode used in the carrier oscil- voltage source.
cially formulated to have a relatively large lator. (For examples of the use of varactors in The primary sources of error in Zener-
range of capacitance values for a modest range oscillators and modulators, see the chapters diode-derived voltages are the variation with
of reverse bias voltages (Fig 3.19). on Mixers, Modulators, and Demodulators load current and the variation due to heat.
As the reverse bias applied to a diode and Oscillators and Synthesizers.) Temperature-compensated Zener diodes are
changes, the width of the depletion layer, available with temperature coefficients as low
and therefore the capacitance, also changes. ZENER DIODES as 5 parts per million per C. If this is unac-
The diode junction capacitance (Cj) under a When the PIV of a reverse-biased diode is ceptable, voltage reference integrated circuits
reverse bias of V volts is given by exceeded, the diode begins to conduct current based on Zener diodes have been developed

3.16 Chapter 3
that include additional circuitry to counteract however, the relative sizes of the collector,
temperature effects. base and emitter regions differ. A common
A variation of Zener diodes, transient volt- transistor configuration that spans a distance
age suppressor (TVS) diodes are designed to of 3 mm between the collector and emitter
dissipate the energy in short-duration, high- contacts typically has a base region that is
voltage transients that would otherwise dam- only 25 m across.
age equipment or circuits. TVS diodes have The operation of the bipolar transistor is
large junction cross-sections so that they can described graphically by characteristic curves
handle large currents without damage. These as shown in Fig 3.22. These are similar to the
diodes are also known as TransZorbs. Since I-V characteristic curves for the two-terminal
the polarity of the transient can be positive, devices described in the preceding sections.
negative, or both, transient protection circuits The parameters shown by the curves depend
can be designed with two devices connected on the type of circuit in which they are mea-
with opposite polarities. sured, such as common emitter or common
collector. The output characteristic shows a
RECTIFIERS set of curves for either collector or emitter
The most common application of a diode current versus collector-emitter voltage at
is to perform rectification; that is, permitting various values of input current (either base
current flow in only one direction. Power rec- or emitter). The input characteristic shows
tification converts ac current into pulsating dc the voltage between the input and common
current. There are three basic forms of power terminals (such as base-emitter) versus the
rectification using semiconductor diodes: input current for different values of output
half wave (1 diode), full-wave center-tapped voltage.
(2 diodes) and full-wave bridge (4 diodes).
These applications are shown in Fig 3.21A, CURRENT GAIN
B and C and are more fully described in the Two parameters describe the relationships
Power Sources chapter. between the three transistor currents at low
The most important diode parameters to frequencies:
consider for power rectification are the PIV
and current ratings. The peak negative volt- I C
= = 1 (8)
ages that are blocked by the diode must be I E
smaller in magnitude than the PIV and the
peak current through the diode when it is for- I C
= (9)
ward biased must be less than the maximum I B
average forward current.
The relationship between and is defined
Rectification is also used at much lower
as
current levels in modulation and demodula-
tion and other types of analog signal process-
=
ing circuits. For these applications, the diodes 1 + (10)
response speed and junction forward voltage
Another designation for is often used:
are the most important ratings.
hFE, the forward dc current gain. (The h
refers to h parameters, a set of transfer
3.3.2 Bipolar Junction parameters for describing a two-port net-
Transistors (BJT) work and described in more detail in the RF
Fig 3.21 Diode rectifier circuits. (A) Techniques chapter.) The symbol, hfe, in
The bipolar transistor is a current-con- Half wave rectifier circuit. Only when the which the subscript is in lower case, is used
trolled device with three basic terminals; emit- ac voltage is positive does current pass for the forward current gain of ac signals.
ter, collector and base. The current between through the diode. Current flows only
the emitter and the collector is controlled by during half of the cycle. (B) Full-wave OPERATING REGIONS
the current between the base and emitter. The center-tapped rectifier circuit. Center-tap
on the transformer secondary is grounded Current conduction between collector
convention when discussing transistor opera- and emitter is described by regions of the
and the two ends of the secondary are
tion is that the three currents into the device 180 out of phase. (C) Full-wave bridge transistors characteristic curves in Fig 3.22.
are positive (Ic into the collector, Ib into the rectifier circuit. In each half of the cycle (References such as common-emitter or com-
base and Ie into the emitter). Kirchhoffs Cur- two diodes conduct. mon-base refer to the configuration of the
rent Law (see the Electrical Fundamentals circuit in which the parameter is measured.)
chapter) applies to transistors just as it does The transistor is in its active or linear region
to passive electrical networks: the total cur- example, if we are interested in the emitter when the base-collector junction is reverse bi-
rent entering the device must be zero. Thus, current, ased and the base-emitter junction is forward
the relationship between the currents into a biased. The slope of the output current, IO,
transistor can be generalized as ( I c + I b ) (7)
Ie = versus the output voltage, VO, is virtually flat,
indicating that the output current is nearly in-
R pm (6)
+ Vnm The back-to-back diode model shown in dependent of the output voltage. In this region,
R m + R pm the output circuit of the transistor can be mod-
Fig 3.9 is appropriate for visualization of
which can be rearranged as necessary. For transistor construction. In actual transistors, eled as a constant-current source controlled by

Analog Basics 3.17


Fig 3.22 Transistor response curve output characteristics. The x-axis is the output
voltage, and the y-axis is the output current. Different curves are plotted for various
values of input current. The three regions of the transistor are its cutoff region, where
no current flows in any terminal, its active region, where the output current is nearly
independent of the output voltage and there is a linear relationship between the input
current and the output current, and the saturation region, where the output current has
large changes for small changes in output voltage.

the input current. The slight slope that does In the OPERATING PARAMETERS section,
exist is due to base-width modulation (known three guaranteed minimum junction break-
as the Early effect). down voltages are listed V(BR)CEO, V(BR)CBO
When both the junctions in the transistor and V(BR)EBO. Exceeding these voltages is
are forward biased, the transistor is said to be likely to cause the transistor to enter avalanche
in its saturation region. In this region, VO is breakdown, but if current is limited, perma-
nearly zero and large changes in IO occur for nent damage may not result.
very small changes in VO. Both junctions in Under ON CHARACTERISTICS are the guar-
the transistor are reverse-biased in the cutoff anteed minimum dc current gain ( or hFE),
region. Under this condition, there is very guaranteed maximum collector-emitter satu-
little current in the output, only the nanoam- ration voltage, VCE(SAT), and the guaranteed
peres or microamperes that result from the maximum base-emitter on voltage, VBE(ON).
very small leakage across the input-to-output Two guaranteed maximum collector cutoff
junction. Finally, if VO is increased to very currents, ICEO and ICBO, are listed under OFF
high values, avalanche breakdown begins as CHARACTERISTICS.
in a PN-junction diode and output current in- The next section is SMALL-SIGNAL CHARAC-
creases rapidly. This is the breakdown region, TERISTICS, where the guaranteed minimum
not shown in Fig 3.22. current gain-bandwidth product, BW or fT,
These descriptions of junction conditions the guaranteed maximum output capacitance,
are the basis for the use of transistors. Various Cobo, the guaranteed maximum input capaci-
configurations of the transistor in circuitry tance, Cibo, the guaranteed range of input im-
make use of the properties of the junctions pedance, hie, the small-signal current gain, hfe,
to serve different purposes in analog signal the guaranteed maximum voltage feedback
processing. ratio, hre and output admittance, hoe are listed.
Finally, the SWITCHING CHARACTERISTICS
OPERATING PARAMETERS section lists absolute maximum ratings for
A typical general-purpose bipolar-transis- delay time, td; rise time, t r; storage time, ts; Fig 3.23FET schematic symbols.
tor data sheet lists important device specifi- and fall time, t f.
cations. Parameters listed in the ABSOLUTE
MAXIMUM RATINGS section are the three junc-
tion voltages (VCEO, VCBO and VEBO), the 3.3.3 Field-Effect Transistors
continuous collector current (IC), the total (FET) to the emitter, and the drain to the collector.
device power dissipation (PD) and the operat- FET devices are controlled by the voltage Symbols for the various forms of FET devices
ing and storage temperature range. Exceeding level of the input rather than the input current, are pictured in Fig 3.23.
any of these parameters is likely to cause the as in the bipolar transistor. FETs have three The FET gate has a very high impedance,
transistor to be destroyed. (The O in the suf- basic terminals, the gate, the source and the so the input can be modeled as an open circuit.
fixes of the junction voltages indicates that the drain. They are analogous to bipolar transis- The voltage between gate and source, VGS,
remaining terminal is not connected, or open.) tor terminals: the gate to the base, the source controls the resistance of the drain-source

3.18 Chapter 3
JFET), rDS(on) is minimum. This describes
the effectiveness of the device as an analog
switch. Channel resistance is approximately
the same for ac and dc signals until at high
frequencies the capacitive reactances inher-
ent in the FET structure become significant.
FETs also have strong similarities to
vacuum tubes in that input voltage between
the grid and cathode controls an output cur-
rent between the plate and cathode. (See the
chapter on RF Power Amplifiers for more
information on vacuum tubes.)
Fig 3.24 JFET input leakage curves for
common source amplifier configuration. FORWARD TRANSCONDUCTANCE
Input voltage (VGS) on the x-axis versus The change in FET drain current caused by
input current (IG) on the y-axis, with two a change in gate-to-source voltage is called Fig 3.26 JFET operating regions. At the
curves plotted for different operating left, ID is increasing rapidly with VGS and
temperatures, 25 C and 125 C. Input forward transconductance, gm.
the JFET can be treated as resistance
current increases greatly when the gate I DS (RDS) controlled by VGS. In the saturation
voltage exceeds the junction breakpoint gm = region, drain current, ID, is relatively
voltage. VGS independent of VGS. As VDS increases
or further, avalanche breakdown begins and
ID increases rapidly.
channel, rDS, and so the output of the FET is I DS =g m VGS (11)
modeled as a current source, whose output
current is controlled by the input voltage. The input voltage, VGS, is measured between
The action of the FET channel is so nearly the FET gate and source and drain current, to the configuration of the circuit in which the
ideal that, as long as the JFET gate does not IDS, flows from drain to source. Analogous parameter is measured.) Transconductance
become forward biased and inject current to a bipolar transistors current gain, the curves relate the drain current, ID, to gate-to-
from the base into the channel, the drain and units of transconductance are Siemens (S) source voltage, VGS, at various drain-source
source currents are virtually identical. For because it is the ratio of current to voltage. voltages, VDS. The FETs forward transcon-
JFETs the gate leakage current, IG, is a func- (Both gm and gfs are used interchangeably ductance, gm, is the slope of the lines in the
tion of VGS and this is often expressed with an to indicate transconductance. Some sources forward transconductance curve. The same
input curve (see Fig 3.24). The point at which specify gfs as the common-source forward parameters are interrelated in a different way
there is a significant increase in IG is called transconductance. This chapter uses gm, the in the output characteristic, in which ID is
the junction breakpoint voltage. Because the most common convention in the reference shown versus VDS for different values of VGS.
gate of MOSFETs is insulated from the chan- literature.) Like the bipolar transistor, FET operation
nel, gate leakage current is insignificant in can be characterized by regions. The ohmic
these devices. OPERATING REGIONS region is shown at the left of the FET output
The dc channel resistance, rDS, is specified The most useful relationships for FETs characteristic curve in Fig 3.26 where ID is in-
in data sheets to be less than a maximum value are the output and transconductance response creasing nearly linearly with VDS and the FET
when the device is biased on (rDS(on)). When characteristic curves in Fig 3.25. (References is acting like a resistance controlled by VGS.
the gate voltage is maximum (VGS = 0 for a such as common-source or common-gate refer As VDS continues to increase, ID saturates and
becomes nearly constant. This is the FETs
saturation region in which the channel of the
FET can be modeled as a constant-current
source. VDS can become so large that VGS no
longer controls the conduction of the device
and avalanche breakdown occurs as in bipolar
transistors and PN-junction diodes. This is the
breakdown region, shown in Fig 3.26 where
the curves for ID break sharply upward. If VGS
is less than VP, so that transconductance is
zero, the FET is in the cutoff region.

OPERATING PARAMETERS
A typical FET data sheet gives ABSOLUTE
MAXIMUM RATINGS for VDS, VDG, VGS and ID,
along with the usual device dissipation (PD)
and storage temperature range. Exceeding
these limits usually results in destruction of
the FET.
Fig 3.25 JFET output and transconductance response curves for common source Under OPERATING PARAMTERS the OFF
amplifier configuration. (A) Output voltage (VDS) on the x-axis versus output current (ID)
on the y-axis, with different curves plotted for various values of input voltage (VGS). (B) CHARACTERISTICS list the gate-source break-
Transconductance curve with the same three variables rearranged: VGS on the x-axis, ID down voltage, VGS(BR), the reverse gate cur-
on the y-axis and curves plotted for different values of VDS. rent, IGSS and the gate-source cutoff voltage,

Analog Basics 3.19


gate-source voltage, VGS, the drain current
can be considered to be constant over a wide
range of drain-source voltages.
Multiple gate MOSFETs are also available.
Due to the insulating layer, the two gates are
isolated from each other and allow two sig-
nals to control the channel simultaneously
with virtually no loading of one signal by the
other. A common application of this type of
device is an automatic gain control (AGC)
amplifier. The signal is applied to one gate
and a rectified, low-pass filtered form of the
output (the AGC voltage) is fed back to the
other gate. Another common application is
as a mixer in which the two input signals are
applied to the pair of gates.
MOSFET Gate Protection
The MOSFET is constructed with a very
thin layer of SiO2 for the gate insulator. This
layer is extremely thin in order to improve the
transconductance of the device but this makes
it susceptible to damage from high voltage
levels, such as electrostatic discharge (ESD)
from static electricity. If enough charge ac-
cumulates on the gate terminal, it can punch
through the gate insulator and destroy it. The
insulation of the gate terminal is so good that
virtually none of this potential is eased by
leakage of the charge into the device. While
Fig 3.27 MOSFET output [(A) and (C)] and transconductance [(B) and (D)] response this condition makes for nearly ideal input
curves. Plots (A) and (B) are for an N-channel depletion mode device. Note that impedance (approaching infinity), it puts the
VGS varies from negative to positive values. Plots (C) and (D) are for an N-channel
enhancement mode device. VGS has only positive values. device at risk of destruction from even such
seemingly innocuous electrical sources as
static electrical discharges from handling.
Some MOSFET devices contain an internal
Zener diode with its cathode connected to
VGS(OFF). Exceeding VGS(BR) will not perma- need input and reverse transconductance the gate and its anode to the substrate. If the
nently damage the device if current is limited. curves. Their output curves (Fig 3.27) are voltage at the gate rises to a damaging level
The primary ON CHARACTERISTIC parameters similar to those of the JFET. The gate acts the Zener diode junction conducts, bleeding
are the channel resistance, rDS, and the zero- as a small capacitance between the gate and excess charges off to the substrate. When volt-
gate-voltage drain current (IDSS). An FETs both the source and drain. ages are within normal operating limits the
dc channel resistance, rDS, is specified in data The output and transconductance curves in Zener has little effect on the signal at the gate,
sheets to be less than a maximum value when Fig 3.27A and 3.27B show that the depletion- although it may decrease the input impedance
the device is biased on (rDS(on)). For ac signals, mode N-channel MOSFETs transconduc- of the MOSFET.
rds(on) is not necessarily the same as rDS(on), but tance is positive at VGS = 0, like that of the This solution will not work for all MOS-
it is not very different as long as the frequency N-channel JFET. Unlike the JFET, however, FETs. The Zener diode must always be reverse
is not so high that capacitive reactance in the increasing VGS does not forward-bias the biased to be effective. In the enhancement-
FET becomes significant. gate-source junction and so the device can mode MOSFET, VGS > 0 for all valid uses
The SMALL SIGNAL CHARACTERISTICS in- be operated with VGS > 0. of the part, keeping the Zener reverse biased.
clude the forward transfer admittance, yfs, the In the enhancement-mode MOSFET, trans- In depletion mode devices however, VGS can
output admittance, yos, the static drain-source conductance is zero at VGS = 0. As VGS is be both positive and negative; when negative,
on resistance, rds(on) and various capacitanc- increased, the MOSFET enters the ohmic a gate-protection Zener diode would be for-
es such as input capacitance, Ciss, reverse region. If VGS increases further, the satura- ward biased and the MOSFET gate would not
transfer capacitance, Crss, the drain-substrate tion region is reached and the MOSFET is be driven properly. In some depletion mode
capacitance, Cd(sub). FUNCTIONAL CHARAC- said to be fully-on, with rDS at its minimum MOSFETs, back-to-back Zener diodes are
TERISTICS include the noise figure, NF, and value. The behavior of the enhancement- used to protect the gate.
the common source power gain, Gps. mode MOSFET is similar to that of the bipolar MOSFET devices are at greatest risk of
transistor in this regard. damage from static electricity when they are
MOSFETS The relatively flat regions in the MOSFET out of circuit. Even though an electrostatic
As described earlier, the MOSFETs gate output curves are often used to provide a con- discharge is capable of delivering little energy,
is insulated from the channel by a thin layer stant current source. As is plotted in these it can generate thousands of volts and high
of nonconductive oxide, doing away with any curves, the drain current, ID, changes very peak currents. When storing MOSFETs, the
appreciable gate leakage current. Because of little as the drain-source voltage, VDS, varies leads should be placed into conductive foam.
this isolation of the gate, MOSFETs do not in this portion of the curve. Thus, for a fixed When working with MOSFETs, it is a good

3.20 Chapter 3
idea to minimize static by wearing a grounded frequencies and add little noise to the signal. to allow it to cross a PN junctions depletion
wrist strap and working on a grounded work- The reason GaAsFETs have gain at these fre- region as current flow through the semicon-
bench or mat. A humidifier may help to de- quencies is the high mobility of the electrons ductor (photoelectricity).
crease the static electricity in the air. Before in GaAs material. Because the electrons are
inserting a MOSFET into a circuit board it more mobile than in silicon, they respond to PHOTOCONDUCTORS
helps to first touch the device leads with your the gate-source input signal more quickly and In commercial photoconductors (also
hand and then touch the circuit board. This strongly than silicon FETs, providing gain at called photoresistors) the resistance can
serves to equalize the excess charge so that higher frequencies (fT is directly proportional change by as much as several kilohms for
little excess charge flows when the device is to electron mobility). The higher electron mo- a light intensity change of 100 ft-candles.
inserted into the circuit board. bility also reduces thermally-generated noise The most common material used in photo-
generated in the FET, making the GaAsFET conductors is cadmium sulfide (CdS), with a
Power MOSFETs especially suitable for weak-signal preamps. resistance range of more than 2 M in total
Power MOSFETs are designed for use as Because electron mobility is always higher darkness to less than 10 in bright light.
switches, with extremely low values of rDS(on); than hole mobility, N-type material is used in Other materials used in photoconductors re-
values of 50 milliohms (m) are common. GaAsFETs to maximize high-frequency gain. spond best at specific colors. Lead sulfide
The largest devices of this type can switch tens Since P-type material is not used to make a (PbS) is most sensitive to infrared light and
of amps of current with VDS voltage ratings gate-channel junction, a metal Schottky junc- selenium (Se) works best in the blue end of
of hundreds of volts. The Component Data tion is formed by depositing metal directly on the visible spectrum.
and References chapter includes a table of the surface of the channel. This type of device
Power FET ratings. The schematic symbol is also called a MESFET (metal-semiconduc- PHOTODIODES
for power MOSFETs (see Fig 3.23) includes tor field-effect transistor). A similar effect is used in some diodes
a body diode that allows the FET to conduct and transistors so that their operation can be
in the reverse direction, regardless of VGS. controlled by light instead of electrical current
This is useful in many high-power switching 3.3.4 Optical Semiconductors biasing. These devices, shown in Fig 3.28, are
applications. Power MOSFETs used for RF In addition to electrical energy and heat called photodiodes and phototransistors. The
amplifiers are discussed in more detail in the energy, light energy also affects the behav- flow of minority carriers across the reverse
RF Power Amplifiers chapter. ior of semiconductor materials. If a device biased PN junction is increased by light fall-
While the maximum ratings for current and is made to allow photons of light to strike ing on the doped semiconductor material. In
voltage are high, the devices cannot withstand the surface of the semiconductor material, the dark, the junction acts the same as any
both high drain current and high drain-to- the energy absorbed by electrons disrupts the reverse biased PN junction, with a very low
source voltage at the same time because of the bonds between atoms, creating free electrons current, ISC, (on the order of 10 A) that is
power dissipated; P = VDS ID. It is important and holes. This increases the conductivity of nearly independent of reverse voltage. The
to drive the gate of a power MOSFET such the material (photoconductivity). The photon presence of light not only increases the current
that the device is fully on or fully off so that can also transfer enough energy to an electron but also provides a resistance-like relationship
either VDS or ID is at or close to zero. When
switching, the device should spend as little
time as possible in the linear region where
both current and voltage are nonzero because
their product (P) can be substantial. This is not
a big problem if switching only takes place
occasionally, but if the switching is repetitive
(such as in a switching power supply) care
should be taken to drive the gate properly and
remove excess heat from the device.
Because the gate of a power MOSFET is
capacitive (up to several hundred pF for large
devices), charging and discharging the gate
quickly results in short current peaks of more
than 100 mA. Whatever circuit is used to drive
the gate of a power MOSFET must be able to
handle that current level, such as an integrated
circuit designed for driving the capacitive load
an FET gate presents.
The gate of a power MOSFET should not
be left open or connected to a high-impedance
circuit. Use a pull-down or pull-up resistor
connected between the gate and the appro-
priate power supply to ensure that the gate
is placed at the right voltage when not being
driven by the gate drive circuit.

GaAsFETs
Fig 3.28 The photodiode (A) is used to detect light. An amplifier circuit changes
FETs made from gallium-arsenide (GaAs) the variations in photodiode current to a change in output voltage. At (B), a photo-
material are used at UHF and microwave transistor conducts current when its base is illuminated. This causes the voltage at
frequencies because they have gain at these the collector to change causing the amplifiers output to switch between ON and OFF.

Analog Basics 3.21


minority carriers. As the reverse voltage is
reduced, the potential barrier to the forward
flow of majority carriers is also reduced.
Since light energy leads to the generation
of both majority and minority carriers, when
the resistance to the flow of majority carriers
is decreased these carriers form a forward
current. The voltage at which the forward
current equals the reverse current is called
the photovoltaic potential of the junction. If
the illuminated PN junction is not connected
to a load, a voltage equal to the photovoltaic
potential can be measured across it as the
terminal voltage, VT, or open-circuit volt-
age, VOC.
Fig 3.29 Photodiode I-V curve. Reverse voltage is plotted on the x-axis and current Devices that use light from the sun to pro-
through diode is plotted on the y-axis. Various response lines are plotted for different duce electricity in this way are called photo-
illumination. Except for the zero illumination line, the response does not pass through voltaic (PV) or solar cells or solar batteries.
the origin since there is current generated at the PN junction by the light energy. A load
line is shown for a 50-k resistor in series with the photodiode.
The symbol for a photovoltaic cell is shown
in Fig 3.30A. The electrical equivalent circuit
of the cell is shown in Fig 3.30B. The cell is
basically a large, flat diode junction exposed
(reverse current increases as reverse voltage between the collector and emitter. Thus the to light. Metal electrodes on each side of the
increases). See Fig 3.29 for the characteris- phototransistor acts as an amplifier whose junction collect the current generated.
tic response of a photodiode. Even with no input signal is light and whose output is cur- When illuminated, the cell acts like a cur-
reverse voltage applied, the presence of light rent. Phototransistors are more sensitive to rent source, with some of the current flowing
causes a small reverse current, as indicated light than the other devices. Phototransistors through a diode (made of the same material
by the points at which the lines in Fig 3.29 have lots of light-to-current gain, but photo- as the cell), a shunt resistance for leakage
intersect the left side of the graph. diodes normally have less noise, so they make current and a series resistor that represents
Photoconductors and photodiodes are gen- more sensitive detectors. The phototransistor the resistance of the cell. Two quantities de-
erally used to produce light-related analog in Fig 3.28B is being used as a detector. Light fine the electrical characteristics of common
signals that require further processing. For ex- falling on the phototransistor causes collector silicon photovoltaic cells. These are an open-
ample, a photodiode is used to detect infrared current to flow, dropping the collector voltage circuit voltage, VOC of 0.5 to 0.6 V and the
light signals from remote control devices as below the voltage at the amplifiers + input output short-circuit current, ISC as above, that
in Fig 3.28A. The light falling on the reverse- and causing a change in VOUT. depends on the area of the cell exposed to
biased photodiode causes a change in ISC that light and the degree of illumination. A mea-
is detected as a change in output voltage. PHOTOVOLTAIC CELLS sure of the cells effectiveness at converting
Light falling on the phototransistor acts When illuminated, the reverse-biased pho- light into current is the conversion efficiency.
as base current to control a larger current todiode has a reverse current caused by excess Typical silicon solar cells have a conversion
efficiency of 10 to 15% although special cells
with stacked junctions or using special light-
absorbing materials have shown efficiencies
as high as 40%.
Solar cells are primarily made from sin-
gle-crystal slices of silicon, similar to diodes
and transistors, but with a much greater area.
Polycrystalline silicon and thin-film cells are
less expensive, but have lower conversion
efficiency. Technology is advancing rapidly
in the field of photovoltaic energy and there
are a number of different types of materials
and fabrication techniques that have promise
in surpassing the effectiveness of the single-
junction silicon cells.
Solar cells are assembled into arrays called
solar panels, shown in Fig 3.30C. Cells are
connected in series so that the combined out-
put voltage is a more useful voltage, such as 12
V. Several strings of cells are then connected
in parallel to increase the available output
current. Solar panels are available with output
powers from a few watts to hundreds of watts.
Note that unlike batteries, strings of solar cells
Fig 3.30 A photovoltaic cells symbol (A) is similar to a battery. Electrically, the cell
can be modeled as the equivalent circuit at (B). Solar panels (C) consist of arrays of can be connected directly in parallel because
cells connected to supply power at a convenient voltage. they act as sources of constant current instead

3.22 Chapter 3
of voltage. (More information on the use of (V) with a series resistor (R) that limits the
solar panels for powering radio equipment can current to the desired level (IF) for the amount
be found in the chapter on Power Sources.) of light to be generated.

LIGHT EMITTING DIODES AND V VF


R= (12)
LASER DIODES IF
In the photodiode, energy from light falling where VF is the forward voltage of the LED.
on the semiconductor material is absorbed to The cathode lead is connected to the lower
create additional electron-hole pairs. When potential, and is specially marked as shown
the electrons and holes recombine, the same in the manufacturers data sheet. LEDs may
amount of energy is given off. In normal di- be connected in series for additional light,
odes the energy from recombination of car- with the same current flowing in all of the
riers is given off as heat. In certain forms of diodes. Diodes connected in parallel without
semiconductor material, the recombination current-limiting resistors for each diode are
energy is given off as light with a mecha- Fig 3.32 The optoisolator consists
likely to share the current unequally, thus the of an LED (input) that illuminates the
nism called electroluminescence. Unlike the series connection is preferred. base of a phototransistor (output).
incandescent light bulb, electroluminescence The laser diode operates by similar prin- The phototransistor then conducts
is a cold (non-thermal) light source that typi- ciples to the LED except that all of the light current in the output circuit. CTR is the
cally operates with low voltages and currents optoisolators current transfer ratio.
produced is monochromatic (of the same col-
(such as 1.5 V and 10 mA). Devices made for or and wavelength) and it is coherent, mean-
this purpose are called light emitting diodes ing that all of the light waves emitted by the
(LEDs). They have the advantages of low device are in phase. Laser diodes generally
power requirements, fast switching times (on somewhat expensive.
require higher current than an LED and will As an example of voltage level shifting, an
the order of 10 ns) and narrow spectra (rela- not emit light until the lasing current level is
tively pure color). optoisolator can be used to allow a low-volt-
reached. Because the light is monochromatic age, solid-state electronic Morse code keyer
The LED emits light when it is forward and coherent, laser diodes can be used for ap-
biased and excess carriers are present. As the to activate a vacuum-tube grid-block keying
plications requiring precise illumination and circuit that operates at a high negative voltage
carriers recombine, light is produced with a modulation, such as high-speed data links,
color that depends on the properties of the (typically about 100 V) but low current. No
and in data storage media such as CD-ROM common ground is required between the two
semiconductor material used. Gallium-arse- and DVD. LEDs are not used for high-speed
nide (GaAs) generates light in the infrared pieces of equipment.
or high-frequency analog modulation because Optoisolators can act as input protection for
region, gallium-phosphide (GaP) gives off of recovery time limitations, just as in regular
red light when doped with oxygen or green circuits that are exposed to high voltages or
rectifiers. transients. For example, a short 1000-V tran-
light when doped with nitrogen. Orange light
is attained with a mixture of GaAs and GaP OPTOISOLATORS sient that can destroy a semiconductor circuit
(GaAsP). Silicon-carbide (SiC) creates a blue will only saturate the LED in the optoisolator,
An interesting combination of optoelec- preventing damage to the circuit. The worst
LED. tronic components proves very useful in
White LEDs are made by coating the inside that will happen is the LED in the optoisolator
many analog signal processing applications. will be destroyed, but that is usually quite a bit
of the LED lens with a white-light emitting An optoisolator consists of an LED optically
phosphor and illuminating the phosphor with less expensive than the circuit it is protecting.
coupled to a phototransistor, usually in an Optoisolators are also useful for isolating
light from a single-color LED. White LEDs enclosed package (see Fig 3.32). The optoiso-
are currently approaching the cost of cold- different ground systems. The input and out-
lator, as its name suggests, isolates different put signals are totally isolated from each other,
florescent (CFL) bulbs and will eventually circuits from each other. Typically, isolation
displace CFL technology for lighting, just even with respect to the references for each
resistance is on the order of 1011 and iso- signal. A common application for optoisola-
as CFL is replacing the incandescent bulb. lation capacitance is less than 1 pF. Maxi-
The LED, shown in Fig 3.31, is very simple tors is when a computer is used to control
mum voltage isolation varies from 1000 to radio equipment. The computer signal, and
to use. It is connected across a voltage source 10,000 V ac. The most common optoisolators even its ground reference, typically contains
are available in 6-pin DIP packages. considerable wide-band noise caused by the
Optoisolators are primarily used for volt- digital circuitry. The best way to keep this
age level shifting and signal isolation. Voltage noise out of the radio is to isolate both the
level shifting allows signals (usually digital signal and its reference; this is easily done
signals) to pass between circuits operating at with an optoisolator.
greatly different voltages. The isolation has The design of circuits with optoisolators is
two purposes: to protect circuitry (and opera- not greatly different from the design of circuits
tors) from excessive voltages and to isolate with LEDs and with transistors. On the input
noisy circuitry from noise-sensitive circuitry. side, the LED is forward-biased and driven
Optoisolators also cannot transfer signals with a series current-limiting resistor whose
with high power levels. The power rating value limits current to less than the maximum
of the LED in a 4N25 device is 120 mW. value for the device (for example, 60 mA is the
Optoisolators have a limited frequency re- maximum LED current for a 4N25). This is
Fig 3.31 A light-emitting diode (LED) sponse due to the high capacitance of the
emits light when conducting forward identical to designing with standalone LEDs.
LED. A typical bandwidth for the 4N25 On the output side, instead of current gain
current. A series current-limiting resistor
is used to set the current through the LED
series is 300 kHz. Optoisolators with band- for a transistor, the optoisolators current
according to the equation. widths of several MHz are available, but are transfer ratio (CTR) is used. CTR is a ratio

Analog Basics 3.23


given in percent between the amount of cur- are placed in the same package and miniature for the demanding military and aerospace
rent through the LED to the output transistors wires are connected between them to make applications.
maximum available collector current. For complete circuits. On top of the P-type substrate is a thin layer
example, if an optoisolators CTR = 25%, Hybrid circuits miniaturize analog elec- of N-type material in which the active and
then an LED current of 20 mA results in the tronic circuits by replacing much of the pack- passive components are built. Impurities are
output transistor being able to conduct up to aging that is inherent in discrete electronics. diffused into this layer to form the appropriate
20 0.25 = 5 mA of current in its collector The term discrete refers to the use of indi- component at each location. To prevent ran-
circuit. vidual components to make a circuit, each in dom diffusion of impurities into the N-layer,
If the optoisolator is to be used for an analog its own package. The individual components its upper surface must be protected. This is
signal, the input signal must be appropriately are attached together on a small circuit board done by covering the N-layer with a layer of
dc shifted so that the LED is always forward or with bonding wires. silicon dioxide (SiO2). Wherever diffusion of
biased. A phototransistor with all three leads Once widespread in electronics, hybrid impurities is desired, the SiO2 is etched away.
available for connection (as in Fig 3.32) is ICs have largely been displaced by fully in- The precision of placing the components on
required. The base lead is used for biasing, tegrated devices with all the components on the semiconductor material depends mainly
allowing the optical signal to create variations the same piece of semiconductor material. on the fineness of the etching. The fourth
above and below the transistors operating Manufacturers often use hybrids when small layer of an IC is made of aluminum (copper
point. The collector and emitter leads are used size is needed, but there is insufficient volume is used in some high-speed digital ICs) and
as they would be in any transistor amplifier to justify the expense of a custom IC. is used to make the interconnections between
circuit. (There are also linear optoisolators A current application for hybrid circuitry is the components.
that include built-in linearizing circuitry.) The UHF and microwave amplifiers they are in Different components are made in a sin-
use of linear optoisolators is not common. wide use by the mobile phone industry. For ex- gle piece of semiconductor material by first
ample, the Motorola MW4IC915N Wideband diffusing a high concentration of acceptor
FIBER OPTICS Integrated Power Amplifier is a complete impurities into the layer of N-type material.
An interesting variation on the optoisolator 15-W transmitting module. Its TO-272 pack- This process creates P-type semiconductor
is the fiber-optic connection. Like the optoiso- age is only about 1 inch long by 38-inch wide. often referred to as P+-type semiconductor
lator, the input signal is used to drive an LED This particular device is designed for use be- because of its high concentration of accep-
or laser diode that produces modulated light tween 750 and 1000 MHz and can be adapted tor atoms that isolates regions of N-type
(usually light pulses). The light is transmit- for use on the amateur 902 MHz band. Other material. Each of these regions is then further
ted in a fiber optic cable, an extruded glass devices available as hybrid circuits include processed to form single components.
fiber that efficiently carries light over long oscillators, signal processors, preamplifiers A component is produced by the diffusion
distances and around fairly sharp bends. The and so forth. Surplus hybrids can be hard to of a lesser concentration of acceptor atoms
signal is recovered by a photo detector (photo adapt to amateur use unless they are clearly into the middle of each isolation region. This
resistor, photodiode or phototransistor). Be- identified with manufacturing identification results in an N-type isolation well that contains
cause the fiber optic cable is nonconductive, such that a data sheet can be obtained. P-type material, is surrounded on its sides by
the transmitting and receiving systems are P+-type material and has P-type material (sub-
electrically isolated. MONOLITHIC INTEGRATED strate) below it. The cross sectional view in
Fiber optic cables generally have far less CIRCUITS Fig 3.33B illustrates the various layers. Con-
loss than coaxial cable transmission lines. In order to build entire circuits on a single nections to the metal layer are often made by
They do not leak RF energy, nor do they piece of semiconductor, it must be possible diffusing high concentrations of donor atoms
pick up electrical noise. Fiber optic cables to fabricate resistors and capacitors, as well into small regions of the N-type well and the
are virtually immune to electromagnetic as transistors and diodes. Only then can the P-type material in the well. The material in
interference! Special forms of LEDs and entire circuit be created on one piece of silicon these small regions is N+-type and facilitates
phototransistors are available with the ap- called a monolithic integrated circuit. electron flow between the metal contact and
propriate optical couplers for connecting to An integrated circuit (IC) or chip is fabri- the semiconductor. In some configurations, it
fiber optic cables. These devices are typically cated in layers. An example of a semiconduc- is necessary to connect the metal directly to
designed for higher frequency operation with tor circuit schematic and its implementation the P-type material in the well.
gigahertz bandwidth. in an IC is pictured in Fig 3.33. The base
layer of the circuit, the substrate, is made Fabricating Resistors and
of P-type semiconductor material. Although Capacitors
3.3.5 Linear Integrated less common, the polarity of the substrate can An isolation well can be made into a resistor
Circuits also be N-type material. Since the mobility of by making two contacts into the P-type semi-
If you look inside a transistor, the actual electrons is about three times higher than that conductor in the well. Resistance is inversely
size of the semiconductor is quite small com- of holes, bipolar transistors made with N-type proportional to the cross-sectional area of the
pared to the size of the packaging. For most collectors and FETs made with N-type chan- well. An alternate type of resistor that can
semiconductors, the packaging takes consid- nels are capable of higher speeds and power be integrated in a semiconductor circuit is
erably more space than the actual semicon- handling. Thus, P-type substrates are far more a thin-film resistor, where a metallic film is
ductor device. Thus, an obvious way to reduce common. For devices with N-type substrates, deposited on the SiO2 layer, masked on its
the physical size of circuitry is to combine all polarities in the ensuing discussion would upper surface by more SiO2 and then etched
more of the circuit inside a single package. be reversed. to make the desired geometry, thus adjusting
Other substrates have been used, one of the resistance.
HYBRID INTEGRATED CIRCUITS the most successful of which is the silicon- There are two ways to form capacitors in
It is easy to imagine placing several small on-sapphire (SOS) construction that has been a semiconductor. One is to make use of the
semiconductor chips in the same package. used to increase the bandwidth of integrated PN junction between the N-type well and the
This is known as hybrid circuitry, a technol- circuitry. Its relatively high manufacturing P-type material that fills it. Much like a varac-
ogy in which several semiconductor chips cost has impeded its use, however, except tor diode, when this junction is reverse biased

3.24 Chapter 3
Fig 3.33 Integrated circuit layout. (A) Circuit containing two diodes, a resistor, a capacitor, an NPN transistor and an N-channel
MOSFET. Labeled leads are D for diode, R for resistor, DC for diode-capacitor, E for emitter, S for source, CD for collector-drain and
G for gate. (B) Integrated circuit that is identical to circuit in (A). Same leads are labeled for comparison. Circuit is built on a P-type
semiconductor substrate with N-type wells diffused into it. An insulating layer of SiO2 is above the semiconductor and is etched
away where aluminum metal contacts are made with the semiconductor. Most metal-to-semiconductor contacts are made with
heavily doped N-type material (N+-type semiconductor).

a capacitance results. Since a bias voltage is larly at lower frequencies, the behavior of an of the well, made of N-type semiconduc-
required, this type of capacitor is polarized, inductor can be mimicked by an amplifier tor, forms the collector, the P-type material
like an electrolytic capacitor. Nonpolarized circuit. In many cases the appropriate design in the well forms the base and a small region
capacitors can also be formed in an integrated of IC amplifiers can reduce or eliminate the of N+-type material formed in the center of
circuit by using thin film technology. In this need for external inductors. the well becomes the emitter. A PNP transis-
case, a very high concentration of donor ions tor is made by diffusing donor ions into the
is diffused into the well, creating an N+-type Fabricating Diodes and Transistors P-type semiconductor in the well to make
region. A thin metallic film is deposited over The simplest form of diode is generated by a pattern with P-type material in the center
the SiO2 layer covering the well and the ca- connecting to an N+-type connection point in (emitter) surrounded by a ring of N-type ma-
pacitance is created between the metallic film the well for the cathode and to the P-type well terial that connects all the way down to the
and the well. The value of the capacitance material for the anode. Diodes are often con- well material (base), and this is surrounded
is adjusted by varying the thickness of the verted from NPN transistor configurations. by another ring of P-type material (collec-
SiO2 layer and the cross-sectional size of Integrated circuit diodes made this way can tor). This configuration results in a large base
the well. This type of thin film capacitor is either short the collector to the base or leave width separating the emitter and collector,
also known as a metal oxide semiconductor the collector unconnected. The base contact causing these devices to have much lower
(MOS) capacitor. is the anode and the emitter contact is the current gain than the NPN form. This is one
Unlike resistors and capacitors, it is very cathode. reason why integrated circuitry is designed
difficult to create inductors in integrated Transistors are created in integrated cir- to use many more NPN transistors than PNP
circuits. Generally, RF circuits that need cuitry in much the same way that they are transistors.
inductance require external inductors to be fabricated in their discrete forms. The NPN FETs can also be fabricated in IC form as
connected to the IC. In some cases, particu- transistor is the easiest to make since the wall shown in Fig 3.33C. Due to its many func-

Analog Basics 3.25


tional advantages, the MOSFET is the most of the individual features that can be created radiation and reception of spurious signals.
common form used for digital ICs. MOS- on the semiconductor wafer. Since the inven- As frequencies increase and wavelengths ap-
FETs are made in a semiconductor chip much tion of the monolithic IC in the mid-1960s, proach the dimensions of the wires in a circuit
the same way as MOS capacitors, described feature size limits have dropped below 100 board, the interconnections act as efficient an-
earlier. In addition to the signal processing nanometers (110th of a millionth of a meter) tennas. The dimensions of the circuitry within
advantages offered by MOSFETs over other as of 2009. Currently, it is not unusual to find an IC are orders of magnitude smaller than
transistors, the MOSFET device can be fab- chips with more than one million transistors in discrete circuitry, thus greatly decreasing
ricated in 5% of the physical space required on them. this problem and permitting the processing of
for bipolar transistors. CMOS ICs can contain In addition to size and reliability of the much higher frequencies with fewer problems
20 times more circuitry than bipolar ICs with ICs themselves, the relative properties of the of interstage interference. Another related ad-
the same chip size, making the devices more devices on a single chip are very predictable. vantage of the smaller interconnections in
powerful and less expensive than those based Since adjacent components on a semiconduc- an IC is the lower inherent inductance of the
on bipolar technology. CMOS is the most tor chip are made simultaneously (the entire wires, and lower stray capacitance between
popular form of integrated circuit. N-type layer is grown at once; a single dif- components and traces.
The final configuration of the switching fusion pass isolates all the wells and another
circuit is CMOS as described in a previous pass fills them), the characteristics of identi- INTEGRATED CIRCUIT
section of this chapter. CMOS gates require cally formed components on a single chip DISADVANTAGES
two FETs, one of each form (NMOS and of silicon are nearly identical. Even if the Despite the many advantages of integrated
PMOS as shown in the figure). NMOS re- exact characteristics of the components are circuitry, disadvantages also exist. ICs have
quires fewer processing steps, and the unknown, very often in analog circuit design not replaced discrete components, even
individual FETs have lower on-resistance than the major concern is how components inter- vacuum tubes, in some applications. There
PMOS. The fabrication of NMOS FETs is the act. For instance, push-pull amplifiers require are some tasks that ICs cannot perform, even
same as for individual semiconductors; P+ perfectly matched transistors, and the gain of though the list of these continues to decrease
wells form the source and drain in a P-type many amplifier configurations is governed by over time as IC technology improves.
substrate. A metal gate electrode is formed on the ratio between two resistors and not their Although the high concentration of com-
top of an insulating SiO2 layer so that the absolute values of resistance. With closely ponents on an IC chip is considered to be an
channel forms in the P-type substrate between matched components on the single substrate, advantage of that technology, it also leads
the source and drain. For the PMOS FET, the this type of design works very well without to a major limitation. Heat generated in the
process is similar, but begins with an N-type requiring external components to adjust or individual components on the IC chip is often
well in the P-type substrate. trim IC performance. difficult to dissipate. Since there are so many
MOSFETs fabricated in this manner also Integrated circuits often have an advantage heat generating components so close together,
have bias (B) terminals connected to the over discrete circuits in their temperature be- the heat can build up and destroy the circuitry.
positive power supply to prevent destructive havior. The variation of performance of the It is this limitation that currently causes many
latch-up. This can occur in CMOS gates be- components on an integrated circuit due to power amplifiers of more than 50 W output to
cause the two MOSFETs form a parasitic heat is no better than that of discrete compo- be designed with discrete components.
SCR. If the SCR mode is triggered and both nents. While a discrete circuit may be exposed Integrated circuits, despite their short inter-
transistors conduct at the same time, large to a wide range of temperature changes, the connection lengths and lower stray inductance,
currents can flow through the FET and destroy entire semiconductor chip generally changes do not have as high a frequency response as
the IC unless power is removed. Just as dis- temperature by the same amount; there are similar circuits built with appropriate dis-
crete MOSFETs are at risk of gate destruction, fewer hot spots and cold spots. Thus, crete components. (There are exceptions to
IC chips made with MOSFET devices have a integrated circuits can be designed to better this generalization, of course. As described
similar risk. They should be treated with the compensate for temperature changes. previously, monolithic microwave integrated
same care to protect them from static electric- A designer of analog devices implemented circuits MMICs are available for op-
ity as discrete MOSFETs. with integrated circuitry has more freedom eration to 10 GHz.) The physical architecture
While CMOS is the most widely used tech- to include additional components that could of an integrated circuit is the cause of this
nology, integrated circuits need not be made improve the stability and performance of the limitation. Since the substrate and the walls
exclusively with MOSFETs or bipolar tran- implementation. The inclusion of compo- of the isolation wells are made of opposite
sistors. It is common to find IC chips designed nents that could cause a prohibitive increase types of semiconductor material, the PN junc-
with both technologies, taking advantage of in the size, cost or complexity of a discrete tion between them must be reverse biased to
the strengths of each. circuit would have very little effect on any of prevent current from passing into the substrate.
these factors in an integrated circuit. Like any other reverse-biased PN junction, a
INTEGRATED CIRCUIT Once an integrated circuit is designed and capacitance is created at the junction and this
ADVANTAGES laid out, the cost of making copies of it is very limits the frequency response of the devices
The primary advantages of using integrated small, often only pennies per chip. Integrated on the IC. This situation has improved over the
circuits as opposed to discrete components circuitry is responsible for the incredible in- years as isolation wells have gotten smaller,
are the greatly decreased physical size of the crease in performance with a corresponding thus decreasing the capacitance between the
circuit and improved reliability. In fact, stud- decrease in price of electronics. While this well and the substrate, and techniques have
ies show that failure rate of electronic circuitry trend is most obvious in digital computers, been developed to decrease the PN junction ca-
is most closely related to the number of in- analog circuitry has also benefited from this pacitance at the substrate. One such technique
terconnections between components. Thus, technology. has been to create an N+-type layer between
using integrated circuits not only reduces vol- The advent of integrated circuitry has also the well and the substrate, which decreases the
ume, but makes the equipment more reliable. improved the design of high frequency circuit- capacitance of the PN junction as seen by the
The amount of circuitry that can be placed ry, particularly the ubiquitous mobile phone well. As a result, analog ICs are now available
onto a single semiconductor chip is a function and other wireless devices. One problem in with gain-bandwidth products over 1 GHz.
of two factors: the size of the chip and the size the design and layout of RF equipment is the A major impediment to the introduction

3.26 Chapter 3
of new integrated circuits, particularly with with specific advantages and disadvantages. tubes in video cameras and film cameras of
special applications, is the very high cost of The vacuum tube, once the dominant signal all types. Liquid crystal displays (LCDs) in
development of new designs for full custom processing component, is relegated to high- laptop computers and standalone computer
ICs. The masking cost alone for a designed power amplifier and display applications and displays have largely displaced the bulky and
and tested integrated circuit can exceed is found only in the RF Power Amplifiers power-hungry cathode-ray tube (CRT), al-
$100,000 so these devices must sell in high chapter of this Handbook. Cathode-ray tubes though it is still found in analog oscilloscopes
volume to recoup the development costs. Over (CRTs) are covered in a supplement on the and some types of video display equipment.
the past decade, however, IC design tools and Handbook CD-ROM. An important consideration in the use of
fabrication services have become available Bipolar transistors, when treated properly, analog components is the future availability
that greatly reduce the cost of IC development can have virtually unlimited life spans. They of parts. At an ever increasing rate, as new
by using predefined circuit layouts and func- are relatively small and, if they do not handle components are developed to replace older
tional blocks. These application-specific in- high currents, do not generate much heat. technology, the older components are dis-
tegrated circuits (ASICs) and programmable They make excellent high-frequency ampli- continued by the manufacturers and become
gate arrays (PGA) are now routinely created fiers. Compared to MOSFET devices they unavailable for future use. ASIC technology,
and used for even limited production runs. are less susceptible to damage from elec- as mentioned earlier, brings the power of cus-
While still well beyond the reach of the indi- trostatic discharge. RF amplifiers designed tom electronics to the radio, but also makes it
vidual amateurs resources, the ASIC or PGA with bipolar transistors in their final ampli- nearly impossible to repair at the level of the
is widely used in nearly all consumer electron- fiers generally include circuitry to protect the IC, even if the problem is known. If field repair
ics and in many pieces of radio equipment. transistors from the high voltages generated and service at the component level are to be
The drawback of the ASIC and PGA is by reflections under high SWR conditions. performed, it is important to use standard ICs
that servicing and repairing the equipment Bipolar transistors and ICs, like all semi- wherever possible. Even so, when demand for
at the integrated circuit level is almost im- conductors, are susceptible to damage from a particular component drops, a manufacturer
possible for the individual without access power and lightning transients. will discontinue its production. This happens
to the manufacturers inventory of parts and There are many performance advantages on an ever-decreasing timeline.
proprietary information. Nevertheless, just to FET devices, particularly MOSFETs. The A further consideration is the trend to-
as earlier amateurs moved beyond replacing extremely low gate currents allow the design ward digital signal processing and software-
individual components to diagnosing ICs, of analog stages with nearly infinite input re- defined radio systems. (See the chapter on
todays amateurs can troubleshoot to the sistance. Signal distortion due to loading is DSP and Software Radio Design.) More
module or circuit-board level, treating them minimized in this way. FETs are less expensive and more analog functions are being per-
as components in their own right. to fabricate in ICs and so are gradually replac- formed by microprocessors and the analog
ing bipolar transistors in many IC applications. signals converted to digital at higher and
The current trend in electronics is porta- higher frequencies. There will always be
3.3.6 Comparison of bility. Transceivers are decreasing in size a need for analog circuits, but the balance
Semiconductor Devices for and in their power requirements. Integrat- point between analog and digital is shifting
Analog Applications ed circuitry has played a large part in this
towards the latter. In future years, radio and
Analog signal processing deals with trend. Extremely large circuits have been
test equipment will consist of a powerful,
changing a signal to a desired form. The three designed with microscopic proportions, in-
general-purpose digital signal processor, sur-
primary types of devices bipolar transis- cluding combinations of analog and digital
circuitry that previously required multiple rounded by the necessary analog circuitry to
tors, field-effect transistors and integrated convert the signals to digital form and supply
circuits perform similar functions, each devices. Charge-coupled devices (CCD)
imaging technology has replaced vidicon the processor with power.

3.4 Analog Systems


Many kinds of electronic equipment are such as way as to perform sequential opera- 3.4.1 Transfer Functions
developed by combining basic analog signal tions on a signal, the individual circuits are The specific way in which the analog
processing circuits, often treating them as called stages. circuit modifies the signal can be described
independent functional blocks. This section The most general way of referring to a mathematically as a transfer function. The
describes several topics associated with build- circuit is as a network. Two basic properties mathematical operation that combines a sig-
ing analog systems from multiple blocks. Al- of analog networks are of principal concern: nal with a transfer function is pictured sym-
though not all basic electronic functions are the effect that the network has on an analog bolically in Fig 3.34. The transfer function,
discussed here, the concepts associated with signal and the interaction that the network has h(t) or h(f), describes the circuits modifica-
combining them can be applied generally. with the circuitry surrounding it. Interfaces tion of the input signal in the time domain
An analog circuit can contain any number between the network and the rest of the net- where all values are functions of time, such
of discrete components (or may be imple- work are called ports. as a(t) or b(t), or in the frequency domain
mented as an IC). Since our main concern is Many analog circuits are analyzed as two- where all values are functions of frequency,
the effect that circuitry has on a signal, we port networks with an input and an output such as a(f) or b(f). The mathematical opera-
often describe the circuit by its actions rather port. The signal is fed into the input port, is tion by which h(t) operates on a(t) is called
than by its specific components. A black box modified inside the network and then exits convolution and is represented as a dot, as in
is a circuit that can be described entirely by from the output port. (See the chapter on RF a(t) h(t) = b(t). In the frequency domain,
the behavior of its interfaces with other blocks Techniques for more information on two-port the transfer function multiplies the input, as
and circuitry. When circuits are combined in networks.) in a(f) h(f) = b(f).

Analog Basics 3.27


function, the ideal cascading of analog cir- a maximum amount of power from the output
cuits results in changes produced only by the of the stage to a load connected to the output.
individual transfer functions. For any number In this case, the output impedance of the stage
of stages that are cascaded, the combination is matched or transformed to that of the load
of their transfer functions results in a new (or vice versa). This allows the stage to oper-
transfer function. The signal that enters the ate at its optimum voltage and current levels.
circuit is changed by the composite transfer In an RF amplifier, the impedance at the input
function to produce the signal that exits in the of the transmission line feeding an antenna is
cascaded circuits. transformed by means of a matching network
While each stage in a series may use feed- to produce the resistance the amplifier needs
back within itself, feedback around more than in order to efficiently produce RF power.
one stage may create a function and resul- In contrast, it is the goal of most analog
tant performance different from any of the signal processing circuitry to modify a signal
included stages. Examples include oscillation rather than to deliver large amounts of energy.
or negative feedback. Thus, an impedance-matched condition may
not be required. Instead, current between
stages can be minimized by using mismatched
3.4.3 Amplifier Frequency impedances. Ideally, if the output impedance
Response of a network is very low and the input imped-
Fig 3.34 Linear function blocks and At higher frequencies a typical amplifier ance of the following stage is very high, very
transfer functions. The transfer function little current will pass between the stages, and
acts as a low-pass filter, decreasing ampli-
can be expressed in the time domain interstage loading will be negligible.
(A) or in the frequency domain (B). The fication with increasing frequency. Signals
transfer function describes how the input within a range of frequencies are amplified
signal a(t) or a(f) is transformed into the consistently but outside that range the ampli-
output signal b(t) or b(f).
3.4.5 Noise
fication changes. At high gains many ampli-
fiers work properly only over a small range Generally we are only interested in spe-
of frequencies. The combination of gain and cific man-made signals. Nature allows many
frequency response is often expressed as a signals to combine, however, so the desired
While it is not necessary to understand gain-bandwidth product. For many ampli- signal becomes combined with many other
transfer functions mathematically to work fiers, gain times bandwidth is approximately unwanted signals, both man-made and natu-
with analog circuits, it is useful to realize that constant. As gain increases, bandwidth de- rally occurring. The broadest definition of
they describe how a signal interacts with other creases, and vice versa. noise is any signal that is not the one in which
signals in an electronic system. In general, the Performance at lower frequencies depends we are interested. One of the goals of signal
output signal of an analog system depends not on whether the amplifier is dc- or ac-coupled. processing is to separate desired signals from
only on the input signal at the same time, but Coupling refers to the transfer of signals be- noise.
also on past values of the input signal. This is tween circuits. A dc-coupled amplifier ampli- One form of noise that occurs naturally and
a very important concept and is the basis of fies signals at all frequencies down to dc. An must be dealt with in low-level processing
such essential functions as analog filtering. ac-coupled amplifier acts as a high-pass filter, circuits is called thermal noise, or Johnson
decreasing amplification as the frequency de- noise. Thermal noise is produced by random
creases toward dc. Ac-coupled circuits usu- motion of free electrons in conductors and
3.4.2 Cascading Stages ally use capacitors to allow ac signals to flow semiconductors. This motion increases as
If an analog circuit can be described with between stages while blocking the dc bias temperature increases, hence the name. This
a transfer function, a combination of analog voltages of the circuit. kind of noise is present at all frequencies
circuits can also be described similarly. This and is proportional to temperature. Natu-
description of the combined circuits depends rally occurring noise can be reduced either
upon the relationship between the transfer 3.4.4 Interstage Loading and by decreasing the circuits bandwidth or by
functions of the parts and that of the combined Impedance Matching reducing the temperature in the system. Ther-
circuits. In many cases this relationship allows Every two-port network can be further de- mal noise voltage and current vary with the
us to predict the behavior of large and complex fined by its input and output impedance. The circuit impedance and follow Ohms Law.
circuits from what we know about the parts of input impedance is the opposition to current, Low-noise-amplifier-design techniques are
which they are made. This aids in the design as a function of frequency, seen when looking based on these relationships.
and analysis of analog circuits. into the input port of the network. Likewise, Analog signal processing stages are char-
When two analog circuits are cascaded the output impedance is similarly defined acterized in part by the noise they add to a sig-
(the output signal of one stage becomes the when looking back into a network through nal. A distinction is made between enhancing
input signal to the next stage) their transfer its output port. existing noise (such as amplifying it) and
functions are combined. The mechanism of If the transfer function of a stage changes adding new noise. The noise added by analog
the combination depends on the interaction when it is cascaded with another stage, we signal processing is commonly quantified by
between the stages. The ideal case is the func- say that the second stage has loaded the first the noise factor, f. Noise factor is the ratio of
tions of the stages are completely indepen- stage. This often occurs when an appreciable the total output noise power (thermal noise
dent. In other words, when the action of a stage amount of current passes from one stage to the plus noise added by the stage) to the amplifier
is unchanged, regardless of the characteristics next. Interstage loading is related to the rela- input noise power when the termination is at
of any stages connected to its input or output. tive output impedance of a stage and the input the standard temperature of 290 K (17 C).
Just as the signal entering the first stage impedance of the stage that is cascaded after it. When the noise factor is expressed in dB, we
is modified by the action of the first transfer In some applications, the goal is to transfer often call it noise figure, NF.

3.28 Chapter 3
NF is calculated as: stage. Noise added by later stages is not mul- tween stages. An intervening stage, a type of
tiplied to the same degree and so is a smaller amplifier called a buffer, is often used for this
PNO contribution to the overall noise at the output. purpose.
NF = 10 log (13)
A PN TH Designers try to optimize system noise fac- Buffers can have high values of amplifi-
where tor by using a first stage with a minimum cation but this is unusual. A buffer used for
PNO = total noise output power, possible noise factor and maximum possible impedance transformation generally has a
A = amplification gain gain. (Caution: A circuit that overloads is of- low or unity gain. In some circuits, notably
PN TH = input thermal noise power. ten as useless as one that generates too much power amplifiers, the desired goal is to deliver
Noise factor can also be calculated as the noise.) See the RF Techniques chapter for a a maximum amount of power to the output
difference between the input and output more complete discussion on noise. Circuit device (such as a speaker or an antenna).
signal-to-noise ratios (SNR), with SNR overload is discussed in the Receivers chapter. Matching the amplifier output impedance to
expressed in dB. the output-device impedance provides maxi-
In a system of many cascaded signal pro- mum power transfer. A buffer amplifier may
3.4.6 Buffering be just the circuit for this type of application.
cessing stages, such as a communications
receiver, each stage contributes to the total It is often necessary to isolate the stages Such amplifier circuits must be carefully de-
noise of the system. The noise factor of the of an analog circuit. This isolation reduces signed to avoid distortion. Combinations of
first stage dominates the noise factor of the the loading, coupling and feedback between buffer stages can also be effective at isolat-
entire system because noise added at the first stages. It is often necessary to connect circuits ing the stages from each other and making
stage is then multiplied by each following that operate at different impedance levels be- impedance transformations, as well.

3.5 Amplifiers
By far, the most common type of analog transistors collector can be considered the device is operated close to some nominal set
circuit is the amplifier. The basic component common part of the circuit, even though in of characteristics such that current and volt-
of most electronics the transistor is actual operation, a dc voltage is applied to it. age interact fairly linearly. The large-signal
an amplifier in which a small input signal Fig 3.35 shows the three basic types of bi- model is used when the device is operated so
controls a larger signal. Transistor circuits are polar transistor amplifiers: the common-base, that it enters its saturation or cut-off regions,
designed to use the amplifying characteristics common-emitter, and common-collector. for example.
of transistors in order to create useful signal The common terminal is shown connected Different frequency ranges also require
processing functions, regardless of whether to ground, although as mentioned earlier, a dc different models. The low-frequency models
the input signal is amplified at the output. bias voltage may be present. Each type of am- used in this chapter can be used to develop
plifier is described in the following sections. circuits for dc, audio and very low RF appli-
Following the description of the amplifier, cations. At higher frequencies, small capaci-
3.5.1 Amplifier Configurations additional discussion of biasing transistors tances and inductances that can be ignored
Amplifier configurations are described by and their operation at high frequencies and at low frequencies begin to have significant
the common part of the device. The word for large signals is presented. effects on device behavior, such as gain or
common is used to describe the connection impedance. In addition, the physical structure
of a lead directly to a reference that is used by of the device also becomes significant as gain
both the input and output ports of the circuit. 3.5.2 Transistor Amplifiers begins to drop or phase shifts between input
The most common reference is ground, but Creating a useful transistor amplifier de- and output signals start to grow. In this region,
positive and negative power sources are also pends on using an appropriate model for the high-frequency models are used.
valid references. transistor itself, choosing the right configura- Amplifiers are also grouped by their oper-
The type of circuit reference used depends tion of the amplifier, using the design equa- ating class that describes the way in which
on the type of device (transistor [NPN or PNP] tions for that configuration and insuring that the input signal is amplified. There are several
or FET [P-channel or N-channel]), which lead the amplifier operates properly at different classes of analog amplifiers; A, B, AB, AB1,
is chosen as common, and the range of sig- temperatures. This section follows that se- AB2 and C.
nal levels. Once a common lead is chosen, quence, first introducing simple transistor The analog class designators specify over
the other two leads are used for signal input models and then extending that knowledge how much of the input cycle the active device
and output. Based on the biasing conditions, to the point of design guidelines for common is conducting current. A class-A amplifiers
there is only one way to select these leads. circuits that use bipolar and FETs. active device conducts current for 100 per-
Thus, there are three possible amplifier con- cent of the input signal cycle, such as shown
figurations for each type of three-lead device. DEVICE MODELS AND CLASSES in Fig 3.6. A class-B amplifier conducts
(Vacuum tube amplifiers are discussed in the Semiconductor circuit design is based on during one-half of the input cycle, class-AB,
chapter on RF Power Amplifiers.) equivalent circuits that describe the phys- AB1, and AB2 some fraction between 50
DC power sources are usually constructed ics of the devices. These circuits, made up and 100 percent of the input cycle, and class-
so that ac signals at the output terminals are of voltage and current sources and passive C for less than 50 percent of the input signal
bypassed to ground through a very low im- components such as resistors, capacitors cycle.
pedance. This allows the power source to be and inductors, are called models. A com- Digital amplifiers, in which the active
treated as an ac ground, even though it may plete model that describes a transistor exactly device is operated as a switch that is either
be supplying dc voltages to the circuit. When over a wide frequency range is a fairly com- fully-on or fully-off, similarly to switchmode
a circuit is being analyzed for its ac behavior, plex circuit. As a result, simpler models are power supplies, are also grouped by classes
ac grounds are usually treated as ground, since used in specific circumstances. For example, beginning with the letter D and beyond. Each
dc bias is ignored in the ac analysis. Thus, a the small-signal model works well when the different class uses a different method of con-

Analog Basics 3.29


Fig 3.35 The three configurations of bipolar transistor amplifiers. Each has a table of its relative impedance and current gain.
The output characteristic curve is plotted for each, with the output voltage along the x-axis, the output current along the y-axis and
various curves plotted for different values of input current. The input characteristic curve is plotted for each configuration with input
current along the x-axis, input voltage along the y-axis and various curves plotted for different values of output voltage. (A) Common
base configuration with input terminal at the emitter and output terminal at the collector. (B) Common emitter configuration with
input terminal at the base and output terminal at the collector. (C) Common collector with input terminal at the base and output
terminal at the emitter.

verting the switchs output waveform to the This section assumes the small-signal, low- rent as its input and collector circuit as its
desired RF waveform. frequency models for the transistors. output.) Current is positive if it flows into a
Amplifier classes, models and their use at device terminal.
high-frequencies are discussed in more detail SMALL-SIGNAL BJT MODEL The transistor can also be treated as a volt-
in the chapter on RF Techniques. In addition, The transistor is usually considered as a age-controlled device in which the transistors
the use of models for circuit simulation is current-controlled device in which the base emitter current, Ie, is controlled by the base-
discussed at length in the Computer-Aided current controls the collector current: emitter voltage, Vbe:
Circuit Design chapter.
I c = I b (14)
= I c I es [e(qVbe /kT) 1] I es e(qVbe /kT) (15)
3.5.3 Bipolar Transistor where where
Amplifiers Ic = collector current q = electronic charge
In this discussion, we will focus on simple Ib = base current k = Boltzmanns constant
models for bipolar transistors (BJTs). This = common-emitter current gain, beta. T = temperature in degrees Kelvin (K)
discussion is centered on NPN BJTs but ap- (The term common-emitter refers to the Ies = emitter saturation current, typically
plies equally well to PNP BJTs if the bias type of transistor circuit described below in 1 1013 A.
voltage and current polarities are reversed. which the transistor operates with base cur-

3.30 Chapter 3
The subscripts for voltages indicate the
direction of positive voltage, so that Vbe in-
dicates positive is from the base to the emit-
ter. It is simpler to design circuits using the
current-controlled device, but accounting for
the transistors behavior with temperature
requires an understanding of the voltage-
controlled model.
Transistors are usually driven by both bi-
asing and signal voltages. Equations 14 and
15 apply to both transistor dc biasing and
signal design. Both of these equations are
approximations of the more complex behav-
ior exhibited by actual transistors. Equation Fig 3.38 The hybrid-pi model for the
15 applies to a simplification of the first bipolar transistor.
Ebers-Moll model in Reference 1. More
sophisticated models for BJTs are described
by Getreu in Reference 2. Small-signal
models treat only the signal components. We THREE BASIC BJT AMPLIFIERS
will consider bias later. Fig 3.39 shows a small-signal model
The next step is to use these basic equa- applied to the three basic bipolar junction
tions to design circuits. We will begin with transistor (BJT) amplifier circuits: common-
small-signal amplifier design and the limits emitter (CE), common-base (CB) and com-
of where the techniques can be applied. Later, mon-collector (CC), more commonly known
Fig 3.36 Bipolar transistor with voltage as the emitter-follower (EF). As defined ear-
well discuss large-signal amplifier design bias and input signal.
and the distortion that arises from operating lier, the word common indicates that the
the transistor in regions where the relation- referenced terminal is part of both the input
ship between the input and output signals is and output circuits.
nonlinear. In these simple models, transistors in both
the CE and CB configurations have infinite
Common-Emitter Model output resistance because the collector current
Fig 3.36 shows a BJT amplifier connected source is in series with the output current. (The
in the common-emitter configuration. (The amplifier circuits output impedance must in-
emitter, shown connected to ground, is com- clude the effects of RL.) The transistor con-
mon to both the input circuit with the voltage nected in the EF configuration, on the other
source and the output circuit with the tran- hand, has a finite output resistance because
sistors collector.) The performance of this the current source is connected in parallel with
circuit is adequately described by equation the base circuits equivalent resistance. Cal-
14. Fig 3.37 shows the most common of all culating the EF amplifiers output resistance
transistor small-signal models, a controlled requires including the input voltage source,
current source with emitter resistance. Vs, and its impedance.
There are two variations of the model The three transistor amplifier configura-
shown in the figure. Fig 3.37B shows the tions are shown as simple circuits in Fig 3.35.
base as a direct connection to the junction Each circuit includes the basic characteristics
of a current-controlled current source (Ic = of the amplifier and characteristic curves for a
Ib) and a resistance, re, the dynamic emit- Fig 3.37 Simplified low-frequency typical transistor in each configuration. Two
ter resistance representing the change in Vbe model for the bipolar transistor, a beta sets of characteristic curves are presented:
generator with emitter resistance.
with Ie. This resistance also changes with re = 26 / le (mA dc).
one describing the input behavior and the
emitter current: other describing the output behavior in each
amplifier configuration. The different tran-
kT 26 sistor amplifier configurations have differ-
re
= (16)
qI 0 I e ent gains, input and output impedances and
This is a good approximation for most silicon phase relationships between the input and
where Ie is the dc bias current in milliamperes. transistors at low frequencies (well below output signals.
The simplified approximation only ap- the transistors gain-bandwidth product, FT) Examining the performance needs of the
plies at a typical ambient temperature of and will be used for the design examples that amplifier (engineers refer to these as the cir-
300 K because re increases with temperature. follow. cuits performance requirements) determines
In Fig 3.37A, the emitter resistance has been As frequency increases, the capacitance which of the three circuits is appropriate. Then,
moved to the base connection, where it has inherent in BJT construction becomes sig- once the amplifier configuration is chosen, the
the value (+1)re. These models are electri- nificant and the hybrid-pi model shown in equations that describe the circuits behavior
cally equivalent. Fig 3.38 is used, adding C in parallel with are used to turn the performance requirements
The transistors output resistance (the the input resistance. In this model the trans- into actual circuit component values.
Thevenin or Norton equivalent resistance fer parameter hie often represents the input This text presents design information for
between the collector and the grounded emit- impedance, shown here as a resistance at low the CE amplifier in some detail, then sum-
ter) is infinite because of the current source. frequencies. marizes designs for the CC and CB ampli-

Analog Basics 3.31


emitter, collector-base or emitter-collector).
The particular combination at which the am-
plifier is operating is its operating point. The
operating point is controlled by the selection
of component values that make up the ampli-
fier circuit so that it has the proper combina-
tion of gain, linearity and so forth. The result
is that the operating point is restricted to a
set of points that fall along a load line. The
operating point with no input signal applied
is the circuits quiescent point or Q-point. As
the input signal varies, the operating point
moves along the load line, but returns to the
Q-point when the input signal is removed.
Fig 3.40 shows the load line and Q-point
for an amplifier drawn on a transistors set
of characteristic curves for the CE amplifier Fig 3.40 A load line. A circuits load
line shows all of the possible operating
circuit. The two end-points of the load line points with the specific component
correspond to transistor saturation (ICsat on values chosen. If there is no input signal,
the IC current axis) and cutoff (VCC on the the operating point is the quiescent or
VCE voltage axis). Q-point.
When a transistor is in saturation, further
increases in base current do not cause a fur-
ther increase in collector current. In the CE
amplifier, this means that VCE is very close circuit does not operate in nonlinear regions,
to zero and IC is at a maximum. In the circuit distorting the signal as shown in Fig 3.6.
of Fig 3.35B, imagine a short circuit across If the circuit behaves differently for ac
the collector-to-emitter so that all of VCC ap- signals than for dc signals, a separate ac load
pears across RL. Increasing base current will line can be drawn as discussed below in the
not result in any additional collector current. section AC Performance for the common-
At cutoff, base current is so small that VCE emitter amplifier. For example, in the preced-
is at a maximum because no collector cur- ing circuit, if RL is replaced by a circuit that
rent is flowing and further reductions in base includes inductive or capacitive reactance,
current cause no additional increase in VCE. ac collector current will result in a different
In this simple circuit, VCE = VCC ICRL voltage drop across the circuit than will dc
and the relationship between IC and VCE is collector current. This causes the slope of
a straight line between saturation and cutoff. the ac load line to be different than that of
This is the circuits load line and it has a the dc load line.
slope of RL = (VCC VCE) / IC. No matter The ac load lines slope will also vary with
what value of base current is flowing in the frequency, although it is generally treated
transistor, the resulting combination of IC as constant over the range of frequencies
and VCE will be somewhere on the load line. for which the circuit is designed to oper-
Fig 3.39 Application of small-signal With no input signal to this simple circuit, ate. The ac and dc load lines intersect at
models for analysis of (A) the CE the transistor is at cutoff where IC = 0 and the circuits Q-point because the circuits ac
amplifier, (B) the CB and (C) the EF (CC) VCE = VCC. As the input signal increases so and dc operation is the same if the ac input
bipolar junction transistor amplifiers. that base current gets larger, the operating signal is zero.
point begins to move along the load line to the
left, so that IC increases and the voltage drop
fiers. Detailed design analysis for all three across the load, ICRL, increases, reducing
amplifiers is described in the texts listed in VCE. Eventually, the input signal will cause
the reference section for this chapter. All of enough base current to flow that saturation
the analysis in the following sections assume is reached, where VCE 0 (typically 0.1 to
the small-signal, low-frequency model and 0.3 V for silicon transistors) and IC VCC /
ignore the effects of the coupling capacitors. RL. If RL is made smaller, the load line will
High-frequency considerations are discussed become steeper and if RL increases, the load
in the RF Techniques chapter and some ad- lines slope is reduced.
vanced discussion of biasing and large signal This simple circuit cannot reproduce
behavior of BJT amplifiers is available on the negative input signals because the transis-
companion CD-ROM. tor is already in cutoff with no input signal.
In addition, the shape and spacing of the
LOAD LINES AND Q-POINT characteristic curves show that the transistor
The characteristic curves in Fig 3.35 show responds nonlinearly when close to satura-
that the transistor can operate with an infinite tion and cutoff (the nonlinear regions) than it Fig 3.41 Fixed-bias is the simplest
number of combinations of current (collec- does in the middle of the curves (the linear or common-emitter (CE) amplifier circuit.
tor, emitter and base) and voltage (collector- active region). Biasing is required so that the

3.32 Chapter 3
COMMON-EMITTER AMPLIFIER We can now determine the circuits voltage
The common-emitter amplifier (CE) is the gain, the variation in output voltage, VC,
most common amplifier configuration of all due to variations in input voltage, VB. Since
found in analog and digital circuits, from VCC is constant and is much greater than
dc through microwaves, made of discrete 1 in our model:
components and fabricated in ICs. If you
understand the CE amplifier, youve made a RC
AV (19)
good start in electronics. re
The CE amplifier is used when modest
Because re is quite small (typically a few
voltage gain is required along with an input
ohms, see equation 16), AV for this circuit
impedance (the load presented to the circuit
can be quite high.
supplying the signal to be amplified) of a few
The circuit load lines end-points are VCE =
hundred to a few k. The current gain of the
VCC and IC = VCC / RC. The circuits Q-point
CE amplifier is the transistors current gain, . Fig 3.42 Emitter degeneration. Adding
is determined by the collector resistor, RC, and RE produces negative feedback to
The simplest practical CE amplifier circuit
resistor R1 that causes bias current to flow into stabilize the bias point against changes
is shown in Fig 3.41. This circuit includes
the base. To determine the Q-point, again use due to temperature. As the bias current
both coupling and biasing components. increases, the voltage drop across RE
KVL starting at the power source and assum-
The capacitors at the input (CIN) and output also increases and causes a decrease
ing that VBE = 0.7 V for a silicon transistors
(COUT) block the flow of dc current to the load in VBE. This reduces bias current and
PN junction when forward-biased.
or to the circuit driving the amplifier. This is stabilizes the operating point.
an ac-coupled design. These capacitors also
cause the gain at very low frequencies to be VCC I BR1 =VB =VBE =0.7 V
reduced gain at dc is zero, for example, In effect, the load resistor is now split be-
so
because dc input current is blocked by CIN. tween RC and RE, with part of the output
Resistor R1 provides a path for bias current VCC 0.7V voltage appearing across each because the
to flow into the base, offsetting the collector IB = (20)
R1 changing current flows through both resistors.
current from zero and establishing the Q-point While somewhat lower than with the emitter
for the circuit. And the Q-point is therefore
connected directly to ground, voltage gain
As the input signal swings positive, more becomes easy to control because it is the ratio
current flows into the transistors base through VCEQ
= VCC I BR C (21a)
of two resistances.
CIN, causing more current to flow from the
collector to emitter as shown by equation 14. and Biasing the CE Amplifier
This causes more voltage drop across RL and Fig 3.43 adds R1 and R2 from a voltage
so the voltage at the collector also drops. The I CQ = I B (21b)
divider that controls bias current by fixing
reverse is true when the input signal swings the base voltage at:
negative. Thus, the output from the CE ampli- The actual VBE of silicon transistors will
fier is inverted from its input. vary from 0.6-0.75 V, depending on the level R2
Kirchoffs Voltage Law (KVL, see the of base current, but 0.7 V is a good compro- VB = VCC
R1 + R 2
Electrical Fundamentals chapter) is used mise value and widely used in small-signal,
to analyze the circuit. Well start with the low-frequency design. Use 0.6 V for very Since
collector circuit and treat the power supply low-power amplifiers and 0.75 V (or more) VB = VBE + (I B + I C ) R E =
as a voltage source. for high-current switch circuits.
This simple fixed-bias circuit is a good 0.7 V + ( + 1) I BR E
Vcc I c R c + Vce
= introduction to basic amplifiers, but is not
entirely practical because the bias current
We can determine the circuits voltage will change due to the change of VBE with
gain, AV, from the variation in output voltage temperature, leading to thermal instability.
caused by variations in input voltage. The In addition, the high voltage gain can lead to
output voltage from the circuit at the transis- instability due to positive feedback at high
tor collector is frequencies.
To stabilize the dc bias, Fig 3.42 adds
RE, a technique called emitter degeneration
Vc VCC I c=
= R c VCC I BR C (17)
because the extra emitter resistance creates
negative feedback: as base current rises, so
It is also necessary to determine how base does VE, the voltage drop across RE. This
current varies with input voltage. Using the reduces the base-emitter voltage and lowers
transistors equivalent circuit of Fig 3.37A, base current. The benefit of emitter degen-
VB eration comes from stabilizing the circuits
IB = dc behavior with temperature, but there is
( + 1) re Fig 3.43 Self-bias. R1 and R2 form a
a reduction in gain because of the increased voltage divider to stabilize VB and bias
so that resistance in the emitter circuit. Ignoring the current. A good rule of thumb is for current
effect of RL for the moment, flow through R1 and R2 to be 10 times the
R desired bias current. This stabilizes bias
Vc =
VCC VB C (18) RC against changes in transistor parameters
+ 1 re AV (22)
RE and component values.

Analog Basics 3.33


base current is cuits primary performance requirements,
including voltage gain, impedances, power
VB 0.7 V consumption and so on. The most common
IB = (23a)
( + 1) R E situation in which a specific voltage gain is
required and the circuits Q-point has been
and Q-point collector current becomes for
selected based on the transistor to be used,
high values of
and using the circuit of Fig 3.43, is as follows:
R2 1) Start by determining the circuits design
VCC 0.7
R1 + R 2 constraints and assumptions: power supply
I CQ =I B (23b)
RE VCC = 12 V, transistor = 150 and VBE =
0.7 V. State the circuits design requirements:
This is referred to as self-bias in which the |AV| = 5, Q-point of ICQ = 4 mA and VCEQ =
Q-point is much less sensitive to variations in 5 V. (A VCEQ 12 VCC allows a wide swing
temperature that affect and VBE. Fig 3.44 Emitter bypass. Adding CE in output voltage with the least distortion.)
A good rule-of-thumb for determining the allows ac currents to flow around RE, 2) Determine the values of RC and RE us-
sum of R1 and R2 is that the current flowing returning ac gain to the value for the ing equation 24: RC + RE = (VCC VCEQ)/
through the voltage divider, VCC/(R1+R2), fixed-bias circuit while allowing RE to
stabilize the dc operating point. ICQ = 1.75 k
should be at least 10 times the bias current,
IB. This keeps VB relatively constant even
with small changes in transistor parameters
and temperature.
Q-point VCEQ must now also account for
the voltage drop across both RC and RE,

VCEQ VCC I B (R C + R E ) (24)

More sophisticated techniques for designing


the bias networks of bipolar transistor circuits
are described in reference texts listed at the
end of this chapter.
Input and Output Impedance
With RE in the circuit, the small changes
in input current, IB, when multiplied by the
transistors current gain, , cause a large volt-
age change across RE equal to IBRE. This
is the same voltage drop as if IB was flowing
through a resistance equal to RE. Thus, the
effect of on impedance at the base is to
multiply the emitter resistance, RE by , as
well. At the transistors base,

Z B ( + 1) R E

The input source doesnt just drive the base,


of course, it also has to drive the combination
of R1 and R2, the biasing resistors. From an ac
point of view, both R1 and R2 can be consid-
ered as connected to ac ground and they can be
treated as if they were connected in parallel.
When R1//R2 are considered along with the
transistor base impedance, ZB, the impedance
Fig 3.45 Amplifier biasing and ac and dc load lines. (A) Fixed bias. Input signal is ac
presented to the input signal source is:
coupled through Ci. The output has a voltage that is equal to VCC IC RC. This signal
is ac coupled to the load, RL, through CO. For dc signals, the entire output voltage is
=Z IN R1 / / R2 / / ( + 1) R E (25) based on the value of RC. For ac signals, the output voltage is based on the value of RC
in parallel with RL. (B) Characteristic curve for the transistor amplifier pictured in (A). The
where // designates in parallel with. slope of the dc load line is equal to 1 / RC. For ac signals, the slope of the ac load line
For both versions of the CE amplifier, the is equal to 1 / (RC // RL). The quiescent-point, Q, is based on the base bias current with
no input signal applied and the point where this characteristic line crosses the dc load
collector output impedance is high enough that line. The ac load line must also pass through point Q. (C) Self-bias. Similar to fixed bias
circuit with the base bias resistor split into two: R1 connected to VCC and R2 connected
Z OUT R C (26) to ground. Also an emitter bias resistor, RE, is included to compensate for changing de-
vice characteristics. (D) This is similar to the characteristic curve plotted in (B) but with
an additional bias curve that shows how the base bias current varies as the device
CE Amplifier Design Example characteristics change with temperature. The operating point, Q, moves along this line
and the load lines continue to intersect it as it changes. If CE was added as in Fig 3.44,
The general process depends on the cir- the slope of the ac load line would increase further.

3.34 Chapter 3
3) AV = 5, so from equation 22: RC = 5 RE, Z IN R1 / / R2 / / re (29) resistor as described by equation 25,
thus 6RE = 1.75 k and RE = 270
4) Use equation 14 to determine the base and =Z IN R1 / / R2 / / ( + 1) R E (32)
bias current, IB = ICQ/ = 27 A. By the
rule of thumb, current through R1 and R2 = Z OUT R C (30) The impedance at the EF amplifiers output
10 IB = 270 A consists of the emitter resistance, RE, in paral-
5) Use equation 23 to find the voltage again neglecting the reactance of the three lel with the series combination of the internal
across R2 = VB = VBE + IC RE = 0.7 + capacitors. emitter resistance, re, the parallel combination
4 mA (0.27 k) = 1.8 V. Thus, R2 = 1.8 V / The power gain, AP, for the emitter-by- of biasing resistors R1 and R2, and the internal
270 A = 6.7 k passed CE amplifier is the ratio of output impedance of the source providing the input
6) The voltage across R1 = VCC VR2 = power, VO2/ZOUT, to input power, VI2/ZIN. signal. In this case, current gain acts to reduce
12 1.8 = 10.2 V and R1 = 10.2 V / 270 A Since VO = VIAV, the effect of the input circuits impedance on
= 37.8 k output impedance:
Use the nearest standard values (RE = R1 / / R2 / / re
270 , R1 = 39 k, R2 = 6.8 k) and circuit A P = A V2 (31) R / / R1 / / R2
RC Z OUT = S / / R E (33)
behavior will be close to that predicted. ( + 1)
AC Performance COMMON-COLLECTOR (EMITTER- In practice, with transistor of 100 or
To achieve high gains for ac signals while FOLLOWER) AMPLIFIER more, ZOUT RS/. However, if a very high
maintaining dc bias stability, the emitter-by- The common-collector (CC) amplifier in impedance source is used, such as an crystal
pass capacitor, CE, is added in Fig 3.44 to Fig 3.46 is also known as the emitter-follower microphone element or photodetector, the
provide a low impedance path for ac signals (EF) because the emitter voltage follows effects of the biasing and emitter resistors
around RE. In addition, a more accurate for- the input voltage. In fact, the amplifier has no must be considered.
mula for ac gain includes the effect of add- voltage gain (voltage gain 1), but is used as Because the voltage gain of the EF ampli-
ing RL through the dc blocking capacitor at a buffer amplifier to isolate sensitive circuits fier is unity, the power gain is simply the ratio
the collector. In this circuit, the ac voltage such as oscillators or to drive low-impedance of input impedance to output impedance,
gain is loads, such as coaxial cables. As in the CE
amplifier, the current gain of the emitter- R1 / / R2 / / ( + 1) R E
R / / RL AP (34)
AV C (27) follower is the transistors current gain, . RE
re It has relatively high input impedance with
Because of the different signal paths for ac and low output impedance and good power gain. EF Amplifier Design Example
dc signals, the ac performance of the circuit The collector of the transistor is connected
The following procedure is similar to the
is different than its dc performance. This is directly to the power supply without a resistor
design procedure in the preceding section for
illustrated in Fig 3.45 by the intersecting load and the output signal is created by the voltage
the CE amplifier, except AV = 1.
lines labeled AC Load Line and DC Load drop across the emitter resistor. There is no
1) Start by determining the circuits design
Line. The load lines intersect at the Q-point 180 phase shift as seen in the CE amplifier;
constraints and assumptions: Vcc = 12 V (the
because at that point dc performance is the the output voltage follows the input signal
power supply voltage), a transistors of 150
same as ac performance if no ac signal is with 0 phase shift because increases in the
and VBE = 0.7 V. State the circuits design
present. input signal cause increases in emitter cur-
requirements: Q-point of ICQ = 5 mA and
The equation for ac voltage gain assumes rent and the voltage drop across the emitter
VCEQ = 6 V.
that the reactances of CIN, COUT, and CE are resistor.
2) RE = (VCC VCEQ)/ICQ = 1.2 k
small enough to be neglected (less than one- The EF amplifier has high input imped-
3) Base current, IB = ICQ/ = 33 A
tenth that of the components to which they ance: following the same reasoning as for
4) Current through R1 and R2 = 10 IB =
are connected at the frequency of interest). the CE amplifier with an unbypassed emitter
330 A (10 IB rule of thumb as with the CE
At low frequencies, where the capacitor re- amplifier)
actances become increasingly large, voltage 5) Voltage across R2 = VBE + IC RE = 0.7 +
gain is reduced. Neglecting CIN and COUT, the 5 mA (1.2 k) = 6.7 V and R2 = 6.7 V / 330
low-frequency 3 dB point of the amplifier, fL, A = 20.3 k (use the standard value 22 k)
occurs where XCE = 0.414 re, 6) Voltage across R1 = VCC 6.7 V = 5.3 V
7) R1 = 5.3 V / 330 A = 16.1 k (use
2.42
fL = (28) 16 k)
2 pre C E 8) ZIN = R1 // R2 // RE( + 1) 8.5 k
This increases the emitter circuit impedance
such that AV is lowered to 0.707 of its mid- COMMON-BASE AMPLIFIER
band value, lowering gain by 3 dB. (This The common-base (CB) amplifier of
ignores the effects of CIN and COUT, which will Fig 3.47 is used where low input impedance
also affect the low-frequency performance of is needed, such as for a receiver preamp with
the circuit.) a coaxial feed line as the input signal source.
The ac input impedance of this version of Complementary to the EF amplifier, the CB
the CE amplifier is lower because the effect Fig 3.46 Emitter follower (EF) amplifier. amplifier has unity current gain and high out-
The voltage gain of the EF amplifier
of RE on ac signals is removed by the by- is unity. The amplifier has high input put impedance.
pass capacitor. This leaves only the internal impedance and low output impedance, Fig 3.47A shows the CB circuit as it is
emitter resistance, re, to be multiplied by the making it a good choice for use as a usually drawn, without the bias circuit resis-
current gain, buffer amplifier. tors connected and with the transistor symbol

Analog Basics 3.35


admittance. The reciprocal of hoe is in the
range of 100 k at low frequencies.
Voltage gain for the CB amplifier is
RC / / R L
AV (38)
re
As a result, the usual function of the CB
amplifier is to convert input current from a
low-impedance source into output voltage at
a higher impedance.
Power gain for the CB amplifier is approxi-
mately the ratio of output to input impedance,
RC
AP (39)
R E / / ( + 1) re
Fig 3.47 The common-base (CB) amplifier is often drawn in an unfamiliar style (A),
but is more easily understood when drawn similarly to the CE and EF amplifiers (B).
The input signal to the CB amplifier is applied to the emitter instead of the base. CB Amplifier Design Example
Because of its usual function as a current-
to-voltage converter, the design process for
the CB amplifier begins with selecting RE
and AV, assuming that RL is known.
Fig 3.48 A practical 1) Start by determining the circuits design
common-base (CB) constraints and assumptions: Vcc = 12 V (the
amplifier. The current gain power supply voltage), a transistors of 150
of the CB amplifier is unity. and VBE = 0.7 V. State the circuits design
It has low input impedance
and high output requirements: RE = 50 , RL = 1 k, ICQ =
impedance, resulting in 5 mA, VCEQ = 6 V.
high voltage gain. The 2) Base current, IB = ICQ/ = 33 A
CB amplifier is used to 3) Current through R1 and R2 = 10 IB =
amplify signals from low- 330 A (10 IB rule of thumb as with the CE
impedance sources, such
amplifier)
as coaxial cables.
4) Voltage across R2 = VBE + IC RE =
0.7 + 5 mA (1.2 k) = 6.7 V and R2 = 6.7 V
/ 330 A = 20.3 k (use the standard value
22 k)
5) Voltage across R1 = VCC 6.7 V = 5.3 V
turned on its side from the usual orientation iC 6) R1 = 5.3 V / 330 A = 16.1 k (use
so that the emitter faces the input. In order A= I =
iE + 1
(35)
16 k)
to better understand the amplifiers function, 7) RC = (VCC ICQ RE VCEQ) / ICQ =
Fig 3.47B reorients the circuit in a more fa- is relatively independent of input and output (12 0.25 5) / 5 mA = 1.35 k (use
miliar style. We can now clearly see that the impedance, providing excellent isolation 1.5 k)
input has just moved from the base circuit to between the input and output circuits. Output 8) AV = (1.5 k // 1 k) / (26 / IE) = 115
the emitter circuit. impedance does not affect input impedance,
Placing the input in the emitter circuit al- allowing the CB amplifier to maintain stable
lows it to cause changes in the base-emitter input impedance, even with a changing load. 3.5.4 FET Amplifiers
current as for the CE and EF amplifiers, Following reasoning similar to that for the The field-effect transistor (FET) is widely
except that for the CB amplifier a positive CE and EF amplifiers for the effect of current used in radio and RF applications. There are
change in input amplitude reduces base cur- gain on RE, we find that input impedance for many types of FETs, with JFETs (junction
rent by lowering VBE and raising VC. As a the CB amp is FET) and MOSFETs (metal-oxide-semicon-
result, the CB amplifier is noninverting, just ductor FET) being the most common types.
like the EF, with output and input signals = Z IN R E / / ( + 1) re (36) In this section we will discuss JFETs, with
in-phase. the understanding that the use of MOSFETs
A practical circuit for the CB amplifier is With high-gain transistors having a b > is similar. (This discussion is based on N-
shown in Fig 3.48. From a dc point of view 100, for typical values of re (about 1 kW) the channel JFETs, but the same discussion ap-
(replace the capacitors with open circuits), all input impedance is approximately RE. If RE plies to P-channel devices if the bias voltages
of the same resistors are there as in the CE is chosen to be 50 W, the result will be a good and currents are reversed.)
amplifier. The input capacitor, CIN, allows input impedance match to 50 W feed lines and
the dc emitter current to bypass the ac input signal sources. SMALL-SIGNAL FET MODEL
signal source and CB places the base at ac The output impedance for the CB amplifier While bipolar transistors are most com-
ground while allowing a dc voltage for bias- is approximately monly viewed as current-controlled devices,
ing. (All voltages and currents are labeled to the JFET, however, is purely a voltage-con-
1 trolled device at least at low frequencies.
aid in understanding the different orientation= Z OUT R C / / R C (37)
of the circuit.) h oe The input gate is treated as a reverse-biased di-
The CB amplifiers current gain, where hoe is the transistors collector output ode junction with virtually no current flow. As

3.36 Chapter 3
circuit (at low frequencies). This simplifies
circuit modeling considerably as biasing of
the FET gate can be done by a simple voltage
divider without having to consider the effects
of bias current flowing in the JFET itself.
The FET has characteristic curves as shown
in Fig 3.25 that are similar to those of a bipolar
transistor. The output characteristic curves
are similar to those of the bipolar transis-
tor, with the horizontal axis showing VDS
Fig 3.49 Small-signal FET model.
instead of VCE and the vertical axis showing
The FET can be modeled as a voltage- ID instead of IC. Load lines, both ac and dc,
controlled current source in its saturation can be developed and drawn on the output
region. The gate is treated as an open- characteristic curves in the same way as for
circuit due to the reverse-biased gate- bipolar transistors.
channel junction. The set of characteristic curves in Fig 3.25
are called transconductance response curves
and they show the relationship between input
with the bipolar transistor amplifier circuits, voltage (VGS), output current (ID) and out-
the circuits in this section are very basic and put voltage (VDS). The output characteristic
more thorough treatments of FET amplifier curves show ID and VDS for different values of
design can be found in the references at the VGS and are similar to the BJT output charac-
end of the chapter. teristic curve. The input characteristic curves
The operation of an N-channel JFET for show ID versus VGS for different values of VDS.
both biasing and signal amplification can be MOSFETs act in much the same way as
characterized by the following equation: JFETs when used in an amplifier. They have a Fig 3.50 In the ohmic region (A), the
2 higher input impedance, due to the insulation
1 VSG FET acts like a variable resistance, RDS,
I D = I DSS (40) between the gate and channel. The insulated with a value controlled by VGS. The alpha
VP gate also means that they can be operated with symbol (a) means is proportional to. In
the polarity of VGS such that a JFETs gate- the saturation region (B), the drain-source
where
channel of the FET can be treated like a
IDSS = drain saturation current channel junction would be forward biased, current source with ID = gmVGS.
VGS = the gate-source voltage beyond VP. Refer to the discussion of deple-
VP = the pinch-off voltage. tion- and enhancement-mode MOSFETs in
the previous section on Practical Semicon-
IDSS is the maximum current that will flow is used in the circuit of Fig 3.50B in which,
ductors.
between the drain and source for a given
value of drain-to-source voltage, VDS. Note
THREE BASIC FET AMPLIFIERS I D = g m VGS (41)
that the FET is a square-law device in which
output current is proportional to the square Just as for bipolar transistor amplifiers,
of an input voltage. (The bipolar transistors there are three basic configurations of amplifi- where gm is the FETs forward transcon-
output current is an exponential function of ers using FETs; the common-source (CS)(cor- ductance.
input current.) responding to the common-emitter), common- If VO is measured at the drain terminal
Also note that VGS in this equation has the drain (CD) or source-follower (corresponding (just as the common-emitter output voltage
opposite sense of the bipolar transistors VBE. to the emitter-follower) and the common-gate is measured at the collector), then
For this device, as VGS increases (making the (CG) (corresponding to the common base).
source more positive than the gate), drain Simple circuits and design methods are pre- VO =
g m VGSR D
current decreases until at VP the channel sented here for each, assuming low-frequency
is completely pinched-off and no drain operation and a simple, voltage-controlled The minus sign results from the output voltage
current flows at all. This equation applies current-source model for the FET. Discus- decreasing as drain current and the voltage
only if VGS is between 0 and VP. JFETs are sion of the FET amplifier at high frequencies drop across RD increases, just as in the CE
seldom used with the gate-to-channel diode is available in the RF Techniques chapter amplifier. Like the CE amplifier, the input and
forward-biased (VGS < 0). and an advanced discussion of biasing FET output voltages are thus 180 out of phase.
None of the terms in Equation 40 depend amplifiers and their large-signal behavior is Voltage gain of the CS amplifier in terms of
explicitly on temperature. Thus, the FET is contained on the companion CD-ROM. transconductance and the drain resistance is:
relatively free of the thermal instability exhib-
COMMON-SOURCE AMPLIFIER
ited by the bipolar transistor. As temperature A V = g m R D (42)
increases, the overall effect on the JFET is The basic circuit for a common-source FET
to reduce drain current and to stabilize the amplifier is shown in Fig 3.50. In the ohmic As long as VGS < 0, this simple CS am-
operation. region (see the previous discussion on FET plifiers input impedance at low frequencies
The small-signal model used for the JFET characteristics), the FET can be treated as a is that of a reverse-biased diode nearly
is shown in Fig 3.49. The drain-source chan- variable resistance as shown in Fig 3.50A infinite with a very small leakage current.
nel is treated as a current source whose output where VGS effectively varies the resistance Output impedance of the CS amplifier is ap-
is controlled by the gate-to-source voltage so between drain and source. However, most proximately RD because the FET drain-to-
that ID = gmVGS. The high input impedance FET amplifiers are designed to operate in the source channel acts like a current source with
allows the input to be modeled as an open saturation region and the model of Fig 3.49 very high impedance.

Analog Basics 3.37


Because of the high impedance of the drain- power supply voltage) and the JFET has an
source channel in the saturation region, the IDSS of 35 mA and a VP of 3.0V, typical of
output impedance of the circuit is: small-signal JFETs. State the circuits design
requirements: |AV| = 10 and IDQ = 10 mA.
Z OUT R D (47) 2) Use equation 45 to determine RS =
139
3) Since |AV| = 10, RD = 10 RS = 1390 .
Designing the Common-Source
Use standard values for RS = 150 and RD
Amplifier
= 1.5 k.
The design of the CS amplifier begins with
selection of a Q-point IDQ < IDSS. Because of AC Performance
variations in VP and IDSS from JFET to JFET, As with the CE bipolar transistor amplifier,
it may be necessary to select devices indivi a bypass capacitor can be used to increase
dually to obtain the desired performance. ac gain while leaving dc bias conditions un-
Fig 3.51 Common-source (CS) amplifier 1) Start by determining the circuits design
with self-bias. changed as shown in Fig 3.52. In the case of
constraints and assumptions: VDD = 12 V (the the CS amplifier, a source bypass capacitor is
placed across RS and the load, RL, connected
through a dc blocking capacitor. In this circuit
Z IN =
and Z OUT R D (43) voltage gain becomes:

A V = g m (R D / / R L ) (48)
As with the BJT, biasing is required to cre-
ate a Q-point for the amplifier that allows re-
Assuming CIN and COUT are large enough
production of ac signals. The practical circuit
to ignore their effects, the low-frequency
of Fig 3.50B is used to allow control of VGS
cutoff frequency of the amplifier, fL, is ap-
bias. A load line is drawn on the JFET out-
proximately where XCS = 0.707 (RD // RL),
put characteristic curves, just as for a bipolar
transistor circuit. One end point of the load 1.414
line is at VDS = VDD and the other at IDS = fL = (49)
2 p (R D / / R L ) CS
VDD / RD. The Q-point for the CS amplifier
at IDQ and VDSQ is thus determined by the as this reduces AV to 0.707 of its mid-band
dc value of VGS. Fig 3.52 Common-source amplifier with value, resulting in a 3 dB drop in output
The practical JFET CS amplifier shown source bypass capacitor, CS, to increase amplitude.
voltage gain without affecting the circuits The low-frequency ac input and output im-
in Fig 3.51 uses self-biasing in which the
dc performance. pedances of the CS amplifier remain
voltage developed across the source resistor,
RS, raises VS above ground by IDRS volts and
VGS = IDRS since there is no dc drop across =Z IN R G and Z OUT R D (50)
RG. This is also called source degeneration.
The presence of RS changes the equation of
COMMON-DRAIN (SOURCE-
voltage gain to
FOLLOWER) AMPLIFIER
gm RD R The common-drain amplifier in Fig 3.53
AV = D (44)
1 + gm RS RS is also known as a source-follower (SF) be-
cause the voltage gain of the amplifier is unity,
The value of RS is obtained by substituting similar to the emitter follower (EF) bipolar
VGS = IDRS into Equation 40 and solving for transistor amplifier. The SF amplifier is used
RS as follows: primarily as a buffer stage and to drive low-
impedance loads.
VP I Fig 3.53 Similar to the EF amplifier,
=RS 1 DQ (45) the common-drain (CD) amplifier has a
At low frequencies, the input impedance of
I DQ I DSS the SF amplifier remains nearly infinite. The
voltage gain of unity, but makes a good
buffer with high input and low output SF amplifiers output impedance is the source
Once RS is known, the equation for voltage impedances. resistance, RS, in parallel with the impedance
gain can be used to find RD.
The input impedance for the circuit of Fig
3.51 is essentially RG. Since the gate of the
JFET is often ac coupled to the input source
through a dc blocking capacitor, CIN, a value Fig 3.54 FET common-gate
(CG) amplifiers are often used
of 100 k to 1 M is often used for RG to as preamplifiers because
provide a path to ground for gate leakage cur- of their high voltage gain
rent. If RG is omitted in an ac-coupled JFET and low input impedance.
amplifier, a dc voltage can build up on the With the proper choice of
gate from leakage current or static electricity, transistor and quiescent-point
affecting the channel conductivity. current, the input impedance
can match coaxial cable
impedances directly.
Z IN = R G (46)

3.38 Chapter 3
of the controlled current source, 1/gm.
1
Z OUT = R S / /
gm
(51)
RS 1
= for g m R S >> 1
gm RS + 1 gm
Design of the SF amplifier follows essen-
tially the same process as the CS amplifier,
with RD = 0.

THE COMMON-GATE AMPLIFIER


The common-gate amplifier in Fig 3.54
has similar properties to the bipolar transistor
common-base (CB) amplifier; unity current
gain, high voltage gain, low input imped-
ance and high output impedance. (Refer to
the discussion of the CB amplifier regarding
placement of the input and how the circuit
schematic is drawn.) It is used as a voltage Fig 3.55 Common buffer stages and some typical input (ZI) and output (ZO)
amplifier, particularly for low-impedance impedances. (A) Emitter follower, made with an NPN bipolar transistor; (B) Source
sources, such as coaxial cable inputs. follower, made with an FET; and (C) Voltage follower, made with an operational
The CG amplifiers voltage gain is amplifier. All of these buffers are terminated with a load resistance, RL, and have an
output voltage that is approximately equal to the input voltage (gain 1).

A V = g m (R D / / R L ) (52)

The output impedance of the CG amplifier 6 V. State the circuits design requirements: input very closely with approximately the
is very high, we must take into account the AV = 10, RL = 1 k and RS = 50 . same voltage and little phase shift between
output resistance of the controlled current 2) Solve equation 52 for RD: 10 = 0.015 the input and output signals.
source, ro. This is analogous to the appearance RD//RL, so RD//RL = 667 . RD = 667 RL /
of hoe in the equation for output impedance of (RL 667) = 2 k.
3) IDQ is determined from equation 55: 3.5.6 Cascaded Buffers
the bipolar transistor CG amplifier.
IDQ = 10 mA. If IDQ places the Q-point in THE DARLINGTON PAIR
Z O ro (g m R S + 1) / / R D (53) the ohmic region, reduce AV and repeat the Buffer stages that are made with single ac-
calculations. tive devices can be more effective if cascaded.
The CG amplifier input impedance is ap- Two types of such buffers are in common
proximately use. The Darlington pair is a cascade of two
3.5.5 Buffer Amplifiers transistors connected as emitter followers as
1 Fig 3.55 shows common forms of buffers shown in Fig 3.56. The current gain of the
Z I = RS / / (54) with low-impedance outputs: the emitter fol- Darlington pair is the product of the current
gm
lower using a bipolar transistor, the source gains for the two transistors, 1 2.
Because the input impedance is quite low, the follower using a field-effect transistor and the What makes the Darlington pair so useful
cascode circuit described later in the section voltage follower, using an operational ampli- is that its input impedance is equal to the load
on buffers is often used to present a higher- fier. (The operational amplifier is discussed impedance times the current gain, effectively
impedance input to the signal source. later in this chapter.) These circuits are called multiplying the load impedance;
Occasionally, the value of RS must be fixed followers because the output follows the
in order to provide a specific value of input =Z I Z LOAD 1 2 (56)
impedance. Solving equation 40 for IDQ re-
sults in the following equation: For example, if a typical bipolar transistor
has = 100 and ZLOAD = 15 k, a pair of these
I DQ =

VP 2 VP
VP + VP 4R SI DSSVP
2R S2 I DSS RS
(55)

Designing the Common-Gate


Amplifier
Follow the procedure for designing a CS
amplifier, except determine the value of RD as Fig 3.56 Darlington pair made with
shown in equation 52 for voltage gain above. two emitter followers. Input impedance,
ZI, is far higher than for a single Fig 3.57 Cascode buffer made with two
1) Start by determining the circuits design transistor and output impedance, NPN bipolar transistors has a medium
constraints and assumptions: VDD = 12 V (the ZO, is nearly the same as for a single input impedance and high output
power supply voltage) and the JFET has a transistor. DC biasing has been omitted impedance. DC biasing has been omitted
gm of 15 mA/V, an IDSS of 60 mA and VP = for simplicity. for simplicity.

Analog Basics 3.39


transistors in the Darlington-pair configura-
HBK0195
tion would have:

Z=
I 15 k 100 100
= 150 M

This impedance places almost no load on


the circuit connected to the Darlington pairs
input. The shunt capacitance at the input of
real transistors can lower the actual impedance
as the frequency increases.
Drawbacks of the Darlington pair include
lower bandwidth and switching speed. The
extremely high dc gain makes biasing very
sensitive to temperature and component toler-
ances. For these reasons, the circuit is usually
used as a switch and not as a linear amplifier.

CASCODE AMPLIFIERS
A common-emitter amplifier followed by
a common-base amplifier is called a cascode
buffer, shown in its simplest form in Fig 3.57.
(Biasing and dc blocking components are
omitted for simplicity replace the tran-
sistors with the practical circuits described Fig 3.58 A pair of transistor driver circuits using a bipolar transistor and a MOSFET.
earlier.) Cascode stages using FETs follow a The input and output signals show the linear, cutoff and saturation regions.
common-source amplifier with a common-
gate configuration. The input impedance and
current gain of the cascode amplifier are ap-
proximately the same as those of the first fully-on or fully-off. The figure shows the current. Youll need to inspect the transistors
stage. The output impedance of the common- waveforms associated with both types of data sheet because decreases as collector
base or gate stage is much higher than that switch circuits. current increases, so use a value for speci-
of the common-emitter or common-source This discussion is written with power con- fied at a collector current at or above IMAX.
amplifier. The power gain of the cascode am- trol in mind, such as to drive a relay or motor or
lamp. The concepts, however, are equally ap- I MAX
plifier is the product of the input stage current IB = (58)
plicable to the much lower-power circuits that
gain and the output stage voltage gain.
As an example, a typical cascode buffer control logic-level signals. The switch should Now inspect the transistors data sheet val-
made with BJTs has moderate input imped- behave just the same switch between on ues for VCEsat and make sure that this value
ance (ZIN = 1 k), high current gain (hfe = and off quickly and completely whether of IB is sufficient to drive the transistor fully
50), and high output impedance (ZOUT = large or small. into saturation at a collector current of IMAX.
1 M). Cascode amplifiers have excellent Increase IB if necessary this is IBsat. The
DESIGNING SWITCHING CIRCUITS
input/output isolation (very low unwanted transistor must be fully saturated to minimize
internal feedback), resulting in high gain with First, select a transistor that can handle the heating when conducting load current.
good stability. Because of its excellent isola- load current and dissipate whatever power is Using the minimum value for the input volt-
tion, the cascode amplifier has little effect on dissipated as heat. Second, be sure that the age, calculate the value of RB:
external tuning components. Cascode circuits input signal source can supply an adequate
input signal to drive the transistor to the re- VIN(min) VBE
are often used in tuned amplifier designs for RB = (59)
these reasons. quired states, both on and off. Both of these I Bsat
conditions must be met to insure reliable
The minimum value of input voltage must
driver operation.
be used to accommodate the worst-case com-
3.5.7 Using the Transistor as To choose the proper transistor, the load
bination of circuit voltages and currents.
a Switch current and supply voltage must both be
Designing with a MOSFET is a little easier
known. Supply voltage may be steady, but
When designing amplifiers, the goal was because the manufacturer usually specifies
sometimes varies widely. For example, a cars
to make the transistors output a replica of its the value VGS must have for the transistor to
12 V power bus may vary from 9 to 18 V, de-
input, requiring that the transistor stay within be fully on, VGS(on). The MOSFETs gate,
pending on battery condition. The transistor
its linear region, conducting some current at being insulated from the conducting channel,
must withstand the maximum supply voltage,
all times. A switch circuit has completely dif- acts like a small capacitor of a few hundred
VMAX, when off. The load resistance, RL,
ferent properties its output current is either pF and draws very little dc current. RG in Fig
must also be known. The maximum steady-
zero or some maximum value. Fig 3.58 shows 3.58 is required if the input voltage source
state current the switch must handle is:
both a bipolar and metal-oxide semiconductor does not actively drive its output to zero volts
field-effect transistor (or MOSFET) switch VMAX when off, such as a switch connected to a
circuit. Unlike the linear amplifier circuits, I MAX = (57) positive voltage. The MOSFET wont turn
RL
there are no bias resistors in either circuit. off reliably if its gate is allowed to float.
When using the bipolar transistor as a switch, If you are using a bipolar transistor, cal- RG pulls the gate voltage to zero when the
it should operate in saturation or in cutoff. culate how much base current is required to input is open-circuited.
Similarly, an FET switch should be either drive the transistor at this level of collector Power dissipation is the next design hurdle.

3.40 Chapter 3
Even if the transistors are turned completely expense of dissipating power continuously
on, they will still dissipate some heat. Just as when the load is drawing current.
for a resistor, for a bipolar transistor switch
the power dissipation is: HIGH-SIDE AND LOW-SIDE
SWITCHING
PD V=
= CE I C VCE(sat) I MAX
The switching circuits shown in Fig 3.58
are low-side switches. This means the switch
where VCE(sat) is the collector-to-emitter is connected between the load and ground.
voltage when the transistor is saturated. A high-side switch is connected between the
Power dissipation in a MOSFET switch is: power source and the load. The same concerns
for power dissipation apply, but the methods
2 of driving the switch change because of the
PD V=
= DSI D R DS(on) I MAX (60) voltage of the emitter or source of the switch-
ing device will be at or near the power supply
RDS(on) is the resistance of the channel from voltage when the switch is on.
drain to source when the MOSFET is on. To drive an NPN bipolar transistor or an
MOSFETs are available with very low on- Fig 3.59 The snubber RC circuit at (A)
absorbs energy from transients with fast
N-channel MOSFET in a high-side circuit
resistance, but still dissipate a fair amount rise- and fall-times. At (B) a kickback requires the switch input signal to be at least
of power when driving a heavy load. The diode protects the switching device when VBE(sat) or VGS(on) above the voltage supplied
transistors data sheet will contain RDS(on) current is interrupted in the inductive to the load. If the load expects to see the full
specification and the VGS required for it to load, causing a voltage transient, by power supply voltage, the switch input signal
be reached. conducting the energy back to the power will have to be greater than the power supply
source.
Power dissipation is why a switching tran- voltage. A small step-up or boost dc-to-dc
sistor needs to be kept out of its linear region. converter is often used to supply the extra
When turned completely off or on, either cur- voltage needed for the driver circuit.
rent through the transistor or voltage across One alternate method of high-side switch-
it are low, also keeping the product of volt- transistor or provide some means of getting ing is to use a PNP bipolar transistor as the
age and current (the power to be dissipated) rid of the heat, such as heat sink. Methods of switching transistor. A small input transistor
low. As the waveform diagrams in Fig 3.58 dissipating heat are discussed in the Electri- turns the main PNP transistor on by con-
show, while in the linear region, both voltage cal Fundamentals chapter. trolling the larger transistors base current.
and current have significant values and so the Similarly, a P-channel MOSFET could also
INDUCTIVE AND CAPACITIVE be employed with a bipolar transistor or FET
transistor is generating heat when changing
LOADS acting as its driver. P-type material generally
from off to on and vice versa. Its important to
make the transition through the linear region Voltage transients for inductive loads, such does not have the same high conductivity as
quickly to keep the transistor cool. as solenoids or relays can easily reach doz- N-type material and so these devices dissipate
The worst-case amount of power dissipated ens of times the power supply voltage when somewhat more power than N-type devices
during each on-off transition is approximately load current is suddenly interrupted. To pro- under the same load conditions.
tect the transistor, the voltage transient must
1 be clamped or its energy dissipated. Where
Ptransition = VMAX I MAX 3.5.8 Choosing a Transistor
4 switching is frequent, a series-RC snubber
circuit (see Fig 3.59A) is connected across With all the choices for transistors Web
assuming that the voltage and current increase
the load to dissipate the transients energy as sites and catalogs can list hundreds select-
and decrease linearly. If the circuit turns on
heat. The most common method is to employ a ing a suitable transistor can be intimidating.
and off at a rate of f, the total average power
kickback diode that is reverse-biased when Start by determining the maximum voltage
dissipation due to switching states is:
the load is energized as shown in Fig 3.59B. (VCEO or VDS(MAX)), current (IMAX) and
f When the load current is interrupted, the diode power dissipation (PD(MAX)) the transistor
PD = VMAX I MAX (61) routes the energy back to the power supply, must handle. Determine what dc current gain,
2
clamping the voltage at the power supply volt- , or transconductance, gm, is required. Then
since there are two on-off transitions per age plus the diodes forward voltage drop. determine the highest frequency at which full
switching cycle. This power must be added Capacitive loads such as heavily filtered gain is required and multiply it by either the
to the power dissipated when the switch is power inputs may temporarily act like short voltage or current gain to obtain fT or hfe.
conducting current. circuits when the load is energized or de- This will reduce the number of choices dra-
Once you have calculated the power the energized. The surge current is only limited matically.
switch must dissipate, you must check to see by the internal resistance of the load capaci- The chapter on Component Data and Ref-
whether the transistor can withstand it. The tance. The transistor will have to handle the erences has tables of parameters for popular
manufacturer of the transistor will specify a temporary overloads without being damaged transistors that tend to be the lowest-cost and
free-air dissipation that assumes no heat-sink or overheating. The usual solution is to select most available parts, as well. You will find that
and room temperature air circulating freely a transistor with an IMAX rating greater to the a handful of part types satisfy the majority
around the transistor. This rating should be at surge current. Sometimes a small current- of your building needs. Only in very special
least 50% higher than your calculated power limiting resistor can be placed in series with applications will you need to choose a cor-
dissipation. If not, you must either use a larger the load to reduce the peak surge current at the responding special part.

Analog Basics 3.41


3.6 Operational Amplifiers
An operational amplifier, or op amp, is positive. For example, an op amp powered on the power supply voltages are coupled
one of the most useful linear devices. While from a single power supply voltage amplifies directly to the op amps internal circuitry. The
it is possible to build an op amp with dis- just as well if the circuit reference voltage op amps ability to ignore those disturbances
crete components, and early versions were, is halfway between ground and the supply is expressed by the power supply rejection
the symmetry of the circuit demanded for high voltage. ratio (PSRR). A high PSRR means that the op
performance requires a close match of many amp circuit will continue to perform well even
components. It is more effective, and much GAIN-BANDWIDTH PRODUCT AND if the power supply is imbalanced or noisy.
easier, to implement as an integrated circuit. COMPENSATION
(The term operational comes from the op An ideal op amp would have infinite fre- INPUT AND OUTPUT VOLTAGE
amps origin in analog computers where it was quency response, but just as transistors have LIMITS
used to implement mathematical operations.) an fT that marks their upper frequency limit, The op amp is capable of accepting and
The op amps performance approaches that the op amp has a gain-bandwidth product amplifying signals at levels limited by the
of an ideal analog circuit building block: an (GBW or BW). GBW represents the maxi- power supply voltages, also called rails. The
infinite input impedance (Zi), a zero output mum product of gain and frequency avail- difference in voltages between the two rails
impedance (Zo) and an open loop voltage able to any signal or circuit: voltage gain limits the range of signal voltages that can be
gain (Av) of infinity. Obviously, practical op frequency = GBW. If an op-amp with a GBW processed. The voltages can be symmetrical
amps do not meet these specifications, but of 10 MHz is connected as a 50 voltage positive and negative voltages (12 V), a posi-
they do come closer than most other types of amplifier, the maximum frequency at which tive voltage and ground, ground and a negative
amplifiers. These attributes allow the circuit that gain could be guaranteed is GBW / gain = voltage or any two different, stable voltages.
designer to implement many different func- 10 MHz / 50 = 200 kHz. GBW is an important In most op amps the signal levels that can
tions with an op amp and only a few external consideration in high-performance filters and be handled are one or two diode forward volt-
components. signal processing circuits whose design equa- age drops (0.7 V to 1.4 V) away from each
tions require high-gain at the frequencies over rail. Thus, if an op amp has 15 V connected
which they operate. as its upper rail (usually denoted V+) and
3.6.1 Characteristics of Older operational amplifiers, such as the ground connected as its lower rail (V), in-
Practical Op-Amps LM301, have an additional two connections put signals can be amplified to be as high as
An op amp has three signal terminals (see for compensation. To keep the amplifier 13.6 V and as low as 1.4 V in most amplifiers.
Fig 3.60). There are two input terminals, the from oscillating at very high gains it is often Any values that would be amplified beyond
noninverting input marked with a + sign and necessary to place a capacitor across the com- those limits are clamped (output voltages that
the inverting input marked with a sign. Volt- pensation terminals. This also decreases the should be 1.4 V or less appear as 1.4 V and
ages applied to the noninverting input cause frequency response of the op amp but increas- those that should be 13.6 V or more appear as
the op amp output voltage to change with the es its stability by making sure that the output 13.6 V). This clamping action was illustrated
same polarity. signal can not have the right phase to create in Fig 3.1.
The output of the amplifier is a single ter- positive feedback at its inputs. Most modern Rail-to-rail op amps have been devel-
minal with the output voltage referenced to the op amps are internally compensated and do oped to handle signal levels within a few tens
external circuits reference voltage. Usually, not have separate pins to add compensation of mV of rails (for example, the MAX406,
that reference is ground, but the op amps capacitance. Additional compensation can be from Maxim Integrated Products processes
internal circuitry allows all voltages to float, created by connecting a capacitor between the signals to within 10 mV of the power sup-
that is, to be referenced to any arbitrary volt- op amp output and the inverting input. ply voltages). Rail-to-rail op-amps are often
age between the op amps power supply volt- used in battery-powered products to allow the
ages. The reference can be negative, ground or CMRR AND PSRR circuits to operate from low battery voltages
One of the major advantages of using an for as long as possible.
op amp is its very high common mode rejec-
tion ratio (CMRR). Common mode signals INPUT BIAS AND OFFSET
are those that appear equally at all terminals. The inputs of an op amp, while very high
For example, if both conductors of an audio impedance, still allow some input current
cable pick up a few tenths of a volt of 60 Hz to flow. This is the input bias current and
signal from a nearby power transformer, that it is in the range of nA in modern op amps.
60 Hz signal is a common-mode signal to Slight asymmetries in the op amps internal
whatever device the cable is connected. Since circuitry result in a slight offset in the op
the op amp only responds to differences in amps out-put voltage, even with the input
voltage at its inputs, it can ignore or reject terminals shorted together. The amount of
common mode signals. CMRR is a measure voltage difference between the op amps in-
of how well the op amp rejects the common puts required to cause the output voltage to
Fig 3.60 Operational amplifier
schematic symbol. The terminal marked mode signal. High CMRR results from the be exactly zero is the input offset voltage,
with a + sign is the noninverting input. symmetry between the circuit halves. CMRR generally a few mV or less. Some op amps,
The terminal marked with a sign is is important when designing circuits that pro- such as the LM741, have special terminals
the inverting input. The output is to cess low-level signals, such as microphone to which a potentiometer can be connected
the right. On some op amps, external audio or the mV-level dc signals from sensors to null the offset by correcting the internal
compensation is needed and leads are
provided, pictured here below the device.
or thermocouples. imbalance. Introduction of a small dc cor-
Usually, the power supply leads are The rejection of power-supply imbalance rection voltage to the noninverting terminal
not shown on the op amp itself but are is also an important op amp parameter. Shifts is sometimes used to apply an offset voltage
specified in the data sheet. in power supply voltage and noise or ripple that counteracts the internal mismatch and

3.42 Chapter 3
centers the signal in the rail-to-rail range. where
DC offset is an important consideration Vo = the output voltage
in op amps for two reasons. Actual op amps Vin = the input voltage.
have a slight mismatch between the inverting
and noninverting terminals that can become The higher the op amps open-loop gain,
a substantial dc offset in the output, depend- the closer will be the voltages at the inverting
ing on the amplifier gain. The op amp output and noninverting terminals when the circuit
voltage must not be too close to the clamping is balanced and the more closely the circuits
limits or distortion will occur. closed-loop gain will equal that of Equation
63. So the negative feedback creates an elec-
A TYPICAL OP AMP tronic balancing act with the op amp increas-
As an example of typical values for these ing its output voltage so that the input error
parameters, one of todays garden-variety signal is as small as possible.
op amps, the TL084, which contains both In the inverting configuration of Fig 3.61B,
JFET and bipolar transistors, has a guaranteed the input signal (Vin) is connected through Ri
minimum CMRR of 80 dB, an input bias cur- to the inverting terminal. The feedback resis-
rent guaranteed to be below 200 pA (1 pA = tor is again connected between the inverting
1 millionth of a A) and a gain-bandwidth terminal and the output. The noninverting
product of 3 MHz. Its input offset voltage is terminal is connected to ground (or the circuit
3 mV. CMRR and PSRR are 86 dB, mean- reference voltage). In this configuration the
ing that an unwanted signal or power supply Fig 3.61 Operational amplifier circuits. feedback action results in the output voltage
imbalance of 1 V will only result in a 2.5 nV (A) Noninverting configuration. (B)
Inverting configuration.
changing to whatever value is needed such
change at the op amps output! All this for 33 that the current through Ri is balanced by an
cents even purchased in single quantities and equal and opposite current through Rf. The
there are four op-amps per package thats gain of this circuit is:
a lot of performance. Fig 3.61A, the input signal is connected to the VO R
op-amps noninverting input. The feedback = f (64)
Vin R in
3.6.2 Basic Op Amp Circuits resistor is connected between the output and
the inverting input terminal. The inverting where Vin represents the voltage input to Rin.
If a signal is connected to the input termi- input terminal is connected to Ri, which is For the remainder of this section, ground
nals of an op amp without any other circuitry connected to ground (or the circuit reference or zero voltage should be understood to
attached, it will be amplified at the devices voltage). be the circuit reference voltage. That volt-
open-loop gain (typically 200,000 for the This circuit illustrates how op amp circuits age may not be earth ground potential. For
TL084 at dc and low frequencies, or 106 dB). use negative feedback, the high open-loop example, if a single positive supply of 12 V
This will quickly saturate the output at the gain of the op amp itself, and the high input is used, 6 V may be used as the circuit refer-
power supply rails. Such large gains are rarely impedance of the op amp inputs to create a ence voltage. The circuit reference voltage
used. In most applications, negative feedback stable circuit with a fixed gain. The signal is a fixed dc voltage that can be considered
is used to limit the circuit gain by providing applied to the noninverting input causes the to be an ac ground because of the reference
a feedback path from the output terminal to output voltage of the op-amp to change with sources extremely low ac impedance.
the inverting input terminal. The resulting the same polarity. That is, a positive input The negative sign in equation 64 indicates
closed-loop gain of the circuit depends sole- signal causes a positive change in the op amps that the signal is inverted. For ac signals, in-
ly on the values of the passive components output voltage. This voltage causes current version represents a 180 phase shift. The
used to form the loop (usually resistors and, to flow in the voltage divider formed by Rf gain of the noninverting configuration can
for frequency-selective circuits, capacitors). and Ri. Because the current into the inverting vary from a minimum of 1 to the maximum
The higher the op-amps open-loop gain, the input is so low, the current through Rf is the of which the op amp is capable, as indicated
closer the circuits actual gain will approach same as Ri. by Av for dc signals, or the gain-bandwidth
that predicted from the component values. The voltage at the summing junction, the product for ac signals. The gain of the invert-
Note that the gain of the op amp itself has not connection point for the two resistors and the ing configuration can vary from a minimum of
changed it is the configuration of the ex- inverting terminal, VINV, is: 0 (gains from 0 to 1 attenuate the signal while
ternal components that determines the overall
Ri gains of 1 and higher amplify the signal) to
gain of the circuit. Some examples of differ-
VINV = VO (62) the maximum of which the device is capable.
ent circuit configurations that manipulate the Ri + Rf The inverting amplifier configuration re-
closed-loop gain follow.
The op amps output voltage will continue sults in a special condition at the op amps
INVERTING AND NONINVERTING to rise until the loop error signal, the differ- inverting input called virtual ground. Because
AMPLIFIERS ence in voltage between the inverting and the op amps high open-loop gain drives the
noninverting inputs, is close to zero. At this two inputs to be very close together, if the
The op amp is often used in either an in-
point, the voltage at the inverting terminal noninverting input is at ground potential, the
verting or a noninverting amplifier circuit as
is approximately equal to the voltage at the inverting input will be very close to ground
shown in Fig 3.61. (Inversion means that the as well and the op amps output will change
output signal is inverted from the input sig- noninverting terminal, Vin, so that VINV =
Vin. Substituting in equation 62, the gain of with the input signal to maintain the inverting
nal about the circuits voltage reference as input at ground. Measured with a voltmeter,
described below.) The amount of amplifica- this circuit is:
the input appears to be grounded, but it is
tion is determined by the two resistors: the VO R f merely maintained at ground potential by the
feedback resistor, Rf, and the input resistor, Ri. = 1 + (63)
Vin Ri action of the op amp and the feedback loop.
In the noninverting configuration shown in This point in the circuit may not be connected

Analog Basics 3.43


to any other ground connection or circuit point between individual op amps to be cancelled
because the resulting additional current flow out or dramatically reduced. In addition, the
will upset the balance of the circuit. external resistors using the same designators
The voltage follower or unity-gain buffer (R2, R3, R4) are carefully matched as well,
circuit of Fig 3.62 is commonly used as a sometimes being part of a single integrated
buffer stage. The voltage follower has the resistor pack. The result is a circuit with better
input connected directly to the noninverting performance than any single-amplifier circuit
terminal and the output connected directly over a wider temperature range.
to the inverting terminal. This configuration
has unity gain because the circuit is balanced Summing Amplifier
when the output and input voltages are the The high input impedance of an op amp
same (error voltage equals zero). It also pro- Fig 3.63 Difference amplifier. This makes it ideal for use as a summing ampli-
vides the maximum possible input impedance operational amplifier circuit amplifies the fier. In either the inverting or noninverting
and the minimum possible output impedance difference between the two input signals. configuration, the single input signal can be
of which the device is capable. replaced by multiple input signals that are
from the differential amplifier response by connected together through series resistors,
Differential and Difference Amplifier as shown in Fig 3.65. For the inverting sum-
taking into account the influence of Rn and
A differential amplifier is a special ap- Rg. If all four resistors have the same value ming amplifier, the gain of each input signal
plication of an operational amplifier (see the difference amplifier is created and VO is can be calculated individually using equation
Fig 3.63). It amplifies the difference between just the difference of the two voltages. 64 and, because of the superposition property
two analog signals and is very useful to cancel of linear circuits, the output is the sum of
noise under certain conditions. For instance, if V= each input signal multiplied by its gain. In
O Vn V1 (67)
an analog signal and a reference signal travel the noninverting configuration, the output is
over the same cable they may pick up noise, the gain times the weighted sum of the m
and it is likely that both signals will have the Instrumentation Amplifier different input signals:
same amount of noise. When the differential Just as the symmetry of the transistors mak- R p1 R p2
amplifier subtracts them, the signal will be ing up an op amp leads to a device with high Vn =Vn1 + Vn2 + ...
unchanged but the noise will be completely values of Zi, Av and CMRR and a low value R1 + R p1 R 2 + R p2
removed, within the limits of the CMRR. The of Zo, a symmetric combination of op amps
equation for differential amplifier operation is R pm
is used to further improve these parameters. + Vnm (68)
This circuit, shown in Fig 3.64 is called an R m + R pm
instrumentation amplifier. It has three parts;
Rf 1 Ri where Rpm is the parallel resistance of all m
= VO + 1 V
n V i (65) each of the two inputs is connected to a nonin-
Ri Rn + 1 Rf resistors excluding Rm. For example, with
verting buffer amplifier with a gain of 1 + R2/
R three signals being summed, Rp1 is the parallel
g R1. The outputs of these buffer amplifiers are combination of R2 and R3.
then connected to a differential amplifier with
which, if the ratios Ri/Rf and Rn/Rg are equal, a gain of R4/R3. V2 is the circuits inverting Comparators
simplifies to: input and V1 the noninverting input. A voltage comparator is another special
Rf The three amplifier modules are usually all form of op amp circuit, shown in Fig 3.66.
VO
= ( Vn V1 ) (66) part of the same integrated circuit. This means It has two analog signals as its inputs and
Ri that they have essentially the same tempera- its output is either TRUE or FALSE depend-
Note that the differential amplifier re- ture and the internal transistors and resistors ing on whether the noninverting or inverting
sponse is identical to the inverting amplifier are very well matched. This causes the subtle signal voltage is higher, respectively. Thus, it
response (equation 64) if the voltage applied gain and tracking errors caused by tempera- compares the input voltages. TRUE generally
to the noninverting terminal is equal to zero. ture differences and mismatched components corresponds to a positive output voltage and
If the voltage applied to the inverting termi-
nal (Vi) is zero, the analysis is a little more
complicated but it is possible to derive the
noninverting amplifier response (equation 62)

Fig 3.62 Voltage follower. This


operational amplifier circuit makes a
nearly ideal buffer with a voltage gain of Fig 3.64 Operational amplifiers arranged as an instrumentation amplifier. The
about one, and with extremely high input balanced and cascaded series of op amps work together to perform differential
impedance and extremely low output amplification with good common-mode rejection and very high input impedance
impedance. (no load resistor required) on both the inverting (V1) and noninverting (V2) inputs.

3.44 Chapter 3
FALSE to a negative or zero voltage. The circuit the negative clamping limit. If the compara- prevent chatter the output of the com-
in Fig 3.66 uses external resistors to generate a tor is comparing an unknown voltage to a parator switching rapidly back and forth when
reference voltage, called the setpoint, to which known voltage, the known voltage is called the input voltage is at or close to the setpoint
the input signal is compared. A comparator the setpoint and the comparator output indi- voltage. There may be noise on the input sig-
can also compare two variable voltages. cates whether the unknown voltage is above nal, as shown in Fig 3.67A, that causes the
A standard operational amplifier can be or below the setpoint. input voltage to cross the setpoint threshold
made to act as a comparator by connecting An op amp that has been intended for use repeatedly. The rapid switching of the output
the two input voltages to the noninverting and as a comparator, such as the LM311, is opti- can be confusing to the circuits monitoring
inverting inputs with no input or feedback mized to respond quickly to the input signals. the comparator output.
resistors. If the voltage of the noninverting In addition, comparators often have open- Hysteresis is a form of positive feedback
input is higher than that of the inverting in- collector outputs that use an external pull-up that moves the setpoint by a few mV in
put, the output voltage will be driven to the resistor, ROUT, connected to a positive power the direction opposite to that in which the
positive clamping limit. If the inverting input supply voltage. When the comparator output input signal crossed the setpoint threshold.
is at a higher potential than the noninverting is TRUE, the output transistor is turned off As shown in Fig 3.67B, the slight shift in the
input, the output voltage will be driven to and the pull-up resistor pulls up the output setpoint tends to hold the comparator output
voltage to the positive power supply voltage. in the new state and prevents switching back
When the comparator output is FALSE, the to the old state. Fig 3.68 shows how the output
transistor is driven into the saturation and of the comparator is fed back to the positive
the output voltage is the transistors VCE(sat). input through resistor R3, adding or subtract-
ing a small amount of current from the divider
Hysteresis and shifting the setpoint.
Comparator circuits also use hysteresis to Some applications of a voltage comparator

Setpoint, VSP

Fig 3.65 Summing operational amplifier


circuits. (A) Inverting configuration. (B)
Noninverting configuration.

Setpoint, VSP

HBK0197

Fig 3.66 A comparator circuit in which Fig 3.67 Chatter (A) is caused by noise when the input signal is close to the setpoint.
the output voltage is low when voltage Chatter can also be caused by voltage shifts that occur when a heavy load is turned on
at the inverting input is higher than the and off. Hysteresis (B) shifts the setpoint a small amount by using positive feedback in
setpoint voltage at the noninverting input. which the output pulls the setpoint farther away from the input signal after switching.

Analog Basics 3.45


are a zero crossing detector, a signal squarer ters. Active filters are discussed in the RF
(which turns other cyclical wave forms into and AF Filters chapter.
square waves) and a peak detector. An ama-
teur station application: Circuits that monitor RECTIFIERS AND PEAK
the CI-V band data output voltage from ICOM DETECTORS
HF radios use a series of comparators to sense The high open-loop gain of the op amp can
the level of the voltage and indicate on which also be used to simulate the I-V characteris-
band the radio is operating. tics of an ideal diode. A precision rectifier
circuit is shown in Fig 3.70 along with the
FILTERS I-V characteristics of a real (dashed lines)
One of the most important type of op amp and ideal (solid line) diode. The high gain of
circuits is the active filter. Two examples of the op amp compensates for the VF forward
op amp filter circuits are shown in Fig 3.69. voltage drop of the real diode in its feedback
The simple noninverting low-pass filter in loop with an output voltage equal to the input
Fig 3.68 Comparator circuit with Fig 3.69A has the same response as a passive voltage plus VF. Remember that the op amps
hysteresis. R3 causes a shift in the single-pole RC low-pass filter, but unlike the output increases until its input voltages are
comparator setpoint by allowing more passive filter, the op amp filter circuit has a balanced. When the input voltage is negative,
current to flow through R1 when the very high input impedance and a very low which would reverse-bias the diode, the op
comparator output is low. output impedance so that the filters frequency amps output cant balance the input because
and voltage response are relatively unaffected the diode blocks any current flow through
by the circuits connected to the filter input the feedback loop. The resistor at the output
and output. This circuit is a low-pass filter holds the voltage at zero until the input voltage
because the reactance of the feedback capaci- is positive once again. Precision half-wave
tor decreases with frequency, requiring less and full-wave rectifier circuits are shown in
output voltage to balance the voltages of the Fig 3.71 and their operation is described in
inverting and noninverting inputs. many reference texts.
The multiple-feedback circuit in Fig 3.69B One application of the precision rectifier
results in a band-pass response while using circuit useful in radio is the peak detector,
only resistors and capacitors. This circuit is shown in Fig 3.72. A precision rectifier is
just one of many different types of active fil- used to charge the output capacitor which

Fig 3.69 Op amp active filters. The


circuit at (A) has a low-pass response
identical to an RC filter. The 3 dB
frequency occurs when the reactance of
CF equals RF. The band-pass filter at (B) is Fig 3.70 Ideal and real diode I-V characteristics are shown at (A). The op amp
a multiple-feedback filter. precision rectifier circuit is shown at (B).

Fig 3.71 Half-wave precision rectifier (A). The extra diode at the output of the op amp prevents the op amp from saturating on
negative half-cycles and improves response time. The precision full-wave rectifier circuit at (B) reproduces both halves of the input
waveform.

3.46 Chapter 3
holds the peak voltage. The output resistor diodes current is exponential in response to
sets the time constant at which the capacitor voltage, the gain of the circuit for large input
discharges. The resistor can also be replaced signals is logarithmic.
by a transistor acting as a switch to reset the
detector. This circuit is used in AGC loops, Voltage-Current Converters
spectrum analyzers, and other instruments Another pair of useful op amp circuits
that measure the peak value of ac waveforms. convert voltage into current and current into
voltage. These are frequently used to convert
LOG AMPLIFIER currents from sensors and detectors into volt-
There are a number of applications in ra- ages that are easier to measure. Fig 3.74A
dio in which it is useful for the gain of an shows a voltage-to-current converter in which
amplifier to be higher for small input signals the output current is actually the current in the
Fig 3.72 Peak detector. Coupling a
than for large input signals. For example, an feedback loop. Because the op amps high
precision diode with a capacitor to store
charge creates a peak detector. The audio compressor circuit is used to reduce open-loop gain insures that its input voltages
capacitor will charge to the peak value the variations in a speech signals amplitude are equal, the current IR1 = VIN / R1. Certainly,
of the input voltage. R discharges the so that the average power output of an AM this could also be achieved with a resistor and
capacitor with a time constant of = RC or SSB transmitter is increased. A log am- Ohms Law, but the op amp circuits high
and can be omitted if it is desired for the plifier circuit whose gain for large signals input impedance means there is little interac-
output voltage to remain nearly constant.
is proportional to the logarithm of the input tion between the input voltage source and the
signals amplitude is shown in Fig 3.73. The output current.
log amp circuit is used in compressors and Going the other way, Fig 3.74B is a current-
limiter circuits. to-voltage converter. The op amps output
At signal levels that are too small to cause will change so that the current through the
significant current flow through the diodes, feedback resistor, R1, exactly equals the in-
the gain is set as in a regular inverting ampli- put current, keeping the inverting terminal at
fier, AV = Rf / Ri. As the signal level increas- ground potential. The output voltage, VO =
es, however, more current flows through the IIN R1. Again, this could be done with just
diode according to the Fundamental Diode a resistor, but the op amp provides isolation
Equation (equation 4) given earlier in this between the source of input current and the
chapter. That means the op amp output volt- output voltage. Fig 3.74C shows an appli-
age has to increase less (lower gain) to cause cation of a current-to-voltage converter in
enough current to flow through Ri such that which the small currents from a photodiode
the input voltages balance. The larger the are turned into voltage. This circuit can be
input voltage, the more the diode conducts used as a detector for amplitude modulated
and the lower the gain of the circuit. Since the light pulses or waveforms.
Fig 3.73 Log amplifier. At low voltages,
the gain of the circuit is Rf/Ri, but as
the diodes begin to conduct for higher-
voltage signals, the gain changes to ln
(Vin) in because of the diodes exponential
current as described in the Fundamental
Diode Equation.

Fig 3.74 Voltage-current converters. The current through R1 in (A) equals Vin/R1 because the op amp keeps both input terminals
at approximately the same voltage. At (B), input current is balanced by the op amp, resulting in VOUT = IINR1. Current through a
photodiode (C) can be converted into a voltage in this way.

Analog Basics 3.47


3.7 Analog-Digital Conversion
While radio signals are definitely analog necessary interfaces and sub-systems to per- the digital value representing the analog volt-
entities, much of the electronics associated form the entire conversion process. Schematic age will be 00, no matter whether the voltage
with radio is digital, operating on binary data symbols for ADCs and DACs are shown in is 0.0001 or 0.24999 V. The range 0.25 to
representing the radio signals and the infor- Fig 3.75. 0.5 V is represented by the digital value 01,
mation they carry. The interface between the Fig 3.76 shows two different representa- and so forth.
two worlds analog and digital is a key tions of the same physical phenomenon; an The process of converting a continuous
element of radio communications systems analog voltage changing from 0 to 1 V. In range of possible values to a limited number of
and the function is performed by analog- the analog world, the voltage is continuous discrete values is called digitization and each
digital converters. and can be represented by any real number discrete value is called a code or a quantiza-
This section presents an overview of the between 0 and 1. In the digital world, the tion code. The number of possible codes that
different types of analog-digital converters number of possible values that can represent can represent an analog quantity is 2N, where
and their key specifications and behaviors. any phenomenon is limited by the number of N is the number of bits in the code. A two-
The chapter on Digital Basics presents in- bits contained in each value. bit number can have four codes as shown in
formation on interfacing converters to digital In Fig 3.76, there are only four two-bit Fig 3.76, a four-bit number can have sixteen
circuitry and associated issues. The chapter on digital values 00, 01, 10, and 11, each cor- codes, an eight-bit number 256 codes, and
DSP and Software Radio Design discusses responding to the analog voltage being within so forth. Assuming that the smallest change
specific applications of analog-digital con- a specific range of voltages. If the analog in code values is one bit, that value is called
version technology in radio communications voltage is anywhere in the range 0 to 0.25 V, the least significant bit (LSB) regardless of
systems. its position in the format used to represent
digital numbers.
3.7.1 Basic Conversion RESOLUTION AND RANGE
Properties The resolution or step size of the conver-
Analog-digital conversion consists of tak- sion is the smallest change in the analog value
ing data in one form, such as digital binary that the conversion can represent. The range
data or an analog ac RF waveform, and creat- of the conversion is the total span of analog
ing an equivalent representation of it in the values that the conversion can process. The
opposite domain. Converters that create a maximum value in the range is called the
digital representation of analog voltages or full-scale (F.S.) value. In Fig 3.76, the con-
currents are called analog-to-digital convert- version range is 1 V. The resolution of the
ers (ADC), analog/digital converters, A/D conversion is
converters or A-to-D converters. Similarly,
range
converters that create analog voltages or cur- resolution =
rents from digital quantities are called digital- 2N
to-analog converters (DAC), digital/analog In Fig 3.76, the conversion resolution is
converters, D/A converters or D-to-A convert- 1
4 1 V = 0.25 V in the figure. If each code had
ers. The word conversion in this first section four bits instead, it would have a resolution of
on the properties of converting information 1 4
2 1 V = 0.083 V. Conversion range does
between the analog and digital domains will Fig 3.76 The analog voltage varies not necessarily have zero as one end point.
apply equally to analog-to-digital or digital- continuously between 0 and 1 V, but
the two-bit digital system only has four
For example, a conversion range of 5 V may
to-analog conversion. values to represent the analog voltage, span 0 to 5 V, 5 to 0 V, 10 to 15 V, 2.5 to
Converters are typically implemented so representation of the analog voltage is 2.5 V, and so on.
as integrated circuits that include all of the coarse. Analog-digital conversion can have a range
that is unipolar or bipolar. Unipolar means a
conversion range that is entirely positive or
negative, usually referring to voltage. Bipolar
means the range can take on both positive and
negative values.
Confusingly, the format in which the bits
are organized is also called a code. The binary
code represents digital values as binary num-
bers with the least significant bit on the right.
Binary-coded-decimal (BCD) is a code in
which groups of four bits represent individual
decimal values of 0-9. In the hexadecimal code,
groups of four bits represent decimal values of
0-15. There are many such codes. Be careful
in interpreting the word code to be sure the
correct meaning is used or understood.
To avoid having to know the conversion
Fig 3.75 Schematic symbols (A) for digital-to-analog converters (DAC) and analog-
to-digital converters (ADC). The general block diagram of a system (B) that digitizes an range to specify resolution, percentage reso-
analog signal, operates on it as digital data, then converts it back to analog form. lution is used instead.

3.48 Chapter 3
resolution ACCURACY CONVERSION RATE AND
%resolution
= 100% (69)
fullscale A companion to resolution, accuracy refers BANDWIDTH
In Fig 3.76, the conversions percentage to the ability of the converter to either assign Another important parameter of the con-
resolution is the correct code to an analog value or create version is the conversion rate or its reciprocal,
the true analog value from a specific code. As conversion speed. A code that represents an
0.25 V with resolution, it is most convenient to refer analog value at a specific time is called a
% resolution = 100% = 25% sample, so conversion rate, fS, is specified
1.0 V to accuracy as either a percentage of full scale
or in bits. Full-scale error is the maximum in samples per second (sps) and conversion
Because each code represents a range of deviation of the codes value or the analog speed as some period of time per sample,
possible analog values, the limited number quantitys value as a percentage of the full such as 1 sec/s. Because of the mechanics
of available codes creates quantization er- scale value. If a converters accuracy is given by which conversion is performed, conver-
ror. This is the maximum variation in analog as 0.02% F.S. and the conversion range is sion speed can also be specified as a number
values that can be represented by the same 5 V, the conversion can be in error by as much of cycles of clock signal used by the digital
code. In Fig 3.76, any value from 0.25 through as 0.02% 5 V = 1 mV from the correct or system performing the conversion. Conver-
0.50 V could be represented by the same expected value. Offset has the same meaning sion rate then depends on the frequency of
code: 01. The quantization error in this case is in conversion as it does in analog electron- the clock. Conversion speed may be smaller
0.25 V. Quantization error can also be speci- ics a consistent shift in the value of the than the reciprocal of conversion rate if the
fied as % full-scale by substituting the value conversion from the ideal value. system controlling the conversion introduces
of error for resolution in equation 69. Linearity error represents the maximum a delay between conversions. For example,
Resolution can also be defined by the num- deviation in step size from the ideal step size. if a conversion can take place in 1 s, but
ber of bits in the conversion. The higher the This is also called integral nonlinearity (INL). the system only performs a conversion once
number of bits, the smaller the resolution as In the converter of Fig 3.76, ideal step size is per ms, the conversion rate is 1 ksps, not the
demonstrated above. Since many converters 0.25 V. If the linearity error for the conversion reciprocal of 1 s/s = 1 Msps.
have variable ranges set by external compo- was given as 0.05% F.S., any actual step size The time accuracy of the digital clock that
nents or voltages, referring to percent resolu- could be in error by as much as 0.05% 5 V controls the conversion process can also af-
tion or as a number of bits is preferred. The fect the accuracy of the conversion. An error
= 2.5 mV. The amount of error is based on
conversions between percent resolution and in long-term frequency will cause all conver-
the full-scale value, not the step-size value.
number of bits are as follows: sions to have the same frequency error. Short-
Differential nonlinearity is a measure of how
term variations in clock period are called jitter
1 much any two adjacent step sizes deviate from
%resolution
= 100% (70) and they add noise to the conversion.
the ideal step size. Errors can be represented
2N as a number of bits, usually assumed to be
According to the Nyquist Sampling Theo-
rem, a conversion must occur at a rate twice
and least significant bits, or LSB, with one bit the highest frequency present in the analog
100% representing the same range as the conver- signal. This minimum rate is the Nyquist rate
log sion resolution.
N= %resolution (71) Accuracy and resolution are particularly
and the maximum frequency allowed in the
log2 analog signal is the Nyquist frequency. In
important in conversions for radio applica- this way, the converter bandwidth is limited
tions because they represent distortion in the to one-half the conversion rate.
Quantization error can also be specified as a
signal being processed. An analog receiver Referring to the process of converting ana-
number of least significant bits (LSB) where
that distorts the received signal creates unde- log signals to digital samples, if a lower rate
each bit is equivalent to the conversions
sired spurious signals that interfere with the is used, called undersampling, false signals
resolution.
desired signal. While the process is not ex- called aliases will be created in the digital rep-
When applied to receivers, dynamic range
actly the same, distortion created by the signal resentation of the input signal at frequencies
is more useful than percent resolution. For
conversion circuitry of a digital receiver also related to the difference between the Nyquist
each additional bit of resolution, the resolu-
causes degradation in performance. sampling rate and fS. This is called aliasing.
tion becomes two times greater, or 6.02 dB.
Sampling faster than the Nyquist rate is called
DISTORTION AND NOISE
Dynamic range (dB)
= N 6.02 dB (72) oversampling.
Distortion and noise in a conversion are Because conversions occur at some maxi-
characterized by several parameters all re- mum rate, there is always the possibility of
For example, a 16-bit conversion has a dy-
lated to linearity and accuracy. THD+N (Total signals greater than the Nyquist frequency
namic range of 16 6.02 = 96.32 dB. This is
Harmonic Distortion + Noise) is a measure of being present in an analog signal undergo-
the conversions theoretical dynamic range
how much distortion and noise is introduced ing conversion or that is being created from
with no noise present in the system. If noise
by the conversion. THD+N can be specified digital values. These signals would result in
is present, some number of the smallest codes
in percent or in dB. Smaller values are better. aliases and must be removed by band-limiting
will contain only noise, reducing the dynamic
SINAD (Signal to Noise and Distortion Ra- filters that remove them prior to conversion.
range available to represent a signal. The noise
tio) is related to THD+N, generally specified The mechanics of the sampling process are
inherent in a particular conversion process can
along with a desired signal level to show what discussed further in the chapter DSP and
be specified as an analog value (mV, A and
signal level is required to achieve a certain Software Radio Design.
so on) or as a number of bits, meaning the
level of SINAD or the highest signal level
number of codes that only represent noise.
at which a certain level of SINAD can be
For example, if a conversion has five bits of 3.7.2 Analog-to-Digital
maintained. Spurious-free Dynamic Range
noise, any value represented by the small- Converters (ADC)
(SFDR) is the difference between the am-
est five codes is considered to be noise. The
plitude of the desired signal and the highest There are a number of methods by which
conversions effective number of bits (ENOB)
unwanted signal. SFDR is generally specified the conversion from an analog quantity to a
describes the number of bits available to con-
in dB and a higher number indicates better set of digital samples can be performed. Each
tain information about the signal.
performance. has its strong points simplicity, speed,

Analog Basics 3.49


Fig 3.78 The successive-approximation
converter creates a digital word as it
varies the DAC signal in order to keep the
comparators noninverting terminal close
to the input voltage. A sample-and-hold
circuit (S/H) holds the input signal steady
while the measurement is being made.

available as quickly as the comparators can


respond and the priority encoder can create
the output code.
Flash converters are the fastest of all ADCs
(conversion speeds can be in the ns range) but
do not have high resolution because of the
number of comparators and reference volt-
ages required.

SUCCESSIVE-APPROXIMATION
CONVERTER
The successive-approximation converter
(SAC) is one of the most widely-used types
of converters. As shown in Fig 3.78, it uses
a single comparator and DAC (digital-to-
analog converter) to arrive at the value of the
input voltage by comparing it to successive
analog values generated by the DAC. This
type of converter offers a good compromise
of conversion speed and resolution.
The DAC control logic begins a conversion
by setting the output of the DAC to 12 of the
Fig 3.77 The comparators of the flash converter are always switching state conversion range. If the DAC output is greater
depending on the input signals voltage. The decoder section converts the array of
than the analog input value, the output of the
converter output to a single digital word.
comparator is 0 and the most significant bit of
the digital value is set to 0. The DAC output
then either increases or decreases by 14 of
the range, depending on whether the value of
resolution, accuracy all affect the deci- signal at its input. The flash converter uses the first comparison was 1 or 0. One test is
sion of which method to use for a particular an array of comparators that compare the made for each bit in digital output code and
application. In order to pick the right type of amplitude of the input signal to a set of refer- the result accumulated in a storage register.
ADC, it is important to decide which of these ence voltages. There is one reference voltage The process is then repeated, forming a se-
criteria most strongly affect the performance for each step. ries of approximations (thus the name of the
of your application. The outputs of the comparator array rep- converter), until a test has been made for all
resent a digital value in which each bit indi- bits in the code.
FLASH CONVERTER cates whether the input signal is greater (1) While the digital circuitry to implement
The simplest type of ADC is the flash or less than (0) the reference voltage for that the SAC is somewhat complex, it is less ex-
converter, shown in Fig 3.77, also called a comparator. A digital logic priority encoder pensive to build and calibrate than the array
direct-conversion ADC. It continually gen- then converts the array of bits into a digital of comparators and precision resistors of the
erates a digital representation of the analog output code. Each successive conversion is flash converter. Each conversion also takes

3.50 Chapter 3
a known and fixed number of clock cycles. need for external band-limiting filters on the a means of holding the input signal steady
The higher the number of bits in the output input signal. while the measurements are being made.
code, the longer the conversion takes, all other This function is performed by the circuit of
things being equal. ADC SUBSYSTEMS Fig 3.82. A high input-impedance buffer
Sample-and-Hold drives the external storage capacitor, CHOLD,
DUAL-SLOPE INTEGRATING ADCs that use a sequence of operations so that its voltage is the same as the input
CONVERTERS to create the digital output code must have signal. Another high input-impedance buffer
The dual-slope integrating ADC is shown
in Fig 3.79. It makes a conversion by measur-
ing the time it takes for a capacitor to charge
and discharge to a voltage level proportional
to the analog quantity.
A constant-current source charges a ca-
pacitor (external to the converter IC) until
the comparator output indicates that the ca-
pacitor voltage is equal to the analog input
signal. The capacitor is then discharged by the
constant-current source until the capacitor is
discharged. This process is repeated continu-
ously. The frequency of the charge-discharge
cycle is determined by the values of R and C
in Fig 3.79. The frequency is then measured
by frequency counter circuitry and converted
to the digital output code.
Dual-slope ADCs are low-cost and rela-
tively immune to noise and temperature varia-
tions. Due to the slow speed of the conversion
(measured in tens of ms) these converters are
generally only used in test instruments, such Fig 3.79 Dual-slope integrating converter. By using a constant-current source to
as multimeters. continually charge a capacitor to a known reference voltage then discharge it, the
resulting frequency is directly proportional to the resistor value.
DELTA-ENCODED CONVERTERS
Instead of charging and discharging a ca-
pacitor from 0 V to the level of the input
signal, and then back to 0 V, the delta-encoded
ADC in Fig 3.80 continually compares the
output of a DAC to the input signal using a
comparator. Whenever the signal changes,
the DAC is adjusted until its output is equal
to the input signal. Digital counter circuits
keep track of the DAC value and generate the
digital output code. Delta-encoded counters
are available with wide conversion ranges
and high resolution.

SIGMA-DELTA CONVERTERS
The sigma-delta converter also uses a Fig 3.80 Delta-encoded converter. The 1-bit DAC is operated in such a way that the
bit stream out of the comparator represents the value of the input voltage.
DAC and a comparator in a feedback loop
to generate a digital signal as shown in
Fig 3.81. An integrator stores the sum of
the input signal and the DAC output. (This
is sigma or sum in the converters name.)
The comparator output indicates whether
the integrator output is above or below the
reference voltage and that signal is used
to adjust the DACs output so that the integra-
tor output stays close to the reference voltage.
(This is the delta in the name.) The stream
of 0s and 1s from the comparator forms
a high-speed digital bit stream that is digi-
tally-filtered to form the output code.
Sigma-delta converters are used where
high resolution (16 to 24 bits) is required at Fig 3.81 Sigma-delta converter. Similar to the delta-encoded converter (Fig 3.80), the
low sampling rates of a few kHz. The digi- converter runs much faster than the output samples and uses a digital filter to derive
tal filtering in the converter also reduces the the actual output value.

Analog Basics 3.51


is used to provide a replica of the voltage on
CHOLD to the conversion circuitry.
When a conversion is started, a digital con-
trol signal opens the input switch, closes the
output switch, and the capacitors voltage is
measured by the converter. It is important that
the capacitor used for CHOLD have low leak-
age so that while the measurement is being
made, the voltage stays constant for the few
ms required. This is of particular important Fig 3.82 Sample-and-hold (S/H). An input buffer isolates the sampled voltage from
in high-precision conversion. the input signal by charging the capacitor CHOLD to that voltage with the input switch
closed and the output switch open. When a measurement is being taken, the input
Single-Ended and Differential Inputs switch is open to prevent the input signal from changing the capacitor voltage, and the
output switch is closed so that the output buffer can generate a steady voltage at its
The input of most ADCs is single-ended, in output.
which the input signal is measured between
the input pin and a common ground. Shown
in Fig 3.83A, this is acceptable for most ap-
plications, but if the voltage to be measured
is small or is the difference between two non-
zero voltages, an ADC with differential inputs Fig 3.83 Single-ended ADC
should be used as in Fig 3.83B. Differential inputs have a single active line
inputs are also useful when measuring cur- and a ground or return line (A).
rent as the voltage across a small resistor in Single-ended ADC input are often
series with the current. In that case, neither susceptible to noise and common
mode signals or any kind of
side of the resistor is likely to be at ground,
disturbance on their ground rails.
so a differential input is very useful. Differ- In (B), the differential inputs used
ential inputs also help avoid the issue of noise help the circuit ignore offsets and
contamination as discussed below. shifts in the input signal.

Input Buffering and Filtering


The input impedance of most ADCs is high
enough that the source of the input signal
is unaffected. However, to protect the ADC
input and reduce loading on the input source,
an external buffer stage can be used. Fig 3.84
shows a typical buffer arrangement with
clamping diodes to protect against electrostat-
ic discharge (ESD) and an RC-filter to prevent
RF signals from affecting the input signal. In
addition, to attenuate higher-frequency sig-
nals that might cause aliases, the input filters
can also act as band-limiting filters.

Analog and Digital Ground


By definition, ADCs straddle the analog
and digital domain. In principle, the signals
remain separate and isolated from each other.
In practice, however, voltages and currents
from the analog and digital circuitry can be
mixed together. This can result in the contami-
nation of an analog signal with components of
digital signals, and rarely, vice versa. Mostly,
this is a problem when trying to measure small
voltages in the presence of large power or
RF signals.
The usual problem is that currents from
high-speed digital circuitry find their way into
analog signal paths and create transients and
other artifacts that affect the measurement
of the analog signal. Thus, it is important to
have separate current paths for the two types
of signals. The manufacturer of the converter
Fig 3.84 Typical ADC input buffer-filter circuit. Unity-gain voltage followers help
will provide guidance for the proper use of the isolate the ADC from the input source. RC filters following the buffers act as band-
converter either in the devices data sheet or as limiting filters to prevent aliasing. Zener diodes are used to clamp the transient voltage
application notes. Look for separate pins on and route the energy of transient into the power supply system.

3.52 Chapter 3
the converter, such as AGND or DGND
that indicate how the two types of signal return
paths should be connected.

3.7.3 Digital-to-Analog
Converters (DAC)
Converting a digital value to an analog
quantity is considerably simpler than the re-
verse, but there are several issues primarily
associated with DACs that affect the selection
of a particular converter.
As each new digital value is converted to
analog, the output of the DAC makes an abrupt
step change. Even if very small, the response
of the DAC output does not respond perfectly Fig 3.85 Summing DAC. The output voltage is the inverted, weighted sum of the
inputs to each summing resistor at the input. Digital data at the input controls the
or instantaneously. Settling time is the amount current into the summing resistors and thus, the output voltage.
of time required for the DACs output to sta-
bilize within a certain amount of the final
value. It is specified by the manufacturer and
can be degraded if the load connected to the
DAC is too heavy or if it is highly reactive.
In these cases, using a buffer amplifier is rec-
ommended.
Monotonicity is another aspect of char-
acterizing the DACs accuracy. A DAC is
monotonic if in increasing the digital input
value linearly across the conversion range, the
output of the DAC increases with every step.
Because of errors in the internal conversion
circuitry, it is possible for there to be some
steps that are too small or too large, leading to
output values that seem out of order. These
are usually quite small, but if used in a preci-
sion application, monotonicity is important.

SUMMING DAC
A summing DAC, shown in Fig 3.85, is a Fig 3.86 R-2R Ladder DAC. This is the most common form of DAC because all of the
resistor values are similar, making it easier to manufacture. The similarity in resistor
summing amplifier with all of the inputs con- values also means that there will be less variation of the comparator with temperature
nected to a single reference voltage through and other effects that affect all resistors similarly.
switches. The digital value to be converted
controls which switches are closed. The larger
the digital value, the more switches are closed.
Higher current causes the summing ampli- The current output DAC functions identi- expensive trimming processes that adjust each
fiers output voltage to be higher, as well. cally to the summing DAC, but does not have resistor to the correct value and that main-
Most summing DACs have a digital signal an op amp to convert current in the digital- tain the powers-of-two relationship over wide
interface to hold the digital value between ly-controlled resistor network to voltage. It temperature variations. For DACs with high-
successive conversions, but not all, so be sure consists only of the resistor network, so an resolutions of 8 bits or more, the R-2R ladder
before you select a particular DAC. external current-to-voltage circuit (discussed DAC of Fig 3.86 is a better design.
The input resistors are binary weighted in the previous section on op amps) is required If the resistances are fairly close in value,
so that the summing network resistors to change the current to a voltage. In some the problems of manufacturing are greatly
representing the more significant bit values applications, the conversion to voltage is not reduced. By using the R-2R ladder shown in
inject more current into the op amps sum- required or it is already provided by some the figure, the same method of varying current
ming junction. Each resistor differs from other circuit. injected into an op amp circuits summing
its neighboring resistors in the amount of junction can be used with resistors of only
current it injects into the summing node by R-2R LADDER DAC two values, R and 2R. In fact, since the op
a factor of two, recreating the effect of The summing and current output DACs amp feedback resistor is also one of the IC
each digital bit inthe output voltage. At both used binary weighted resistors to convert resistors, the absolute value of the resistance
high resolutions, this becomes a problem the binary digital value into the analog output. R is unimportant, as long as the ratio of R:2R
because of the wide spread in resistor values The practical limitation of this design is the is maintained. This simplifies manufacturing
a 12-bit DAC would require a spread large difference in value between the smallest greatly and is an example of IC design being
of 2048 betweenthe largest and smallest resis- and largest resistor. For example, in a 12-bit based on ratios instead of absolute values. For
tor values. Summing DACs are generally DAC, the smallest and largest resistors dif- this reason, most DACs use the R-2R ladder
only available with low resolution for that fer by a factor of 2048 (which is 212-1). This design and the performance differences lie
reason. can be difficult to fabricate in an IC without mostly in their speed and accuracy.

Analog Basics 3.53


3.7.4 Choosing a Converter and nonlinearities. If an ADC is going to be and software can support the required data
From the point of view of performance, used for receiving applications, the spur-free rates, too!
choosing a converter, either an ADC or a dynamic range may be more important than Having established the conversion per-
DAC, comes down to resolution, accuracy high precision. (The chapter on DSP and formance requirements, the next step is to
and speed. Begin by determining the per- Software Radio Design goes into more de- consider cost, amount of associated circuitry,
cent resolution or the dynamic range of the tail about the performance requirements for power requirements, and so forth. For ex-
converter. Use the equations in the preceding ADCs used in these applications.) ample, a self-contained ADC is easier to use
section to determine the number of bits the The remaining performance criterion is and takes up less PC board space, but is not
converter must have. Select from converters the speed and rate at which the converter as flexible as one that allows the designer to
with the next highest number of bits. For can operate. Conversions should be able to use an external voltage reference to set the
example, if you determine that you need 7 bits be made at a minimum of twice the highest conversion range. Many DACs with current
of resolution, use an 8-bit converter. frequency of signal you wish to reproduce. output in their description are actually R-2R
Next, consider accuracy. If the converter is (The signal is assumed to be a sinusoid. If DACs with additional analog circuitry to pro-
needed for test instrumentation, youll need a complex waveform is to be converted, vide the current output. DACs are available
to perform an error budget on the instru- you must account for the higher frequen- with both current and voltage outputs, as well.
ments conversion processes, include errors cy signals that create the nonsinusoidal Other considerations, such as the nature of
in the analog circuitry. Once you have cal- shapes.) If the converter will be running near the required digital interface, as discussed in
culated percent errors, you can determine its maximum rate, be sure that the associ- the next section, can also affect the selection
the requirements for FS error, offset error, ated digital interface, supporting circuitry, of the converter.

3.8 Miscellaneous Analog ICs


The three main advantages of designing a Driver arrays, such as the ULN2000-series
circuit into an IC are to take advantage of the devices shown in Fig 3.87 are very useful
matched characteristics of its components, in creating an interface between low-power
to make highly complex circuitry more eco- circuits such as microprocessors and higher-
nomical, and to miniaturize the circuit and power loads and indicators. Each driver con-
reduce power consumption. As circuits stan- sists of a Darlington pair switching circuit
dardize and become widely used, they are as described earlier in this chapter. There are
often converted from discrete components different versions with different types and
to integrated circuits. Along with the op amp arrangements of resistors and diodes.
described earlier, there are many such classes Many manufacturers offer driver arrays.
of linear ICs. They are available with built-in kickback di-
odes to allow them to drive inductive loads,
such as relays, and are heavy enough to source
3.8.1 Transistor and or sink current levels up to 1 A. (All of the
Driver Arrays drivers in the array can not operate at full
The most basic form of linear integrated load at the same time, however. Read the data
circuit and one of the first to be implemented sheet carefully to determine what limitations
is the component array. The most common on current and power dissipation may exist.)
of these are the resistor, diode and transis-
tor arrays. Though capacitor arrays are also
possible, they are used less often. Component 3.8.2 Voltage Regulators
arrays usually provide space saving but this and References
is not the major advantage of these devices. One of the most popular linear ICs is the
Fig 3.87 Typical ULN2000-series driver
They are the least densely packed of the inte- array configuration and internal circuit. voltage regulator. There are two basic types,
grated circuits because each device requires The use of driver array ICs is very popular the three-terminal regulator and the regulator
a separate off-chip connection. While it may as an interface between microprocessor controller. Examples of both are described in
be possible to place over a million transistors or other low-power digital circuits and the Power Sources chapter.
on a single semiconductor chip, individual loads such as relays, solenoids or lamps. The three-terminal regulator (input,
access to these would require a total of three ground, output) is a single package designed
million pins and this is beyond the limits of to perform all of the voltage regulation func-
practicability. More commonly, resistor and tions. The output voltage can be fixed, as in
diode arrays contain from five to 16 individual obtaining this feature. The components within the 7800-series of regulators, or variable, as
devices and transistor arrays contain from an array can be internally combined for spe- in the LM317 regulator. It contains a voltage
three to six individual transistors. The advan- cial functions, such as termination resistors, reference, comparator circuits, current and
tage of these arrays is the very close matching diode bridges and Darlington pair transistors. temperature sensing protective circuits, and
of component values within the array. In a A nearly infinite number of possibilities exists the main pass element. These ICs are usu-
circuit that needs matched components, the for these combinations of components and ally contained in the same packages as power
component array is often a good method of many of these are available in arrays. transistors and the same techniques of thermal

3.54 Chapter 3
management are used to remove excess heat.
Regulator controllers, such as the popular
723 device, contain all of the control and volt-
age reference circuitry, but require external
components for the main pass element, cur-
rent sensing, and to configure some of their
control functions.
Voltage references such as the Linear Tech-
nology LT1635 are special semiconductor
diodes that have a precisely controlled I-V
characteristic. A buffer amplifier isolates the
sensitive diode and provides a low output
impedance for the voltage signal. Voltage
references are used as part of power regula-
tors and by analog-digital converter circuits.

3.8.3 Timers (Multivibrators)


A multivibrator is a circuit that oscillates
between two states, usually with a square Fig 3.88 Internal NE555 timer components. This simple array of components
wave or pulse train output. The frequency combine to make one of the most popular analog ICs. The 555 timer IC uses ratios of
of oscillation is accurately controlled with internal resistors to generate a precise voltage reference for generating time intervals
the addition of appropriate values of external based on charging and discharging a capacitor.
resistance and capacitance. The most com-
mon multivibrator in use today is the 555
timer IC (NE555 by Signetics [now Philips]
or LM555 by National Semiconductor). This
very simple eight-pin device has a frequen- ignores any other changes. An inverter makes T = 1.1 R C1 (73)
cy range from less than one hertz to several the 555 output high when Q is low and vice
hundred kilohertz. Such a device can also be versa this makes the timer circuit easier to Notice that the timing is independent of
used in monostable operation, where an input interface with external circuits. the absolute value of VCC the output pulse
pulse generates an output pulse of a different The transistor connected to Q acts as a width is the same with a 5 V supply as it is
duration, or in a stable or free-running opera- switch. When Q is high, the transistor is on and with a 15 V supply. This is because the 555
tion, where the device oscillates continuously. acts as a closed switch connected to ground. design is based on ratios and not absolute
Other applications of a multivibrator include a When Q is low, the transistor is off and the voltage levels.
frequency divider, a delay line, a pulse width switch is open. These simple building blocks
voltage divider, comparator, flip-flop and THE ASTABLE MULTIVIBRATOR
modulator and a pulse position modulator.
(These can be found in the ICs data sheet or in switch build a surprising number of use- The complement to the monostable circuit
the reference listed at the end of this chapter.) ful circuits. is the astable circuit in Fig 3.90. Pins 2, 6
Fig 3.88 shows the basic components of a and 7 are configured differently and timing
555. Connected between power input (Vcc) THE MONOSTABLE OR resistor is now split into two resistors, R1
and ground, the three resistors labeled R at ONE-SHOT TIMER and R2.
the top left of the figure form a voltage divider The simplest 555 circuit is the monostable Start from the same state as the monostable
that divides VCC into two equal stepsone circuit. This configuration will output one circuit, with C completely discharged. The
at 23 VCC and one at 13 VCC. These serve as fixed-length pulse when triggered by an input monostable circuit requires a trigger pulse to
reference voltages for the rest of the circuit. pulse. Fig 3.89 shows the connections for initiate the timing cycle. In the astable circuit,
Connected to the reference voltages are this circuit. the trigger input is connected directly to the
blocks labeled trigger comparator and Starting with capacitor C discharged, the capacitor, so if the capacitor is discharged,
threshold comparator. (Comparators were flip-flop output, Q, is high, which keeps the then the trigger comparator output must be
discussed in a preceding section.) The trig- discharge transistor turned on and the voltage high. Q is low, turning off the discharge tran-
ger comparator in the 555 is wired so that its on C below 23 VCC. The circuit is in its stable sistor, which allows C to immediately begin
output is high whenever the trigger input is state, waiting for a trigger pulse. charging.
less than 13 VCC and vice versa. Similarly, the When the voltage at the trigger input drops C charges toward VCC, but now through
threshold comparator output is high whenever below 13 VCC, the trigger comparator output the combination of R1 and R2. As the ca-
the threshold input is greater than 23 VCC. changes from low to high, which causes Q pacitor voltage passes 23 VCC, the threshold
These two outputs control a digital flip-flop to toggle to the low state. This turns off the comparator output changes from low to high,
circuit. (Flip-flops are discussed in the Digital transistor (opens the switch) and allows C to resetting Q to high. This turns on the discharge
Basics chapter.) begin charging toward VCC. transistor and the capacitor starts to discharge
The flip-flop output, Q, changes to high or When C reaches 23 VCC, this causes the through R2. When the capacitor is discharged
low when the state of its set and reset input threshold comparator to switch its output below 13 VCC, the trigger comparator changes
changes. The Q output stays high or low (it from low to high and that resets the flip-flop. from high to low and the cycle begins again,
latches or toggles) until the opposite input Q returns high, turning on the transistor and automatically. This happens over and over,
changes. When the set input changes from discharging C. The circuit has returned to its causing a train of pulses at the output while
low to high, Q goes low. When reset changes stable state. The output pulse length for the C charges and discharges between 13 and
from low to high, Q goes high. The flip-flop monostable configuration is:
2
3 VCC as seen in the figure.

Analog Basics 3.55


QS0306-HOR02 QS0306-HOR03

Fig 3.89 Monostable timer. The timing capacitor is discharged Fig 3.90 Astable timer. If the capacitor discharge process
until a trigger pulse initiates the charging process and turns initiates the next charge cycle, the timer will output a pulse
the output on. When the capacitor has charged to 2/3 VCC, the train continuously.
output is turned off, the capacitor is discharged and the timer
awaits the next trigger pulse.

The total time it takes for one complete Use of analog switches at RF through mi- ment varies from a few ohms to more than
cycle is the charge time, Tc, plus the discharge crowave frequencies requires devices spe- 100 ohms. Check the switch data sheet to
time, Td: cifically designed for those frequencies. The determine the limits for how much power
Analog Devices ADG901 is a switch usable to and current the switches can handle. Switch
T = Tc + Td = 0.693 (R1 + R 2 ) C + 0.693 R 2C 2.5 GHz. It absorbs the signal when off, acting arrays, because of the physical size of the ar-
= 0.693 (R1 + 2R 2 ) C (74) as a terminating load. The ADG902 instead ray, can have significant coupling or crosstalk
reflects the signal as an open circuit when off. between signal paths. Use caution when using
and the output frequency is: Arrays of three switches called tee-switches analog switches for high-frequency signals as
are used when very high isolation between the coupling generally increases with frequency
1 1.443 input and output is required. and may compromise the isolation required
f
= = (75)
T (R1 + 2 R 2 ) C Multiplexers or muxes are arrays of SPST for high selectivity in receivers and other RF
switches configured to act as a multiposition signal processing equipment.
When using the 555 in an application in or switch that connects one of four to sixteen
around radios, it is important to block any RF input signals to a single output. Demultiplex-
signals from the IC power supply or timing ers (demuxes) have a single input and mul- 3.8.5 Audio Output Amplifiers
control inputs. Any unwanted signal present tiple outputs. Multiplexer ICs are available as While it is possible to use op amps as low
on these inputs, especially the Control Volt- single N-to-1 switches (the MAX4617 is an power audio output drivers for headphones,
age input, will upset the timers operation 8-to-1 mux) or as groups of N-to-1 switches they generally have output impedances that
and cause it to operate improperly. The usual (the MAX4618 is a dual 4-to-1 mux). are too high for most audio transducers such as
practice is to use a 0.01 F bypass capacitor Crosspoint switch arrays are arranged so speakers and headphones. The LM380 series
(shown on pin 5 in both Fig 3.89 and 3.90) that any of four to sixteen signal inputs can of audio driver ICs has been used in radio cir-
to bypass ac signals such as noise or RF to be connected to any of four to sixteen output cuits for many years and a simple schematic
ground. Abrupt changes in VCC will also signal lines. The Analog Devices AD8108 is for a speaker driver is shown in Fig 3.91.
cause changes in timing and these may be an 8-by-8 crosspoint switch with eight inputs The popularity of personal music play-
prevented by connecting filter capacitors at and eight outputs. These arrays are used when ers has resulted in the creation of many new
the VCC input to ground. it is necessary to switch multiple signal sourc- and inexpensive audio driver ICs, such as
es among multiple signal receivers. They are the National Semiconductor LM4800- and
most commonly used in telecommunications. LM4900-series. Drivers that operate from
3.8.4 Analog Switches All analog switches use FET technology as voltages as low as 1.5 V for battery-powered
and Multiplexers the switching element. To switch ac signals, devices and up to 18 V for use in vehicles are
Arrays of analog switches, such as the most analog switches require both positive now available.
Maxim MAX312-series, allow routing of and negative voltage power supplies. An alter- When choosing an audio driver IC for
audio through lower frequency RF signals native is to use a single power supply voltage communications audio, the most important
without mechanical switches. There are and ground, but bias all inputs and output parameters to evaluate are its power re-
several types of switch arrays. Independent at one-half the power supply voltage. This quirements and power output capabilities.
switches have isolated inputs and outputs and requires dc blocking capacitors in all signal An overloaded or underpowered driver will
are turned on and off independently. Both paths, both input and output, and loading re- result in distortion. Driver ICs intended for
SPST and SPDT configurations are available. sistors may be required at the device outputs. music players have frequency responses well
Multiple switches can be wired with common The blocking capacitors can also introduce in excess of the 3000 Hz required for com-
control signals to implement multiple-pole low-frequency roll-off. munications. This can lead to annoying and
configurations. The impedance of the switching ele- fatiguing hiss unless steps are taken to reduce

3.56 Chapter 3
the circuits frequency response.
Audio power amplifiers should also be
carefully decoupled from the power supply
and the manufacturer may recommend spe-
cific circuit layouts to prevent oscillation or
feedback. Check the devices data sheet for
this information.

3.8.6 Temperature Sensors


Active temperature sensors use the temper-
ature-dependent properties of semiconductor
devices to create voltages that correspond to
absolute temperature in degrees Fahrenheit
(LM34) or degrees Celsius (LM35). These
Fig 3.91 Speaker driver. The LM380-series of audio output drivers are well-suited for
sensors (of which many others are available low-power audio outputs, such as for headphones and small speakers. When using IC
than the two examples given here) are avail- audio output drivers, be sure to refer to the manufacturers data sheet for layout and
able in small plastic packages, both leaded power supply guidelines.
and surface-mount, that respond quickly
to temperature changes. They are available
with 1% and better accuracy, requiring only
a source of voltage at very low current and
ground. Complete application information 3.8.7 Electronic Subsystems couplers, mixers, attenuators, oscillators and
is available in the manufacturer data sheets. As a particular technology becomes popu- so forth. In addition, other wireless technolo-
Thermistors, a type of passive temperature lar, a wave of integrated circuitry is developed gies such as data transmission provide oppor-
sensor, are discussed in the Electrical Fun- to service that technology and reduce its cost of tunities for manufacturers to create integrated
damentals chapter. Temperature sensors are production and service. A good example is the circuits that implement radio-related functions
used in radio mostly in cooling and thermal wireless telephony industry. IC manufactur- at low cost. Analog ICs used in the construc-
management systems. ers have developed a large number of devices tion of various radio systems and supporting
targeting this industry; receivers, transmitters, equipment are discussed in the appropriate
chapters of this book.

3.9 Analog Glossary


AC ground A circuit connection point through a semiconductor device in Cascade Placing one analog stage after
that presents a very low impedance to ac response to an applied voltage beyond another to combine their effects on the
signals. the devices ability to control or block signal.
Accuracy The ability of an analog-to- current flow. Cathode The element of an analog
digital conversion to assign the correct Base The terminal of a bipolar device that emits electrons or from
code to an analog value or create the true transistor in which control current flows. which electrons are emitted or repelled.
analog value from a specific code. Beta ( ) The dc current gain of a Characteristic curve A plot of the
Active A device that requires power to bipolar transistor, also designated hFE. relative responses of two or three
operate. Biasing The addition of a dc voltage analog-device parameters, usually of an
Active region The region in the or current to a signal at the input of an output with respect to an input. (Also
characteristic curve of an analog device analog device, changing or controlling called I-V or V-I curve.)
in which it is capable of processing the the position of the devices operating Class For analog amplifiers (Class
signal linearly. point on the characteristic curve. A, B, AB, C), a categorization of the
Amplification The process by which Bipolar transistor An analog device fraction of the input signal cycle during
amplitude of a signal is increased. Gain made by sandwiching a layer of doped which the amplifying device is active.
is the amount by which the signal is semiconductor between two layers of the For digital or switching amplifiers (Class
amplified. opposite type: PNP or NPN. D and above), a categorization of the
Analog signal A signal that can have Black box Circuit or equipment that is method by which the signal is amplified.
any amplitude (voltage or current) value analyzed only with regards to its external Clipping A nonlinearity in amplification
and exists at any point in time. behavior. in which the signals amplitude can no
Analog-to-digital converter (ADC) Bode plot Graphs showing amplitude longer be increased, usually resulting in
Circuit (usually an IC) that generates a response in dB and phase response distortion of the waveform. (Also called
digital representation of an analog signal. in degrees versus frequency on a clamping or limiting.)
Anode The element of an analog device logarithmic scale. Closed-loop gain Amplifier gain with
that accepts electrons or toward which Buffer An analog stage that prevents an external feedback circuit connected.
electrons flow. loading of one analog stage by another. Collector The terminal of a bipolar
Attenuation The process of reducing Carrier (1) Free electrons and holes in transistor from which electrons are
the amplitude of a signal. semiconductor material. (2) An unmod- removed.
Avalanche breakdown Current flow ulated component of a modulated signal.

Analog Basics 3.57


Code One possible digital value which electrons are removed. Junction FET (JFET) A field-effect
representing an analog quantity. (Also Dynamic range The range of signal transistor whose gate electrode forms a
called quantization code.) levels over which a circuit operates PN junction with the channel.
Common A terminal shared by more properly. Usually refers to the range over Linearity Processing and combining
than one port of a circuit or network. which signals are processed linearly. of analog signals independently of
Common mode Signals that appear Emitter The terminal of a bipolar trans- amplitude.
equally on all terminals of a signal port. istor into which electrons are injected. Load line A line drawn through a family
Comparator A circuit, usually an Enhancement mode An FET with of characteristic curves that shows the
amplifier, whose output indicates the a channel that does not conduct with operating points of an analog device for a
relative amplitude of two input signals. zero gate-to-source voltage and whose given load or circuit component values.
Compensation The process of conductivity is progressively increased Loading The condition that occurs
counteracting the effects of signals that as forward bias is applied. when the output behavior of a circuit is
are inadvertently fed back from the Feedback Routing a portion of an affected by the connection of another
output to the input of an analog system. output signal back to the input of a circuit to that output.
Compensation increases stability and circuit. Positive feedback causes the Low-side A switch or controlling device
prevents oscillation. input signal to be reinforced. Negative connected between a load and ground.
Compression Reducing the dynamic feedback results in partial cancellation Metal-oxide semiconductor (MOSFET)
range of a signal in order to increase the of the input signal. A field-effect transistor whose gate is
average power of the signal or prevent Field-effect transistor (FET) An analog insulated from the channel by an oxide
excessive signal levels. device with a semiconductor channel layer. (Also called insulated gate FET or
Conversion efficiency The amount whose width can be modified by an IGFET)
of light energy converted to electrical electric field. (Also called unipolar Multivibrator A circuit that oscillates
energy by a photoelectric device, transistor.) between two states.
expressed in percent. Forward bias Voltage applied across NMOS N-channel MOSFET.
Conversion rate The amount of time a PN junction in the direction to cause N-type impurity A doping atom with an
in which an analog-digital conversion current flow. excess of valence electrons that is added
can take place. Conversion speed is the Forward voltage The voltage required to semiconductor material to act as a
reciprocal of conversion rate. to cause forward current to flow through source of free electrons.
Coupling (ac or dc) The type of a PN junction. Network General name for any type of
connection between two circuits. DC Free electron An electron in a circuit.
coupling allows dc current to flow semiconductor crystal lattice that is not Noise Any unwanted signal, usually
through the connection. AC coupling bound to any atom. random in nature.
blocks dc current while allowing ac Frequency response A description of Noise figure (NF) A measure of the
current to flow. a circuits gain (or other behavior) with noise added to a signal by an analog
Cutoff frequency Frequency at which a frequency. processing stage, given in dB. (Also
circuits amplitude response is reduced Gain see Amplification. called noise factor.)
to one-half its mid-band value (also Gain-bandwidth product The Open-loop gain Gain of an amplifier
called half-power or corner frequency). relationship between amplification and with no feedback connection.
Cutoff (region) The region in the frequency that defines the limits of Operating point Values of a set of
characteristic curve of an analog device the ability of a device to act as a linear circuit parameters that specify a devices
in which there is no current through the amplifier. In many amplifiers, gain times operation at a particular time.
device. Also called the OFF region. bandwidth is approximately constant. Operational amplifier (op amp) An
Degeneration (emitter or source) Gate The control electrode of a field- integrated circuit amplifier with high
Negative feedback from the voltage effect transistor. open-loop gain, high input impedance,
drop across an emitter or source resistor High-side A switch or controlling and low output impedance.
in order to stabilize a circuits bias and device connecting between a power Optoisolator A device in which current
operating point. source and load. in a light-emitting diode controls the
Depletion mode An FET with a channel Hole A positively charged carrier that operation of a phototransistor without
that conducts current with zero gate-to- results when an electron is removed a direct electrical connection between
source voltage and whose conductivity is from an atom in a semiconductor crystal them.
progressively reduced as reverse bias is structure. Oscillator A circuit whose output varies
applied. Hysteresis In a comparator circuit, the continuously and repeatedly, usually at a
Depletion region The narrow region at practice of using positive feedback to single frequency.
a PN junction in which majority carriers shift the input setpoint in such a way as P-type impurity A doping atom with
have been removed. (Also called space- to minimize output changes when the a shortage of valence electrons that is
charge or transition region.) input signal(s) are near the setpoint. added to semiconductor material to
Digital-to-analog converter (DAC) Integrated circuit (IC) A semiconductor create an excess of holes.
Circuit (usually an IC) that creates device in which many components, such Passive A device that does not require
an analog signal from a digital as diodes, bipolar transistors, field-effect power to operate.
representation. transistors, resistors and capacitors are Peak inverse voltage (PIV) The highest
Diode A two-element semiconductor fabricated to make an entire circuit. voltage that can be tolerated by a reverse
with a cathode and an anode that Isolation Eliminating or reducing biased PN junction before current
conducts current in only one direction. electrical contact between one portion of is conducted. (See also avalanche
Drain The connection at one end of a circuit and another or between pieces breakdown.)
a field-effect-transistor channel from of equipment. Photoconductivity Phenomenon in

3.58 Chapter 3
which light affects the conductivity of Reverse bias Voltage applied across a device is operating in its active region.
semiconductor material. PN junction in the direction that does not Source The connection at one end of the
Photoelectricity Phenomenon in cause current flow. channel of a field-effect transistor into
which light causes current to flow in Reverse breakdown The condition in which electrons are injected.
semiconductor material. which reverse bias across a PN junction Stage One of a series of sequential
PMOS P-channel MOSFET. exceeds the ability of the depletion signal processing circuits or devices.
PN junction The structure that forms region to block current flow. (See also Substrate Base layer of material on
when P-type semiconductor material avalanche breakdown.) which the structure of a semiconductor
is placed in contact with N-type Roll-off Change in a circuits amplitude device is constructed.
semiconductor material. response per octave or decade of Superposition Process in which two or
Pole Frequency at which a circuits frequency. more signals are added together linearly.
transfer function becomes infinite. Safe operating area (SOA) The region Total harmonic distortion (THD)
Port A pair of terminals through which of a devices characteristic curve in A measure of how much noise and
a signal is applied to or output from a which it can operate without damage. distortion are introduced by a signal
circuit. Sample A code that represents the value processing function.
Quiescent (Q-) point Circuit or devices of an analog quantity at a specific time. Thermal runaway The condition in
operating point with no input signal Saturation (Region) The region in the which increasing device temperature
applied. (Also called bias point.) characteristic curve of an analog device increases device current in a positive
Pinch-off The condition in an FET in in which the output signal can no longer feedback cycle.
which the channel conductivity has been be increased by the input signal. See Transconductance Ratio of output
reduced to zero. Clamping. current to input voltage, with units of
Products Signals produced as the result Schottky barrier A metal-to- Siemens (S).
of a signal processing function. semiconductor junction at which a Transfer characteristics A set of
Rail Power supply voltage(s) for a depletion region is formed, similarly to a parameters that describe how a circuit
circuit. PN junction. or network behaves at and between its
Range The total span of analog values Semiconductor (1) An element such as signal interfaces.
that can be processed by an analog-to- silicon with bulk conductivity between Transfer function A mathematical
digital conversion. that of an insulator and a metal. (2) An expression of how a circuit modifies an
Recombination The process by which electronic device whose function is input signal.
free electrons and holes are combined created by a structure of chemically- Unipolar transistor see Field-effect
to produce current flow across a PN modified semiconductor materials. transistor (FET).
junction. Signal-to-noise ratio (SNR) The ratio of Virtual ground Point in a circuit
Recovery time The amount of time the strength of the desired signal to that maintained at ground potential by
required for carriers to be removed from of the unwanted signal (noise), usually the circuit without it actually being
a PN junction devices depletion region, expressed in dB. connected to ground.
halting current flow. Slew rate The maximum rate at which Zener diode A heavily-doped PN-
Rectify Convert ac to pulsating dc. a device can change the amplitude of its junction diode with a controlled reverse
Resolution Smallest change in an output. breakdown voltage, used as a voltage
analog value that can be represented in Small-signal Conditions under which reference or regulator.
a conversion between analog and digital the variations in circuit parameters due Zero Frequency at which a circuits
quantities. (Also called step size.) to the input signal are small compared transfer function becomes zero.
to the quiescent operating point and the

3.10 References and Bibliography


REFERENCES Hayward, W., Introduction to Radio Analog-Digital Conversion Handbook (by
1. Ebers, J., and Moll, J., Large-Signal Frequency Design (ARRL, 2004) the staff of Analog Devices, published
Behavior of Junction Transistors, Hayward, Campbell and Larkin, by Prentice-Hall)
Proceedings of the IRE, 42, Dec 1954, Experimental Methods in RF Design Safe-Operating Area for Power
pp 1761-1772. (ARRL, 2009) Semiconductors (ON Semi), www.
2. Getreu, I., Modeling the Bipolar Millman and Grabel, Microelectronics: onsemi.com/pub_link/Collateral/
Transistor (Elsevier, New York, 1979). Digital and Analog Circuits and Systems AN875-D.PDF
Also available from Tektronix, Inc, (McGraw-Hill, 1988) Selecting the Right CMOS Analog
Beaverton, Oregon, in paperback form. Mims, F., Timer, Op Amp & Optoelectronic Switch, Maxim Semiconductor,
Must be ordered as Part Number Circuits & Projects (Master Publishing, www.maxim-ic.com/appnotes.cfm/
062-2841-00. 2004) an_pk/638
Jung, W. IC Op Amp Cookbook (Prentice- Hyperphysics Op-Amp Circuit Tutorials,
FURTHER READING Hall, 1986) hyperphysics.phy-astr.gsu.edu/Hbase/
Alexander and Sadiku, Fundamentals of Hill and Horowitz, The Art of Electronics Electronic/opampvar.html#c2
Electric Circuits (McGraw-Hill) (Cambridge University Press, 1989)

Analog Basics 3.59