Section F2: Classes of Amplifiers
As mentioned in the introduction, an important role of the final stage of an
amplification system is to ensure a low output resistance so that the output
signal may be delivered to a low resistance load without loss of gain. Many
times, the gain stages provide the output stage with signals that are large
enough that the small-signal transistor model no longer applies (or must be
used with caution). However, it is still of paramount importance that
linearity be preserved – that is, signal distortion must be kept to an
absolute minimum. Another output stage requirement is that it delivers the
required power to the load efficiently. This means that the power dissipated
in output stage transistors must be kept to an absolute minimum, both to
ensure maximum power delivered to the load and, probably more
importantly, to ensure that the transistor temperature remain below
specified levels. Although MOS power transistors (to be discussed later this
semester) offer significant advantages over BJTs such as the reduction or
elimination of certain breakdown mechanisms found in BJTs, the reduction of
the large drive currents required for BJT power amplifiers and a generally
higher speed of operation, we will not address the MOSFET implementation
this semester due to increased the device complexity required for high
power applications.
Amplification stages are classified according to the characteristics of the
collector (output) current waveform with an applied input. These classes are
defined as Class A, Class B, Class AB and Class C. We will be
investigating each of the classifications in the remainder of this section of
our studies and will be concentrating on the discrete-circuit BJT
implementation. The remainder of this discussion will involve a brief
introduction into each class of amplifier, with subsequent discussions
investigating each in greater detail.
Class A Operation
The Class A output stage
reproduces the input signal
in its entirety as shown in
the figure to the right
(Figure 8.1 in your text).
The transistor of a Class A
amplifier conducts for the
entire cycle of the input
signal or, as your author
states, the collector current
is non-zero 100% of the
time (except for the instantaneous zero crossings). The output, ic(t), is an
amplified reproduction of the applied input ib(t). The letter designations (a)
through (e) in the figure serve to represent the same instants in time on the
ib(t) and ic(t) curves.
Class A operation may be implemented with a single transistor and is what
we studied for BJT amplifiers in Sections C and D last semester (chapters 4
and 5 of your text). Class A operation will be introduced for FET amplifiers in
Section J (chapter 6) of our studies. In the figure above, the bias point is
shown in the middle of the ac load line for maximum possible output swing.
This is not absolutely necessary for Class A operation, the only requirement
is that ICQ be greater than the amplitude of the signal current to avoid
nonlinear distortion. As a point of future discussion, remember that the
maximum power dissipated by the transistor is equal to ICQVCEQ.
Class B Operation
Class B operation requires the use of two transistor amplifiers to produce a
complete output waveform. One amplifier is used to amplify the positive half
cycle of the input signal while the second is used to amplify the negative
half-cycle. Each transistor is biased at a zero quiescent current (ICQ=0)
instead of in the active region as in a Class A stage. The advantage of this
configuration is that each transistor dissipates zero, or almost zero, power in
the quiescent condition. The Class B configuration is also known as pushpull or complementary symmetry, and it is extremely important that the
two transistors are perfectly matched so that the positive and negative
portions of the input are amplified by the same amount.
A representation of the combined characteristic curves for a Class B stage is
shown below (Figure 8.2 in your text). Note that the orientation of the
curves is such that the ac load line may be presented continuously. Also
note that the current designations in the figure below have been slightly
changed from Figure 8.2. In your text, the author refers to the currents of
transistor 1 with the subscript Q1 and the currents of transistor 2 with the
subscript Q2. This does not mean quiescent conditions, but is a common
means of identifying transistors. As we have seen (and will see much more),
multiple transistor circuits identify individual transistors by the letter Q and
sequential numbers, letters or symbols. I have attempted to avoid any
possible misunderstandings and have labeled currents simply with ‘1’ or ‘2’
as appropriate. In this figure, the letter designations are intended to convey
a time relationship between input and output – when transistor 1 is on, the
input is ib1(t) and the output is ic1(t) for the time sequence (a), (b), and
(c). Similarly, when transistor 2 is on, the input is ib2(t) and the output is
ic2(t) for the time sequence (d), (e), and (f).
A significant disadvantage
of Class B operation is
related to the advantage
mentioned above. Since the
transistors are biased at
ICQ=0, the operating range
includes
the
nonlinear
cutoff region. This results in
some
output
signal
distortion,
known
as
crossover distortion, as
illustrated in Figure 8.3 and
to the right. As may be observed, the output resembles the input sinusoid,
but is distorted near the collector current zero crossings.
Class AB Operation
In the introductory comments, I made mention of the two requirements for
an output amplification stage – linearity and efficiency. Well, from the
above discussions, we can see that the Class A configuration offers the
better linearity (smaller distortion), but has pretty lousy efficiency (power
dissipation). Conversely, the Class B stage does well with efficiency, but has
rotten linearity. So…enter the Class AB configuration!
Class AB operation still uses two transistor amplifiers, one for the positive
going portion of the input and one for the negative going input, but the bias
of the individual transistors is between the extremes of Class A and Class B.
In this strategy, each transistor is biased at a Q-point that is slightly above
the cutoff region. Biasing each transistor in this fashion does two things:
¾ the Q-point remains in the linear region of the characteristic curves,
which avoids the nonlinear distortion of the cutoff region; and
¾ each transistor conducts for an interval slightly longer than a half-cycle.
The output currents from the two transistors are combined at the load, just
like for the Class B (push-pull) configuration. The difference is that now,
during the zero crossing interval, both transistors are conducting and the
crossover distortion is reduced or eliminated.
The Class AB configuration is a compromise – the efficiency of the Class A
and the linearity of the Class B stage have been significantly (although not
perfectly) improved.
Class C Operation
For completeness, we will briefly introduce the Class C amplifier. These
amplifiers are usually employed in RF (radio-frequency) power amplification
and are capable of providing large amounts of power, but are somewhat
application specific. Briefly, a transistor in a Class C stage is biased such that
it conducts for an interval of less than a half-cycle of the input. The result is
a waveform that pulsates periodically with the period of the input signal. If
the input is a sinusoid, the output is a series of “blips” that contain the
frequency of the input as the fundamental, plus higher frequency harmonics.
The output may then be passed through an LC (inductor-capacitor) circuit
tuned to the frequency of the input sinusoid. This tuned circuit acts as a
bandpass filter and yields a sinusoid at the output that is approximately the
same frequency as the input (the narrower the passband, the closer the
approximation).
Power amplifier circuits contain transistors that are capable of handling high
power (usually defined as greater than 1 Watt) and may possess
significantly higher voltage and/or current ratings than conventional lowpower transistors. In addition, protective circuits that limit current and
techniques to dissipate heat are generally required. However, for our
purposes, we are not going to focus on these specialized devices or
techniques and will retain the small signal model for circuit analysis.