ELECTRONICS LABORATORY NOTES

Harmonic Oscillators

Objective...........................................................................................................................1
Harmonic Oscillators........................................................................................................1
Tuned Circuit' Oscillator...................................................................................................2
Experiment # 55.1.................................................................................................3
Amplitude Limiting...........................................................................................................3
Experiment # 55.2.................................................................................................3
Phase-Shift Oscillator.......................................................................................................4
Experiment # 55.3.................................................................................................4
Wien Bridge Oscillator.....................................................................................................4
Experiment # 55.4.................................................................................................5
Quadrature Oscillator........................................................................................................5
Experiment # 55.5.................................................................................................6

Objective
The purpose of this experiment is to provide introductory laboratory experience with several harmonic
oscillator circuits.

Harmonic Oscillators
One effective way of generating a sinusoidal oscillation is to generate a convenient periodic waveform and
then filter this waveform to eliminate the unwanted harmonics. The laboratory function generator, for
example, generates a square-wave output, integrates this to obtain a triangular-wave output, and then filters
the triangular wave (diode shaping) to obtain a very low distortion sinusoidal output. The focus for this
experiment however is on the generation of sinusoidal oscillations using (nearly) linear circuits.

There are three essential constituents of any harmonic oscillator:

a) An amplifier to transfer energy (usually from a DC source) into the oscillation
to replace inevitable circuit losses and energy supplied to a load, thus sustaining
the oscillation;
b) Circuit components which establish the frequency of the oscillation;
c) A mechanism to define the amplitude of oscillation.
If the circuit used were truly linear there would be no amplitude limiting mechanism other than that
associated with the average energy initially stored in the system. In a linear system signal amplitudes can
be scaled (multiplied by a constant) without affecting frequency, and will be whatever the available energy
will sustain. In a practical oscillator there is a continual infusion of energy from the amplifier and signal
amplitude grows until inherent circuit nonlinearities constrain further growth. Often a controlled
nonlinearity is introduced deliberately for this purpose, rather than depending on the unpredictable effect of
inherent nonlinearities to limit oscillation amplitude.

A productive method of studying linear oscillators is to view them as feedback amplifiers for which the
signal fedback provides the total input necessary to produce the output. This is not really a 'bootstrap'
affair; the oscillation starts because of signals generated by random electron movement (currents)
associated with thermal excitation. And, as noted before, the oscillation energy is obtained from an energy
source, typically a DC source, which is an integral part of the circuit.

Also as noted before a circuit nonlinearity, inherent or specifically introduced, must limit signal amplitude.
However a linear analysis is applicable until the signal amplitude is large enough for nonlinearity to be
significant, and so may be applied usefully to determining the conditions for the onset of oscillation. And,

Harmonic Oscillators V2 55-1 Copyright © 1996 M H Mille

provided the nonlinear limiting is not too severe, a general continuity in nature suggests the results of the
linear analysis will be 'close' in some useful sense to the actual circuit performance.

The unity loop gain condition for the onset of oscillations involves two distinct requirements. The
amplitude of the net loop gain must be 1, i.e., the signal fed back from the output must have precisely the
amplitude which causes the output to provide the signal. In addition it is necessary that the phase of the
loop gain must be 0 (or a multiple of 360°). Together these two requirements form what are called the
Barkhausen conditions for the onset of oscillations.

For a particular linear circuit to support an oscillation the network determinant must, by definition, have
conjugate complex poles on the imaginary axis. Hence to make an oscillator we must start with a circuit
which involves two poles and arrange for these poles to be placed properly on the imaginary axis. However
a circuit with only two poles is not sufficient. The root locus for a two-pole system simply does not cross
the imaginary axis whatever the circuit element values used. At least one more singularity, either a pole or a
zero, must be present as a minimal requirement. Oscillators meeting the minimal condition are studied first.
Then an illustration of a more involved system of singularities, providing certain special benefits, is studied.

A linear system may be scaled in frequency without changing the relative amplitudes of circuit voltages or
currents. Frequency is involved only as a factor in a product with either a circuit inductance or capacitance,
and only the product affects the voltage and current amplitudes. Hence the condition of unity loop gain
magnitude can be maintained while frequency is scaled arbitrarily. (Simply scale the inductances and
capacitances by the inverse of the factor the frequency is to be scaled.) It follows then that it is the phase
of the loop gain, and only the phase, which can determine the frequency of oscillation; the oscillation
frequency must be such that there is no net phase shift around the loop. The loop gain requirement, on the
other hand, determines if an oscillation can occur.

The 'stability' of the frequency of oscillation is often of special concern. 'Stability' here refers to the extent
to which a perturbation of the oscillator circuit, perhaps because of an environmental temperature change or
because of component tolerances, causes the frequency of oscillation to change. A measure of this stability
is rate of change of loop gain phase with frequency. Since the oscillation frequency is determined entirely
by the phase shift requirement the larger this rate of change the smaller the frequency change required to
compensate for a given phase shift change.

'Tuned Circuit' Oscillator

The parallel tuned circuit ( or 'tank') illustrated below
is often used in one form or another to satisfy the
phase shift condition for a sinusoidal oscillator. The
admittance Y looking into the tank is shown to the
right of the tuned circuit.

An oscillator circuit using the tank is shown in Fig.

55-2. The amplifier on the left provides variable
output voltage, of which a fraction G/Y is passed on
to the voltage follower and returned to drive the
amplifier; Y is the admittance of the parallel
combination of G, L, and C.

Assuming the amplifier does not contribute any

phase shift, i.e., simply provides a voltage gain A, the
Barkhausen requirements are AG/Y = 1

Note that Y contributes two poles and a zero, to provide the minimum requirement for oscillation to be
achieved. The conditions are satisfied at a (radian) frequency ωo, and the (theoretical) gain requirement is
simply A ≥ 1.

Harmonic Oscillators V2 55-2 Copyright © 1996 M H Mille

The rate of change of phase of the tank at its resonent frequency is:

Values for the tank Q vary from perhaps 10 or so when discrete components are used, to several thousand
for a crystal equivalent circuit. If, for example, the amplifier introduced some phase shift the resonant
circuit would have to provide a compensating change in its phase shift to maintain the net phase shift zero.
The higher the Q the less the change in oscillation frequency needed to correct for such a parasitic phase
perturbation.

Experiment # 55.1: Tuned Circuit Oscillator

Assemble a tuned-circuit oscillator using nominal values of G = 0.1 millimho, L =10 millihenry, and C =
0.03 microfarad. A 5 kΩ potentiometer provides a gain adjustment for the noninverting amplifier. The
diode branch is used for the next experiment; it should be left disconnected for the present. Adjust the
potentiometer wiper so that oscillations just start, and concurrently observe the signal at the output of each
amplifier. Measure the signal frequency and compare to the calculated tank resonent frequency. Observe
that one of the signals is considerably more distorted than the other. Explain this difference in distortion
between the signals. Hint: Note that the 10 kΩ resistor in series with the parallel LC branches forms a
frequency-sensitive 'low-pass' voltage divider.

Where (approximately) is the dominant pole for the 741 amplifier you are using? What does this imply
about the phase shift introduced by the opamp, which is assumed to be zero in the analysis above? What is
the effect expected on the frequency of oscillation? Increase the amplifier gain (i.e., less feedback and so
more phase shift) until the circuit drops out of oscillation. (Parasitic capacitances may keep the circuit
oscillating after a fashion, but with a heavily distorted waveform indicating the intended oscillation has
ceased.) Why does increasing the gain cause the oscillations to stop?

Amplitude Limiting
The nonlinear mechanism limiting oscillation amplitude in the test oscillator circuit is saturation of the
amplifier when the signal amplitude is large enough. This produces a distorted waveform as the amplifier
output. Less distortion may be obtained by adding a controlled nonlinearity, 'soft' limiting signal amplitude
to a smaller value. The nonlinearity is provided by the paralleled diodes; when the oscillation amplitude
becomes large enough the diodes conduct at the peaks of the cycle modifying the feedback (gain) of the
amplifier. The 10 kΩ potentiometer can be adjusted so that the oscillation gain requirement will not be met
at the very peaks of the cycle. The oscillation stops briefly at the waveform peaks, limiting signal growth
but with limited waveform distortion. The signal amplitude then decays a small amount to a point where
thediodes are cutoff, and the oscillation is restarted.

Experiment # 55.2: Amplitude Control

Harmonic Oscillators 55-3 Copyright © 1996 M H Mille

Add the diode-controlled nonlinearity and observe the circuit waveforms at the output of each amplifier.
Compare the signal distortions with and without the diode branch active.

Phase-Shift Oscillator
The tuned circuit oscillator makes use of the minimum number of singularities (three) necessary to form an
oscillator, in particular two poles and a zero. Since two poles are necessary in any case, since complex
poles must occur in conjugate pairs, a three pole system corresponds to the only other combination of three
singularities that can be used. The 'phase shift' oscillator uses three RC filter sections to obtain the phase
shift necessary to meet the Barkhausen requirements. A representative oscillator configuration is drawn
below.

The inverting amplifier on the left provides gain to compensate for signal attenuation in traversing the
ladder filter, and 180° phase shift to compensate for the 180° of the phase-shift circuit. The voltage
follower on the right side buffers the ladder output from the input resistance of the inverting amplifier. The
loop is closed by connecting the output of the voltage follower to the amplifier input. Assuming adequate
gain the circuit will oscillate at a frequency for which the phase shift across the ladder is 180°

Assume, as will be the case in the test circuit, that the amplifiers operate without contributing excess phase
shift, i.e., the signal frequencies involved are well below the amplifier poles. Analysis of the ladder filter
shows that

Setting s=jω and applying the oscillation condition that the phase shift be 180° leads to the requirement
3(ωRC)2 = 1 and the gain requirement Amplifier Gain = -8

Experiment # 55.3: Phase Shift Oscillator

Assemble the circuit shown using as nominal values R =10 kΩ and C = 0.03 µf. Add the diode nonlinear
gain-limiting network used before for amplitude control. Observe the amplifier output waveforms and
compare the observed frequency of oscillation with the calculated value. Decrease the amplifier gain until
the oscillation stops. Determine the approximate amplifier gain at which this occurs and compare to the
theoretical gain requirement for oscillation.

Wien Bridge Oscillator

A number of frequency-selective networks used in oscillators take the form of a bridge network. The
particular network drawn below is called a Wien Bridge, and a distinguishing characteristic is that the
voltage transfer ratio Vout/Vin involves four singularities, two zeros and two poles; the figure also shows
the loop gain expression.

Harmonic Oscillators 55-4 Copyright © 1996 M H Mille

With the proper choice of the resistor divider ratio k the parameter Q* -> ∞ (theoretically), which has the
result of moving the zeros of the transfer expression to the imaginary axis at ±jωo. The root locus of the
poles of the expression converges on these zeros, adding stability to the oscillation frequency. This
stabilization is indicated by the stability measure used before. Thus at resonance (ω = ωo) the stability
measure as defined before is

and theoretically approaches infinity for the special case cited.

Experiment # 55.4: Wien Bridge Oscillator

A representative Wien Bridge oscillator circuit diagram is drawn below. Assemble the circuit and observe
the signal waveforms; describe your observations in your report. In practice one often finds R1=R2 and
C1=C2, and the value of k for the optimum stability then is 1/3.
The frequency of oscillation is ωo=1/RC.

Quadrature Oscillator
The description of the preceding oscillators has been in the frequency domain, i.e., in terms of poles and
zeros. A relatively straightforward way of obtaining a sinusoidal oscillaton is by assembling a circuit
whose operation in the time domain mimics a second-order differential equation with constant coefficients;
the natural solution of such an equation is sinusoidal. Actually it is easier to integrate the equation twice to
obtain a second order integral equation; assembling good integrators generally is easier than assembling
differentiators. The circuit diagram below illustrates such an oscillator configuration.

Harmonic Oscillators 55-5 Copyright © 1996 M H Mille

The stage to the left is an inverting Miller integrator (with a nonlinear diode amplitude limiting network
added across the feedback capacitor). The other stage is a noninverting integrator. (It can be shown that
the current through the capacitance in the second stage is as indicated on the diagram, and the integrating
property of the amplifier follows directly from this.)

The output of the second stage provides the input of the first, leading to a circuit with the integral equation
operating description desired. Note that because the output of one stage is the integral of the output of the
other stage both sine and cosine signals are available; hence the name 'quadrature' oscillator.

Experiment # 55.5: Quadrature Oscillator (Optional)

Assemble a quadrature oscillator using as nominal values C=0.03 µf and R=10 kΩ. Observe and describe
the circuit waveforms. Compare the calculated and observed frequencies of oscillation.

Harmonic Oscillators 55-6 Copyright © 1996 M H Mille