This article on audio amplifier testing is extracted from the original four part series on Getting The Best From Your Oscilloscope.
The effects of harmonics on a sine wave are best understood by performing a short, but extremely profitable, experiment. The only instrument required (apart from an oscilloscope) is an audio sine-wave oscillator with a frequency range of 90 to 410 Hz. Actually, it is not necessary to have a variable frequency oscillator; fixed frequencies of 2 times, 3 times, etc., the mains frequency are all that is required. The type of oscillator used should be of the sine wave variety. Any oscillator described as 'phase shift' or 'Wien bridge' will be suitable - many other types of audio oscillator do not give a pure sine wave and are therefore unsuitable.
A simple circuit which enables the effects of adding harmonics to be studied.
We now come to the experiment itself. A voltage of 6.3, obtained from the mains via a heater transformer, is applied to the Y terminals of the oscilloscope by way of potentiometer VR1, as shown above. The audio oscillator is applied to the slider and earthy end of this potentiometer, with the result that the signal level it injects into the circuit is adjustable. The value of VR1 may be between 5 and 50kΩ.
We next set up the oscilloscope timebase to show two or three cycles of the 50 Hz input. If this is a perfect sine wave it will have the appearance shown in waveform (a) below, but there may be slight irregularities due to the presence of harmonics. The oscillator is set up to 100 Hz, the second harmonic, and an output voltage of 0.6 (which will correspond to approximately 10% of second harmonic) is then injected. The result will be a waveform similar to that shown in (b). If the second harmonic input is varied slightly in phase, the irregularity it introduces can be made to travel along the fundamental. See (c). Also, the amplitude of the second harmonic can be varied by adjusting VR[1}. It will be noticed that harmonic voltages below about 5% are almost unnoticeable.
(a) The fundamental sine wave
(b) Adding approximately 10% second harmonic
(c) The effect given by second harmonic with a different phase relationship
(d) Adding third harmonic
(e) Third harmonic with a different phase relationship
(f) The markedly asymmetric waveform given by adding fourth harmonic.
The effect of adding the third harmonic, at 150 Hz, can next be examined, and this will give a trace similar to that shown in (d), or, with a different phase relationship, to that illustrated in (e).
Note the basic difference between even and odd harmonics. Odd harmonics do not alter the symmetry of the fundamental, whereas even harmonics introduce marked asymmetry. (f) shows the result of adding the fourth harmonic, at 200 Hz.
The effect of injecting higher harmonics can also be checked with the aid of this circuit. It is possible, simply by connecting a second audio generator in series with the first, to study the effects of two harmonics on the fundamental.
Audio Amplifier Testing
A typical high quality audio amplifier circuit. The circled figures indicate the points at which oscilloscope checks may be made and the order in which they should be carried out.
AF amplifiers can be tested with an oscilloscope by making sequential stage checks, using a sine wave input. The circuit of a typical 10 Watt high quality valve amplifier is given above, and the circled numbers in this diagram indicate the test points to which the oscilloscope may be applied, and the order in which the checks should be made. It will be seen that tests commence at the speaker transformer secondary, and next proceed back to the input. Points 11 to 16 are bypassed, and little signal should be present at these points. A large signal amplitude indicates a faulty bypass capacitor, which must be replaced before further tests can be made. Points 17 and 18 are also bypassed, but it has to be remembered that a small level of signal will be present here due to the negative feedback from the speaker transformer secondary. The final test is at point 19, which is the positive terminal of the reservoir capacitor for the HT rectifier. Some ripple voltage will be evident at this point, and also at point 13. Experience with serviceable amplifiers will assist in determining the acceptable level for such ripple and will also give a good working idea of the waveform amplitudes to be expected at other points in the amplifier.
If the amplifier is faulty it may be necessary to adjust the volume control to the level where distortion is most apparent. Be careful not to overload the amplifier, or this will result in distortion, such as is shown in diagrams (a) and (b) below.
(a) The clipping which results from overloading
(b) Severe overloading may cause the appearance of a waveform which approaches that for a square wave
(c) An effect given by too low a bias voltage
(d) Too high a bias voltage may give this waveform.
If any stage introduces distortion, the signal at its output will be a distorted version of that at its input, and this fact will enable the stage to be located.
The oscilloscope will also give traces which may indicate the nature of the fault. (c) shows the result of too low a bias voltage in a valve amplifier stage. Clipping occurs on one half of the sine wave due to the onset of grid current. Too high a bias voltage can result in the waveform shown in (d), in which one half of the sine wave undergoes greater amplification than does the other. These last two waveforms may require a fairly high signal input (but below that which would; normally cause the amplifier to overload) to give a recognisable indication of the fault.
A further series of waveforms is given in in the diagram below. That shown in (a) illustrates the effect given due to noise introduced by a faulty component. A parasitic oscillation is shown in (b), this occurring at one point of the sine wave only. The parasitic oscillation may only occur at some input signal levels. (c) shows the clipping given by too low a screen-grid voltage in a voltage amplifier. The amplitude of this waveform will also indicate a low level of amplification in the associated valve. The ripple voltage appearing across the amplifier reservoir capacitor can have the appearance shown in (d).
(a) The result of noise due to a faulty component
(b) The appearance of a parasitic oscillation
(c) A clipping effect, combined with low amplification, given by too low a screen-grid voltage
(d) The ripple across the reservoir capacitor of the audio amplifier may have this appearance.
Square Wave Testing
Sine waves can be used, as has been shown, to check for distortion in audio amplifiers, but a more definite test is afforded by square waves. The effects of square waves can be seen in the diagrams below..
(a) The square wave fed to the amplifier input
(b) Waveform distortion of the type given by a differentiator circuit
(c) The type of distortion given by an integrator circuit
(d) The output given by an audio amplifier with a reasonably flat response
(e) Increased sloping indicates poor LF response
(f) Good LF response but poor HF. response
(g) The output resulting from HF peaking or a tendency towards parasitic oscillation.
Diagrams (b) and (c) show the effects of typical capacitor-resistor circuits of various time constants on a square wave.
Typical amplifier output responses are given in (d) to (g). (d) indicates that the amplifier frequency response is reasonably flat, there being only slight sloping of the square wave. In (e) the sloping is considerably more pronounced, this indicating a good HF response and a poor LF response. Fig. 1 (f) illustrates a reverse effect, in which the rise and fall sections of the square wave are rounded. This corresponds to a poor HF response and a good LF response.
(g) shows a damped train of oscillations after each rise and fall section of the square wave. This effect is the result of peaking at a high frequency or of a tendency towards parasitic oscillation at that frequency. A useful measure of the frequency at which the damped train of oscillations occurs is given by counting the number of cycles in the damped train and comparing these with square wave frequency.
For general amplifier testing a square wave of the order of 400 Hz is useful for practical work. High and low square wave frequencies may be used for more specialised testing.
Square wave checks are capable of showing amplifier response shortcomings because a square wave may be analysed into a set of harmonically related sine waves. If any of these harmonics are amplified or attenuated excessively then the wave becomes distorted. Since the oscilloscope amplifier may not be capable of reproducing the square wave exactly, it is desirable to initially couple the square wave generator directly to the oscilloscope and note the result.
It is of particular interest to pass square waves through the differentiating and integrating networks shown in (b) and (c). Depending on the values of R and C the square wave may be deformed in various ways. from a series of pulses of almost no duration to a wave similar to that shown for a long CR in (c). Many of the simpler tone circuits in audio amplifiers are, in fact, integrator or differentiator circuits, hence square waves are most useful in obtaining an idea of their overall effect.