WITH a knowledge of the somewhat critical and specialized circuits used in hi-fi equipment (and also of the peculiarities of temperament of the enthusiastic hi-fi owner!) the service technician will soon find himself as much at home with hi-fi equipment as with radio or television receivers.
Basically, the hi-fi chassis is far less complex than a modern television chassis, but because of the delicately balanced circuits in the former and the subjective nature of hi-fi, servicing compromises are never worth adopting.
For example, if an anode-load resistor is found to be open-circuit, and it is a close-tolerance component valued at 50,000 ohms, then replacement should be made with a component of identical characteristics. Whilst a preferred-value 47,000-ohm resistor would restore operation of the amplifier, and the service technician may feel that the performance is then well up to standard, the owner who is highly sensitive to every characteristic of his amplifier will soon sense that something is not quite right.
He will probably possess the circuit diagram of the unit and may eventually find the incorrectly valued resistor. Immediately he will attribute to this the shortcoming in performance. He will obtain and himself fit the correct component, and whilst technically the results may not be improved, the enthusiast will believe that they are and will be quite satisfied that the amplifier is now up to its former standard. Word will soon get around the local hi-fi world about the wicked ways of the unfortunate service technician, and he may have difficulty in regaining the confidence of the local enthusiasts.
COMPLETE FAILURE
This is one of the easiest of faults to locate. The first check would be to establish the connection of power to the equipment. If the valves and pilot bulb are not alight, the fault is almost certainly in the mains input circuit.
The plug and socket connections at both ends of the mains lead should be examined carefully, and it should be established that power is actually present on the mains socket. There may be a break in one of the conductors of the mains supply lead. This often happens if the equipment is moved around extensively, but the trouble is usually of an intermittent nature.
The next move would be to check the amplifier fuse or fuses for continuity. If these are in order the on/off switch, associated connecting cables, connections to the primary of the mains transformer and the voltage-selector plug and socket should be carefully examined. Sometimes a poor or intermittent connection exists on the voltage-selector connector or a dry joint develops on the mains circuit connections. Testing along these lines will soon reveal the cause of the trouble.
If it is found that the fuse is open-circuit, a check for short-circuits on the h.t., l.t. and mains circuits should be made before a new fuse is fitted and the amplifier switched on. Fuses are fitted to protect these circuits, and a fuse rarely blows without provocation. Test for shorts can be made with a simple ohmmeter. A check on the l.t. circuits should first lead to removal of all the valves and connection of the ohmmeter across the heater line, bearing in mind that the line is shunted by the heater winding on the mains transformer and possibly a humdinger control. Removal o( these components may also be called for if the meter used cannot indicate low ohms.
To check for a h.t. line short, the meter should be connected between the chassis and the h.t. line, bearing in mind the charging and discharging kicks promoted by the electrolytic capacitors. If a reading of some hundreds of ohms, or less, is given, the probe of the instrument should be transferred to the various h.t. feeds until the source of the low resistance or short-circuit is brought to light. Typical faults in this respect are shorting smoothing electrolytics, a short in the h.t. rectifier, a winding-to-core short in the smoothing choke, and valve-holder shorts to chassis.
If the circuits appear completely free from excessive leakage resistance, the fuse should be replaced (one of stipulated value is essential for optimum protection) and the amplifier re-connected to the mains and switched on.
While it is warming up the valves, particularly the h.t. rectifier and output valves, should be carefully observed for signs of an internal flashover. A heavy flashover of this nature will immediately cause failure of the replacement fuse. In this event, both the valve and the fuse should be replaced.
If the valves are alight the program signal should be disconnected or the volume control fully backed-off. At this point it should be made clear that the audio-frequency voltages developed across the primary winding of the output transformer rise to a high level if a signal is conveyed through the amplifier at normal level with the loudspeaker load removed. It is thus essential to establish continuity of the loudspeaker circuit with the signal removed. Many a good and expensive output transformer has been damaged by operating the amplifier without a correct load. The amplifier can be run at full power, of course, by using a suitable load resistor in place of the loudspeaker; a wire-wound component rated at the full output power of the unit should be used in this case.
It is a simple matter to check loudspeaker and connecting-lead continuity and resistance by disconnecting the loudspeaker wires from the speaker terminals on the amplifier and then connecting the leads to the terminals of a battery-operated ohmmeter. A crackle will be heard in the speaker on connection and disconnection of the wires.
Once it has been established that the loudspeaker is, in fact, acting as a load and is in good condition, the volume control can be advanced to its normal setting without fear of damaging the output transformer. At this stage, the program-selector switch can be turned over the various positions, which, provided the appropriate program signals are available, will indicate whether or not the failure is common to all channels.
Assuming that all channels are dead, tests should be made to find out whether the trouble lies in the pre-amplifier or power amplifier. This is a simple matter with two-unit amplifiers, it being necessary to unplug the pre amplifier from the power amplifier and apply one of the program signals direct to the input socket on the power amplifier; the pick-up signal is usually suitable for this test. Whether or not the power amplifier will be fully loaded by this signal will depend upon the overall sensitivity of the amplifier and the level of the pick-up signal-depending upon the type of pick-up used. At this stage, however, we are not interested in the quality or quantity of the sound, and provided we get a reasonable form of reproduction, it is fairly safe to assume that the trouble lies in the pre-amplifier section.
With a single-unit combination amplifier, a similar test can be made by applying the signal between the chassis and control grid of the valve immediately prior to the phase-splitter stage.
POWER AMPLIFIER FAILURE
FIG. 4.1. Circuit diagram of the RD Junior amplifier.
When making signal tests on certain power amplifiers with the pre amplifier or control unit disconnected, it may be necessary to short-circuit the points on the control-unit socket corresponding to the mains on/off switch, since this switch is usually situated on the control unit. Fig. 4.1 shows the circuit diagram of the well-known RD Junior power amplifier, in which points 7 and 8 on the control-unit socket are those associated with the mains on/off switch; shorting these will complete the mains input circuit. Referring to the same circuit, the signal would be applied across points 5 and 6 of the same socket, with the earthy side of the signal source to point 6.
A probable cause of the lack of response is failure of the h.t. circuits, and this can be established quite rapidly by testing the temperature of the output valves, V2 and V3, with a finger. The output valves usually operate at a fairly high temperature, and if they are only just warm there is a strong likelihood of a burnt-out h.t. rectifier valve, V4; failing that, the h.t. feed resistor R14 should be subjected to a continuity test. It is surprising how much can be done in the way of simple servicing and diagnosis without instruments--merely by applying a little well-concentrated thought.
It is extremely unlikely that the fault would be caused by simultaneous failure of both output valves; whilst failure of one of the output valves would result in the offending valve losing temperature, the good valve would remain too hot for comfortable touch, and the amplifier would reproduce after a fashion. The same applies with regard to the output-transformer primary windings--it is most unlikely that both sections would go open-circuit at the same time.
There is one more possibility, however, and that is open-circuit of the common cathode resistor R12. This trouble would cause both valves to lose temperature, though it is possible that the bypass capacitor C6--being a low voltage electrolytic-would leak heavily and give some sort of cathode circuit continuity. In this case, the amplifier would reproduce, but the distortion would be high. Some amplifiers have separate cathode resistors and, again, both would hardly fail simultaneously, though there is a remote chance of this happening! If both output valves are working at fairly high temperature, and there is a very slight trace of normal residual mains hum from the loudspeaker, one can be fairly certain that the voltage-amplifier/phase-splitter stage, V1, is defective. To check the phase-splitter section, a signal could be applied to its grid-pin 7 in the circuit of Fig. 4.1 ; a pick-up signal may not be strong enough at this high-level point (depending upon the output voltage of the pick-up) and it may be necessary to bring into service an audio oscillator.
Failing this, however, one side of the heater line could be connected to the grid through an 0.1 mF capacitor. This action will inject into the grid circuit a 60-hz mains signal (at about 3 volts) and, if the phase-splitter section is operational, will give rise to a very loud mains hum from the loudspeaker.
The remaining stage is the first triode section of VI-the voltage amplifier.
Open-circuit of the anode-load resistor R3 or the coupling capacitor C2 represent the most likely causes of the trouble. However, first a valve change and then a check of anode voltage will soon bring to light the trouble.
The same simple tests are all that are necessary if, for instance, the previous tests indicate trouble in the phase-splitter stage.
PRE-AMPLIFIER FAILURE
Fig. 4.2 shows the circuit diagram of the RD Junior control unit (pre amplifier). If it is found that the power amplifier passes a signal, but some fault is preventing its passage through the pre-amplifier, it is best to make tests with the two units connected together in the normal manner. However, before delving too deeply into the pre-amplifier circuit from the servicing aspect, it often saves considerable time to ensure that the signal-carrying conductors of the multi-core pre-amplifier connecting-cable not only possess continuity, but that they are also in good electrical connection with the tags on the plugs.
Fig. 4.2. Circuit diagram of the RD Junior control unit, Mk. II.
It is best to work back from the tone-control valve, V2b, to the first voltage-amplifier, VI. The signal fed to the power amplifier, by way of point 5 on the octal cable plug, is developed across the volume control P7, being picked-up from the anode of V2b. To check the goodness of stage V2b, the volume control should be advanced about three-quarters of maximum, and pin 7 of V2 touched with the blade of a screwdriver, with the blade making contact with a finger. The other hand should be kept well away from the amplifier, preferably in a pocket to avoid the risk of electric shock. If all is well, a loud hum will emit from the loudspeaker, as the result of the small mains signal being picked up by the body and injected into the grid. This test can be repeated at the grid of V2a (pin 2) and the grid of V 1 (pin 9). If there is a loud hum at pin 7 of V2 and no hum (or a very weak hum) at pin 2 of V2, the trouble lies either in V2a, in R22, R23 or in the coupling capacitor C23. The valve is the most likely cause, and should at least be checked by substitution. If there is a loud hum at the grid of V2a, but no hum at the grid of VI, VI itself, R8, R7 and C4 should be checked in that order.
TRACING AND CLEARING FAULTS IN AMPLIFIERS
If R5 (the screen-feed resistor) appears to be overheating, suspect a short in C2. A few simple voltage and resistance checks will soon bring to light the component responsible.
H.t. power for the pre-amplifier is applied from point 3 on the octal cable plug. Make sure that h.t. is present here, and that it is getting past R32 and R9 (filter resistors). Overheating of R32 would indicate a short in C30, while the same trouble in C1 would cause R9 to overheat.
There is usually no need to set up elaborate instruments to diagnose for total failure if the tests outlined above are followed logically. Once the defective section has been revealed, the problem is virtually solved, for it is then only a matter of testing a few small components and the voltage at a couple of key points.
Instead of relying on the hum method of testing, the signal from an audio oscillator or generator can be applied to the various stages in turn until the point is reached where the signal is blocked; but generally speaking, the hum method is the quickest, and just as reliable. Alternatively, a pair of headphones, or an ear-piece, can be used to trace the signal through an amplifier up to the stage or component which is preventing it getting any farther. This method of testing calls for a normal input signal from one of the program sources and average settings of the various controls. The phones can be used to trace the signal from the program source right up to the point of the trouble. For more complex faults, test instruments are usually required.
DISTORTION
Distortion in one form or other probably accounts for the majority of troubles in hi-fi amplifiers. The symptom ranges from a very low-level distortion, which invariably demands some curious instinct to detect, to a very high-level distortion, whose presence is obvious to any listener.
The reader should understand that there is no such thing as a completely distortionless reproducing channel. Somehow, somewhere, in the electro-acoustic link between the live program in the studio or concert hall and the ear of the listener at the loudspeaker end, the original sound will be altered slightly in character. It may be "colored" by the position of the microphones in the studio and by the position of the loudspeaker and room acoustics at the listening end of the link. It will most definitely be modified during its passage in electrical form through the various electronic circuits If a number of microphones are used close to the instruments of an orchestra, the pick-up of direct sound will be far in excess of the pick-up of reflected sound and the reproduced sound will lack "atmosphere"; it will not sound the same from the loudspeaker as it would in the middle of the concert hall. Little can be done by the enthusiast to correct this trouble, however.
FIG. 4.3. Arrangement of instruments for checking frequency-response
and over loading.
At the reproducing end, the room acoustics will obviously differ from those at the transmitting end, and even if a desirable degree of "atmosphere" is introduced, the final result will be further colored by the listening-room acoustics. If "atmosphere" is purposely excluded by the sound engineer, it is most unlikely that the acoustics of the listening room will resemble those expected of a concert hall. A compromise is necessary along these lines, and this is the main reason why hi-fi amplifiers use elaborate tone-control circuits.
FREQUENCY DISTORTION
Frequency distortion is present when the output signal deviates widely in amplitude as a constant-amplitude input signal is altered in frequency over the entire audio spectrum. Almost all hi-fi amplifiers are substantially flat in response over, and beyond, the audio spectrum, as we have already discovered, and they are rarely troubled with this form of distortion. How ever, at high power outputs, the response may not be quite as flat as suggested by the appropriate response curves.
In Fig. 4.3 is shown an arrangement of instruments which can be used for frequency-response checking and plotting. An audio oscillator or generator is coupled to the input of the amplifier under test, ensuring that it is correctly matched to the input channel selected, a load resistor of suitable value and rating is employed in place of the loudspeaker and the voltage (a.c.) across it is measured by the output meter. The output signal is also monitored on an oscilloscope.
For high-level testing, the amplifier volume control is turned to maxi mum, the tone controls to the "flat" position, the filters switched out, the generator tuned to 1,000 hz and the generator gain control adjusted for maximum power of the amplifier as given on the output meter. The waveform is synchronized on the oscilloscope to ensure that it is not highly distorted owing to overloading of the amplifier by too great an input signal.
With the various controls set, the generator should be tuned to about 20--30 hz, the oscilloscope re-synchronized to that frequency and the wave form checked to ensure that it is still free from distortion. Normally, a pure sine wave will be displayed, depending upon the signal given by the generator, but if the peaks of the wave appear to be flattened, the input signal should be decreased until the distortion disappears, a note being made first of the original setting of the gain control. The output level should be noted at each point as the test is made over the audio spectrum, up to the limit of the generator, and plotted against frequency to give the response curve.
If it was necessary to decrease the input signal at the lower-frequency end, the gain should be advanced progressively up to 1,000 hz, ensuring each time a test is performed that the signal is not overloading the amplifier.
For low-level testing, the same procedure is adopted, but this time the input signal is adjusted to give about I watt power output. In this case, there will be little danger of overloading the amplifier, and an oscilloscope is not essential.
If a proper output meter calibrated in watts of power is used, it will probably incorporate its own load resistor, but it must be ascertained that this represents the correct match to the amplifier; the mete1 should also have a level response itself over the audio spectrum. It is similarly pointless making such tests with an audio generator whose output voltage varies greatly over the band; if the instrument does not have a voltage-output indicator of its own, then its response should be plotted on a curve, which can later be used to correct the amplifier response curve.
If an output meter is not available. a high-resistance level-response a.c. voltmeter can be used equally well. The power output can be computed by using the expression:
W=E2/R,
... where E is the voltage and R is the resistance of the load in ohms.
The oscilloscope (which is in valuable for many tests on hi-fi equipment) should possess a good low-frequency response in relation to its Y amplifier (preferably from d.c. to I M hz or above), have a linear timebase and ease of synchronizing the test signal. An instrument highly suitable for this work is the Serviscope, by Telequipment.
Among many other refinements, this ....
-------
FIG. 4.4. The Serviscope, by Telequipment.
FIG. 4.5. Input-voltage/output-voltage characteristic: A-Bfor a distortionless
amplifier: A-Cfor a practical non-linear amplifier.
----------
.... has a Y amplifier which is substantially flat from d.c. to 6 M hz and a trigger-type timebase which obviates complex synchronizing re-adjustments on altering frequency (see Fig. 4.4).
Incidentally, an audio oscillator having provision for a square-wave output is most desirable, since many tests can be made by injecting a square-wave signal into the input and observing its form after passing through an amplifier.
A INPUT VOLTAGE. An amplifier suffering from frequency distortion is characterized by its somewhat "mellow" tone, which is caused by severe attenuation of the higher frequencies in relation to the low frequencies.
NON-LINEAR DISTORTION AND HARMONICS
Owing to the curvature of valve characteristics, deficiencies of the output and coupling transformers, etc., the input voltage/output voltage characteristic of any amplifier deviates from a straight line over the major portion of its range. The effect is shown graphically in Fig. 4.5; here line A-B would represent the characteristic of a distortionless amplifier, but in practice the characteristic takes the form of the broken line A-C. From this it will be seen that the non-linearity is aggravated as the input voltage is increased.
There is no such thing as a distortionless amplifier! The generation of harmonics of the fundamental frequency of the input signal is one of the by-products of this non-linearity. For example, if the input signal is a pure sine wave of frequency 250 hz, the output signal will consist of the fundamental 250 hz signal, a second harmonic at 500 hz, a third harmonic at 750 hz, a fourth harmonic at 1,000 hz, and so on. The magnitude of the spurious harmonic signals will depend upon the extent of the non-linearity, and they are usually expressed in the form of a percentage of the magnitude of the fundamental signal. Hence, if the power of the fundamental signal is 20 watts, and there is a second harmonic power of 1 watt, it could be said that the amplifier has a second harmonic distortion of 5 percent. Usually, however, it is the total harmonic distortion of an amplifier which is given in the form of a percentage. With push-pull amplifiers, as we have already seen, the second and even harmonics are largely precluded by cancellation in the balanced load, and the third and possibly higher-order odd harmonics are the troublesome ones.
Harmonic distortion when present in a large degree is characterized by the harsh, "rough" nature of the reproduction. At lower levels, it is ...
FIG. 4.6. Various modes of second-harmonic distortion.
(Left) FIG. 4.7. Third-harmonic distortion. (Right)
FIG. 4.8. Severe harmonic distortion caused by iron saturation in an output transformer due to an overload.
... difficult to define objectively, but its presence has a fatiguing effect on the listener; hi-fi enthusiasts are able to sense that something is not quite as it should be, and are glad to get out of audible range! Some harmonics are distinctly unpleasing, to say the least, particularly those which are dissonant with the fundamental frequency, such as the seventh, ninth, eleventh, etc. of a fundamental of 250 hz.
Conversely, the emphasis or suppression of certain harmonics of certain sounds tends to enhance the original sound, and in some cases makes a displeasing sound more pleasing. This effect may be created by the use of the various tone controls and filter controls on the amplifier.
The deformation of the waveform produced by the harmonic components depends upon the phase of the harmonic relative to the fundamental.
In Fig. 4.6a is shown, in broken line, a fundamental and second harmonic, which combine to form the distorted wave in full line. The combined wave is obtained by adding or subtracting the instantaneous values of the two waves. In (b) the harmonic component is displaced from the fundamental by 45 deg., resulting in a combined wave of somewhat different character, while in (c) the harmonic is displaced by 135 deg., which has the effect of inverting the distorted combined wave.
A third-harmonic component has the effect of distorting the waveform like that shown in Fig. 4.7. A characteristic of waves distorted by odd harmonics is that the positive and negative halves of the combined wave are similar, while with even harmonics the positive and negative half-waves are mirror images. In Fig. 4.8 is shown severe harmonic distortion created by iron saturation as the result of overloading of an output transformer.
INTERMODULATION DISTORTION
Another effect of non-linearity is intermodulation distortion. The effect occurs when more than one input frequency is applied to the amplifier, giving rise either to the production of sum or difference frequencies, or to the amplitude modulation of one frequency by the other.
Although in practice there are a host of input frequencies applied simultaneously, an illustration of the first effect is afforded by considering the application of only two frequencies, one at, say, 40 hz and the other at 1,000 hz. A whole string of sum and difference frequencies will be formed; for instance, the 40 hz note will add to and subtract from the 1,000 hz note, thus producing spurious signals at 1,040 els and 960 hz. If the 40 hz signal is of greater amplitude than the 1,000 hz signal (tests are usually made in a ratio of magnitude of 4 to I), harmonics of the 40 hz signal will also add to and subtract from the 1,000 hz signal, thereby giving spurious signals at 1,080 hz, 920 hz, I, 120 hz, 880 hz, and so on. Harmonics of the 1,000 hz signal may also come into play if the non-linearity is severe, and make matters even worse! Intermodulation distortion of this type is very unpleasing to the listener, being far more disconcerting than simple harmonic distortion because the spurious sum-and-difference tones are not harmoniously related to the fundamental frequencies. Such distortion is characterized by a "buzz" or "rough harshness" in the reproduction, and is apparent to almost any listener.
Amplitude modulation of one frequency by another has been illustrated admirably by G. A. Briggs in his book, "Sound Reproduction". Instead of being connected to a source of d.c., the field coil of an early-type loudspeaker was inadvertently connected to a source of "raw" a.c. and, to quote the writer: "when a record was played through this equipment, music came out of the loudspeakers but it was almost unrecognizable, sounding as though it had been chopped up in a high-speed slicing machine. One effect of inter modulation is similar to this but not quite so bad." Another example of intermodulation of this kind is the reproduction of a choir with an organ accompaniment. The author has had occasion to investigate this effect; a tape-recording made during a choir practice was said to have a curious "warbling" characteristic. This was because the choir was amplitude-modulated by the organ! Harmonic distortion is checked on a wave analyzer which is connected to the output of an amplifier. A sine-wave input is applied to the amplifier, and the wave analyzer removes the fundamental frequency, and passes only the harmonic components, which are measured as a percentage of the fundamental.
An intermodulation analyzer is required for measuring intermodulation distortion. The analyzer usually supplies the two input signals over a selected range of test frequencies, and has an attenuator which alters the ratio of the two signal voltages over the range of 1:1 to 10:1. The composite signal is applied to the input of the amplifier under test, and the output of the amplifier is fed through a high-pass filter which eliminates the low-frequency test signal. The remaining signal, consisting of the high-frequency test signal plus the intermodulation, is demodulated. The high-frequency test signal is also eliminated in another filter, and the spurious intermodulation signals are fed to a measuring instrument which usually reads in terms of percentage intermodulation.
PHASE AND TRANSIENT DISTORTION
Phase distortion causes the output waveform to differ from the input waveform, due to alteration of the phase angle between the fundamental frequency and an associated harmonic, and of the phase angle between any two component frequencies of a complex wave. Although phase distortion has an effect on the reproduction of transients, it would, generally speaking, appear to be the least troublesome distortion encountered in audio work, though its presence is clearly visible in television receivers. With a hi-fi amplifier of wide frequency response, phase distortion is usually negligible, but it rises somewhat by the inclusion of filters which serve to limit the frequency range.
Distortion of the transients which, with certain kinds of music, occur at very high level, tends to impair the "attack" performance of the equipment.
Transients are representative of sounds of short duration, such as those produced by certain string instruments-the piano, for example-and by percussive instruments. The general effect is that such reproduced sounds tend to "hang-on" after the energizing pulse or waveform has decayed, and where the distortion is severe, the frequency emitted during the period of decay may differ from that of the actual energizing waveform. Reproduction becomes very "slurred" on peaks.
FIG. 4.9. The input square wave shown at (a) promotes severe "ringing",
as shown at (b), Slight "ringing", as at (c), usually has little
effect on the transient response.
Apart from a good transient response, depending to a large degree on both the electrical and acoustical damping of the loudspeaker system, the amplifier should also possess (1) a wide frequency response, extending beyond the limit of audibility, (2) no phase distortion, (3) a high output damping factor and (4) all resonant circuits, such as tone-control networks, filters and transformers should be sufficiently damped to avoid "ringing". A circuit which is subject to damped or supersonic oscillations will be triggered into transient distortion by transient pulses, and the spurious oscillation-at frequencies depending upon tuned frequencies of the offending circuits-will become superimposed on the signal waveform.
Testing for transient distortion is performed by injecting a square-wave signal into the input of an amplifier and observing its character on the screen of an oscilloscope connected across the output terminals; the amplifier should be properly matched at both the input and output terminals for this test. Fig. 4.9a shows the input square wave. Diagram (b) shows a very high degree of "ringing", which may be of such high amplitude on the peaks as to cause overdriving of one or more stages of the amplifier. The waveform at (c) shows a trace of "ringing" which, in practice, may have little significant effect on the reproduction.
OTHER SQUARE-WAVE TESTS
Square waves can tell us other things about an amplifier; they are useful because their formation depends upon the fact that they comprise harmonic components of their fundamental frequency extending well above the limit of audibility.
The low-frequency performance of an amplifier can be checked by applying a square wave having a fundamental frequency of, say, 50 hz.
Fig. 4.10a shows the usual resultant waveform on the screen of the oscilloscope. There will be some slope on the top of the waveform, but provided it is no greater than 40-50 percent of the height of the waveform, the low frequency performance can be considered satisfactory. It is important to check the oscilloscope on the square wave direct, however, to ensure that the wave is, in fact, square and that the Y amplifier itself (if used) is not responsible for distortion.
FIG. 4.10. The frequency response of an amplifier can be checked by
square waves; (a) reasonable low-frequency response: (b) poor high-frequency
response: (c) very poor high-frequency response: (d) good high-frequency
response.
Having first ensured that the square wave given by the generator is maintained in accuracy over the whole of the audio spectrum, in terms of generator and oscilloscope performance, the square-wave signal applied to the amplifier can be varied in frequency up to 20,000 hz, and the display observed at various intermediate frequencies.
With a good amplifier, the square-wave display should remain essentially uniform up to about 5,000 hz, at which point very slight "ringing" may be evidenced. If the amplifier has a poor high-frequency response, the waveform will deteriorate to that shown in Fig. 4.10b. As the input frequency is increased up to 10,000 hz, a good amplifier will maintain a reasonable square wave display, similar to that of Fig. 4.10d, but with a poor amplifier the display may deteriorate from waveform (b) to waveform (c). Increasing the input frequency up to 20,000 hz really tests the upper-frequency response of the amplifier, but with a good hi-fi unit the waveform should differ little from that shown at (d).
PHASE-SHIFT TESTS
If two voltages of the same frequency are applied to the X and Y terminals of an oscilloscope, the result on the screen is either a straight diagonal line or an ellipse. The straight line is produced when the two frequencies are exactly in phase; an ellipse is produced when the signals differ in phase, but the dimensions of the ellipse will depend on the phase angle and the relative amplitudes of the voltages. However, should either of the signals not obey the sine law, the displays will be irregular in appearance.
Here, then, we not only have a method of checking the phase shift between the input and output terminals of an amplifier, but also, if we apply a pure sine wave to the input terminals, we can obtain an idea of the distortion given by the amplifier. The sequence of patterns shown in Fig. 4.11 illustrates such a display of two pure sine-wave signals of equal amplitude and frequency, but differing in phase angle from in-phase to 180 deg. out-of-phase.
When the amplitudes of the signals are equal and there is a 90-deg. phase shift a perfect circle will result, and intermediate ellipses will occur either side of this point.
The sine-wave signal can be applied to the input terminals of the amplifier under test in the usual manner and a sample of the signal at this point applied to the Y terminals of the oscilloscope. The oscilloscope's time base should be switched off and disconnected, and the output signal of the amplifier--that appearing across a correct-value load--applied to the X terminals. An attenuator may be required at this point so that the Y and X signals can be balanced. The degree of phase shift occurring over the pass band of the amplifier will be revealed on the screen as the sine-wave generator is tuned over the audio spectrum.
Fig. 4.11. Sequence of patterns illustrating two pure sine signals
of equal amplitude and frequency but differing in phase angle from in-phase
to 180 deg. out-of-phase.
FIG. 4.12. Phase-shift patterns. The waveforms at (a) to (e) reveal
clipping of the output signal due to overloading or incorrect operating
conditions of the valves: (f) and (g) show the presence of harmonic distortion
as the result of excessive non-linearity: waveforms (h) and(i) show the
presence of a spurious signal.
Deviation in the symmetry of the display will result if the amplifier output signal is distorted on one cycle only, while if both cycles are equally distorted, the trace will remain symmetrical, but distorted in shape. Spurious oscillations in the system will also be shown on the trace. The various effects are illustrated by the waveforms in Fig. 4.12.
CORRECTION OF DISTORTION
Having first established that an amplifier is, in fact, producing distortion, and that the distortion is not present on the actual program signal or caused by maladjusted controls, steps can be taken to locate and remedy the cause of the trouble. It is a good idea to work from a pure sine wave given by an audio oscillator or generator. and have this signal fed through the amplifier under test and monitored on the screen of an oscilloscope. To ensure correct balance of the circuits, both the input and output terminals should be terminated by the impedance (resistance) specified in the maker's handbook, and an output indicator should be connected across the output load. In this way the output power can be observed in relation to the distortion, and it can be immediately observed whether or not the distortion varies in magnitude as the strength of the input signal is varied.
If it is found that distortion is present only towards the maximum output limit of the amplifier, the most likely cause is overloading of a valve resulting in its being driven into the non-linear portion of its characteristic curve.
Low h.t. voltage, due to a low-emission rectifier, or impaired emission of one of the output valves, is a possible cause of the trouble.
If the h.t. voltage is normal, the oscilloscope can be removed from across the output load resistor and the signal at the input and output of the phase-splitter checked for distortion. If there is no distortion at the input of the phase-splitter, but distortion is present at the output, the phase splitter itself may be responsible. However, there is a possibility that grid current in the output valves is affecting the signal here, and if this is suspected, a test should be made with the phase-splitter coupling components disconnected from the output valves. If the signal from the phase-splitter is free from distortion after this action has been taken, and it remains virtually distortion-less when the input signal is increased, there is little doubt that the trouble lies in the output stage.
The valves should be checked for emission and balance, and the cathode and grid resistors should also be checked for balance. If all seems well here, and the valve test is normal, the output transformer should be suspected for shorting turns. Shorting turns or trouble in the output transformer, apart from an open-circuited winding, is not always an easy fault to diagnose, and a suspect usually calls for a substitution test.
The chief symptom in this respect is lack of power, and if the amplifier is opened-up towards full volume, the reproduction becomes progressively more distorted without an apparent increase in output power; also, the faulty transformer tends to overheat. A short-circuit in one half of the primary winding promotes unbalance in the output stage with a resulting increase in second-harmonic distortion.
CHECK FOR OUTPUT STAGE BALANCE
Amplifiers with an adjustment for output-stage balance have an arrangement whereby the bias of one valve can be altered slightly in relation to the bias of the other valve. One method of achieving this is shown in the circuit in Fig. 4.13. When the stage is in balance, the current in either valve is the same, and the current is equal but opposite in each half-section of the primary of the output transformer. Thus, a voltmeter connected across the two anodes, as shown, will indicate zero voltage when the "balance" control is adjusted correctly.
Fig. 4.13. With a voltmeter connected across the anodes of the valves,
as shown, the "balance" control should be adjusted for zero
reading.
It should be stressed that this adjustment serves only from the d.c. aspect of the circuit, in which case the circuit can be considered as a balanced bridge.
In practice, there is little difference between the setting corresponding to optimum d.c. balance and that corresponding to optimum signal balance, but to secure optimum results in the latter case the "balance" control should be adjusted for minimum distortion, as indicated on a suitable distortion meter.
As the "balance" control usually has a very limited range, its inability to balance the circuit should first lead one to suspect low emission of one of the output valves. If both valves show reasonable balance on a valve tester, the remaining circuit elements should be checked for balance. If a valve tester is not available, and it is found that the voltmeter pointer remains one side of zero over the full range of the "balance" control, the position of the output valves should be reversed. A low-emission valve is definitely responsible if this change causes the voltmeter pointer to remain on the other side of zero over the full range of the "balance" control.
If changing the valves in this way does not reverse the movement of the pointer in relation to zero, a change in value of a component should be suspected. Some amplifiers use separate cathode resistors as well as a "balance" control, in which case the trouble may well be caused by a change in value of one of these. There is also a possibility that one half of the primary of the output transformer has a bad short, causing a decrease in resistance of one half with respect to the other half. This trouble would promote severe distortion and lack of power, as already described.
A leak or poor insulation-resistance of one of the coupling capacitors (C1 and C2 in Fig. 4.13) would also seriously affect the balance of the stage.
These capacitors are in connection with a source of d.c. at the phase-splitter side, so that poor insulation would cause the control grid of the associated output valve to go positive. The negative bias given by the cathode circuit would thus be neutralized, and the affected valve would pass considerably more current than the other. Heavy distortion would occur, and it is most likely that the anode of the affected valve would glow a dull red; in any event the temperature of the valve would be considerably higher than normal. The valve would not last very long under this condition, and the resulting abnormally heavy current would most likely cause failure of the h.t. fuse.
Controls available for balancing the output stage should never be used as a means of neutralizing severe unbalance caused by alteration in the characteristic of a valve or value of a component. The control serves essentially to permit a little extra reduction in distortion content which would otherwise not be possible. Such controls are rarely found in equipment for sound reinforcement and public-address use, where the distortion content is in any case greater than that associated with hi-fi equipment.
Optimum balance of the output stage also serves to minimize the residual mains hum. In fact, balance is sometimes made in this respect; a sensitive a.c. voltmeter is connected across the output load, and with the signal input terminals short-circuited the "balance" control is adjusted for minimum hum voltage. It will be remembered that a similar form of adjustment was recommended for the "humdinger" control.
CHECK FOR SIGNAL BALANCE
If the output stage balances correctly from the d.c. aspect, but slight distortion is present across the load in spite of a distortion-free drive signal, the drive signal should be checked for balance at the control grid of each output valve. A square-wave or sine-wave signal should be applied to the input of the amplifier, and the signal amplitude measured at each grid in turn, relative to chassis, on the screen of an oscilloscope. A high-resistance or valve voltmeter can be used if an oscilloscope is not available and, if necessary, a suitable signal can be obtained from the heater supply.
The signals should be almost identical, but opposite in phase. If considerable deviation in amplitude is observed, the phase-splitter valve and associated components should be carefully checked for value and balance.
With a signal applied, the coupling capacitors (C1 and C2, Fig. 4.13) can be checked for balance by measuring the voltage across them with an oscilloscope or valve voltmeter. Unbalance of these components may not affect the high- and medium-frequency performance, but may incite harmonic distortion at the low frequencies as the result of the reactance of one capacitor differing considerably from that of the other.
Harmonic distortion may rise above that stipulated for the amplifier by open-circuit or low-value of the bypass capacitor across a common cathode resistor. If such a capacitor is not used even harmonics will be injected into the grid circuit. This does not apply where separate cathode resistors are used.
CHECK FOR NEGATIVE FEEDBACK
If the distortion is sudden and severe, an investigation should be made of the main negative-feedback loop if the tests outlined above have failed to reveal the cause of the trouble. If the negative-feedback loop is in order, there should be a distinct increase in output, with the input signal kept at a constant level, on disconnecting the loop either at the cathode, where it is applied, or at the secondary of the output transformer. No apparent increase, or only a very slight increase, in output would indicate that the feedback loop is either open-circuit completely or that the loop resistor has increased in value. A few simple tests will establish the defective component in this case.
PARASITIC OSCILLATION
FIG. 4.14. Circuit diagram of the Pye Mozart combined pre-amplifier
and power amplifier.
FIG. 4.15. The Pye Mozart amplifier, Model HFIO.
Although most amplifiers of hi-fi type have a reasonable margin of feedback stability, an increase in value of the cathode resistor where the feedback loop is connected or a reduction in value of the loop series resistor may increase the feedback above the safety margin and incite parasitic oscillation.
There is a possibility that the frequency of oscillation will be above the audio spectrum, in the supersonic region, where its presence will not be audible as such, but will play havoc with the quality of reproduction. High frequency parasitic oscillation will immediately be revealed on an oscilloscope test of the output signal, but where such an instrument is not to hand, and the trouble is suspected, a milliammeter connected in series with the h.t. feed to output valves can be used as to indicate oscillation. A definite drop in current reading when the feedback loop is disconnected is indicative of trouble of this nature.
If the feedback components appear to be of reasonable tolerance, the output valves themselves should come under suspicion, since a severe unbalance of emission has been known to promote oscillation. In certain amplifiers low-value anti-parasitic resistors are sometimes connected in series with the anode and grid circuits of the output valves, and it should be ascertained that these are in good order.
Other expedients for maintaining stability over the very wide frequency range characteristic of modern equipment are (1) a capacitor and resistor in series in the anode circuit of the first valve of the power amplifier, which serve to reduce the gain at the unstable frequency within the amplifier's passband, and (2) a capacitor in parallel with the feedback loop resistor. The latter component promotes a phase shift opposite to that of the output transformer at the high-frequency resonance of this component, and thus prevents the feedback from turning positive at this frequency. Such devices are sometimes adopted in the Williamson amplifier. These components should be checked for value.
If a replacement output transformer introduces parasitic oscillation, then it may be necessary to modify slightly the value of the phase-shift feedback capacitor. The optimum value is best found by trial and error, and if an oscilloscope and a square-wave generator are available, the value giving the least distortion and "ringing" at 20,000 hz should be chosen. The correct value feedback resistor and phase-shift capacitor must be used for the output impedance selected.
Apart from supersonic oscillations, very low-frequency oscillation may result from a fall in value of an electrolytic decoupling or filter capacitor; this may not be directly associated with the h.t. supply, but serve as a low-pass filter in a voltage amplifier. The effect is usually described as "motor-boating", but in certain instances the oscillation may be less than 10 hz and inaudible.
If all the filter capacitors are up to standard, the output valves should be checked for balance, as also should any push-pull driver valves.
If the feedback connections on the secondary of the output transformer are reversed, the feedback will be positive instead of negative, and very bad oscillation will occur immediately the amplifier warms up. This trouble will not normally be encountered unless the transformer has been replaced and incorrectly connected.
TRACING DISTORTION THROUGH THE AMPLIFIER
As a basis for our tests we shall now refer to the circuit diagram of a commercial amplifier. Fig. 4.14 shows the circuit diagram of the well-known Pye "Mozart" combined control unit and power amplifier (Model HF10); see also Fig. 4.15. This is a remarkable amplifier with a single-ended output stage, having an output of 10 watts with a total harmonic distortion content of about 0ยท3 percent at 9 watts. It has three inputs-"tape", "radio" and "pick-up", and an output for connecting to a tape recorder. In addition, it has a comprehensive tone-control system, a four-position filter and the Pye "Dialomatic" pick-up compensation, which permits easy matching to any pick-up. The single-ended output stage is worthy of note, since the design follows an ultra-linear arrangement centered around a grain-orientated output transformer.
If distortion is well in evidence, and the tests already described eliminate the output stage, there are two general methods which can be adopted to locate the source of the distortion. The procedure, of course, applies to all amplifiers.
The oscilloscope, having been adjusted for distortion tests and aided by a distortion-free input signal from an oscillator or generator, can be moved from the output load to the control grid of each preceding valve in turn, working towards the input signal. For example, if distortion is present across the output load, the oscilloscope should be connected to the control grid of the output valve, the Y-gain adjusted accordingly, and the quality of the waveform noted. If distortion is still present, the signal should be monitored at the grid of V2b, then at the grid of V2a, and so on until a point is reached where the waveform is free from distortion.
Of course, the gain of the Y amplifier will need to be increased as the signal is traced towards the low-level sections of the amplifier. It is also important to avoid overloading the amplifier, and it is best to set the signal level to the point where distortion just occurs-it is assumed that this is well below the maximum power output of the amplifier.
Let us suppose that distortion is present at the grid of V2b, but not at the grid of V2a. It is obvious that the distortion is being produced by mis operation of V2a; a likely cause would be low emission of the valve section itself, though an increase in value of R9 or a leak in CI4 would also cause the trouble. Attention should also be paid to the components associated with the cathode circuit, these being related to the feedback network. In this way the signal can be traced back to its source and any deviation in wave-shape observed at each point of test.
If an oscilloscope is not available, an actual program signal can be applied to the amplifier by way of its appropriate channel, and the signal monitored at each grid in turn from a pair of headphones or earpiece. In order to avoid interference from the loudspeaker, the loudspeaker can be disconnected and its place taken by a suitable resistive load. The point at which the distortion occurs will quickly be traced by this method, and then the circuit section can be analyzed in detail.
Unfortunately, low-level distortion cannot usually be traced easily by this method, since headphones are rarely able to detect distortion at a level of, say, 5 percent. Indeed, one has to be a very critical listener to detect distortion at such low level by way of the loudspeaker-program material often contains distortion above this figure!
TONE CONTROL AND EQUALIZATION FAULTS
Maladjustment of the various tone controls, equalizers and filters is possibly responsible for the majority of reports of impaired performance at the high or low frequencies. Although the purist may be correct in the assumption that his amplifier has an absolutely flat response only when both the bass and treble controls are adjusted to the center of their range, the controls should, nevertheless, be varied from this ideal setting as a means of securing a better balance of sound in relation to the listening room and associated equipment. It is surprising how some enthusiasts are extremely reluctant to use tone controls for the purpose for which they were designed.
It is also possible, however, that the overall frequency response may be far from linear at the center setting of the controls, even though the designer may have intended a center balance. Slight alteration in value of components associated with the tone-control circuits may shift the "linear" point well towards the end of the range of one (or both) of the tone controls. This possibility should be suspected if there appears to be a boost of bass or treble when the controls are set to the center of their range. In extreme cases, it may be desirable to plot the frequency response of the amplifier to prove this point.
Maladjustment of the loudness control (if fitted) will incite excessive bass boost and possibly low-frequency distortion. If this trouble is suspected the loudness control should be switched out of circuit, or turned right off.
and the volume and tone controls then adjusted in the ordinary manner.
Finally, the loudness control can be brought back into circuit and adjusted for the correct level of sound, which will automatically give the correct degree of bass lift.
Some loudspeakers like more bass and/or treble than others, and the same applies to the listening room, as governed by the acoustics. It is quite in order to swing the tone controls over the whole of their range to get the "feeling" of the acoustics of the room and the response of the loudspeaker, after which the controls can be re-set more critically to give the results most pleasing to the listener, and most suitable for the program material.
Newcomers to hi-fi may be tempted to turn on too much bass or treble; this should be avoided. As G. A. Briggs points out, "if you notice the bass in the reproduction, or if the extreme 'top' is prominent, then there is something wrong because you do not notice bass and treble emphasis at a concert". Too little bass is sometimes caused by mis-phasing of the loudspeakers when two separate units are used on the same amplifier. Usually, the speakers are marked at their terminals with a blob of red paint or a positive and negative sign so that they can easily be connected together in correct phase.
When in parallel, the red terminals should be connected together (positive to positive and negative to negative); when in series, a positive terminal should be connected to a negative terminal, as when connecting batteries. If in doubt, a small cycle-lamp battery should be connected across the loudspeaker terminals and the resulting movement of the cone observed. The terminal of the loudspeaker which is connected to the positive tag on the battery to cause the cone to move, say, forward, should be clearly indicated with a blob of red paint.
Distortion of bass is invariably caused by core saturation of the output transformer. Such trouble is promoted by unbalanced output valves, causing a greater current in one half of the primary winding than in the other half.
The output stage should be checked for balance by the method already described.
Another cause of this trouble is low value of one of the output-valve coupling capacitors. Here the capacitive reactance of the defective component will be considerably below that of the good component at low frequencies, thus resulting in overdriving of one output valve with respect to the other.
In addition, the phase of the signal on the grid of one valve at the lower frequencies will deviate from that provided by the phase-splitter, and the phase difference will not be maintained at the ideal 180 deg. This will result in insufficient cancellation of harmonic distortion in the output load.
If the program material possesses an abundance of bass, the amplifier itself may be overloaded at the lower frequencies. Some amplifiers have a fixed high-pass filter to preclude this trouble, while others have a switched "rumble" filter to cut the bass at the extreme end of the spectrum, essentially to obviate transmission of gram motor rumble through the amplifier.
Troubles in the treble may be caused by faulty components in the tone control circuits, and this should first be suspected if the tone controls them selves appear not to be operating as they should. It should also be ensured that the equalization control is adjusted to suit the record being played.
Matching of the pick-up and the various program signals to the amplifier is most important if the correct response is to be maintained throughout the system. (Simple pick-up equalizers are considered in a later section.) It should be remembered that the response of certain loudspeakers is affected somewhat by the damping applied to them from the amplifier.
Insufficient damping-for example, by maladjustment of the damping control (if fitted)-will sometimes lead to a rise in the low-frequency resonance of the loudspeaker and an accompanying increase in the bass response. The bass in this case is of a purely synthetic nature.
HUM TROUBLES
Audio equipment is subject to two kinds of hum. There is the residual hum which is injected into the h.t. feed circuits as the result of a defective component associated with the smoothing and filter networks-this being synonymous with the hum experienced in radio receivers due to a breakdown of one of the electrolytic smoothing or filter capacitors. Then there is the hum caused by an alternating mains field being in proximity to the low-level stages of the amplifier. Here the radiated hum signals are picked up by the highly sensitive signal circuits, amplified by the equipment along with the required signal, and emitted by the loudspeaker in the characteristic manner.
Hum which is carried by the h.t. circuits usually presents little difficulty in remedying. The trouble is invariably caused by a reduction in value or open-circuit of one or more of the electrolytic filter capacitors. If the effect is present on a two-unit amplifier, the pre-amplifier should be disconnected from the power amplifier, and the residual hum level of the power amplifier noted. If the hum level is little different from that given when both units are connected, the power amplifier should receive attention.
If the hum is fairly loud, all the large-value capacitors associated with the h.t. supply should be checked either on a capacitor bridge or by substituting with good components. The connections on the capacitors should be examined and re-made if necessary, and if an electrolytic unit relies for negative connection upon clamp-contact with its case a check should be made to ensure that there is, in fact, a good low-resistance contact between the two points concerned.
As almost all hi-fi amplifiers use a full-wave h.t. rectifying circuit, residual h.t. hum will have a frequency twice that of the mains supply (100 hz in Great Britain and 120 hz in America), and it will also probably contain several harmonics of this frequency, thereby distinguishing it from normal 50 hz to 60 hz ripple.
Hum on the h.t. line can be traced with an oscilloscope or a.c. voltmeter isolated from the d.c. component by a paper capacitor. The hum level at the output of the h.t. rectifier should be noted, and then compared with the hum level at the other side of the smoothing choke, and so on through the filter chain. The hum reading should diminish considerably from section to section.
There is the possibility of a shorting turn in the smoothing choke in cases where the hum persists. If the main filter capacitors are in order, the a.c. reading across the choke should be approximately equal to that across the output of the rectifier; a lesser voltage should lead one to suspect choke trouble. Smaller amplifiers of the 10-watt rating often use a wire-wound resistor in place of a choke, and a test should be made to ascertain that this part is of the stipulated value.
Unbalance of the rectifier valve can also lead to excessive hum, as can a shorting turn in one half of the h.t. winding on the mains transformer; in the latter event, the transformer will overheat and emit wax or pitch.
If the hum is just about audible with the signal input socket shorted, connect a sensitive a.c. voltmeter or output meter across the loudspeaker to register the hum level and adjust the humdinger control for minimum reading.
If this does not reduce the level sufficiently, try adjusting the "balance" control, as unbalance of the output valves is another cause of high residual hum level.
If the hum becomes prominent only with the pre-amplifier connected to the power amplifier, impaired h.t. filtering in the pre-amplifier is a most likely cause, particularly if the hum is present with the volume control at zero. Electrolytic capacitors should be checked as before and if the pre amplifier has a separate humdinger control, this should be adjusted for minimum hum, as already described.
If the hum is not reduced, poor heater-to-cathode insulation in the final pre-amplifier valve may well be responsible. The best check is by valve replacement. The possibility of a hum voltage being induced into the pre amplifier/power amplifier connecting lead should also be considered, especially if the lead has been increased in length for any reason and if the output of the pre-amplifier is at high impedance. A low-impedance cathode follower output circuit is far less susceptible to such spurious pick-up.
FIG. 4.16. Circuit of input stage, showing common connecting points.
FIG. 4.17. A bus-bar taking all the "return" circuits minimizes
hum, providing it is connected at one point only to the chassis.
If the hum level increases as the volume control is advanced, one can be certain that the hum is getting into the stages preceding the volume control.
Make sure that it is not being induced into an open-circuit signal input socket by shorting the socket appropriate to the setting of the selector switch.
If the hum is still present, check all electrolytic capacitors, and all valves for heater-to-cathode insulation. Suspect hum pick-up from stray fields.
Induced hum has been dealt with in Section 2, but there are one or two additional points which are worthy of note. Having first ascertained that the program material is free from hum, and that hum is not being picked up on the program-source connecting leads, attention should be paid to such things as high-resistance connections between "earthing tags" and chassis, magnetic and electrostatic screens (including valve screens), misplaced grid or heater leads (particularly if the wiring has been disturbed during a previous servicing operation), the proximity of mains cables to grid circuits, etc.
It is surprising how much hum can be induced into an amplifier if it happens to be standing on the floor with the base cover removed, and if there is a mains cable running beneath the floor at this point! Even if the amplifier is lifted on to a table in similar proximity to the mains cable, the hum level may still be well above normal. Never run high-gain amplifiers with the screens removed, for it is remarkable how much a.c. mains field exists under domestic conditions. One can prod for hours trying to clear a slight hum which suddenly disappears on re-orientating the amplifier! To avoid hum voltages being introduced into a low-level stage from the "earthy" points of the circuit, a chassis connection common to the associated circuits is often adopted in commercial and home-built equipment. The idea is shown in the circuit in Fig. 4.16; apart from a common chassis point, it will be seen that there is also a common cathode point.
Owing to circulating alternating currents in the chassis itself, particularly if it carries the power transformer and smoothing choke, the common "earthing" device is often taken a stage farther. A heavy-gauge h.t. negative bus-bar is used for all the earth-return connections (Fig. 4.17), and this is "tied" to the chassis at one point only. In this way, there is no danger of the difference in a.c. mains potential, which may exist between two points on the chassis when circulating currents are present, being reflected back into the grid circuits of the low-level stages.
Hum troubles may also be experienced if the pre-amplifier and power amplifier are earthed separately as well as being connected together electric ally. As before, this results in a circulating alternating current, but this time in the loop between the two earth points and the conductor connection between the two units. The disturbance can be reduced considerably by disconnecting the earth from the pre-amplifier.
Whilst this condition may not be incited purposely, it may exist in slightly different form between, say, a pick-up and pre-amplifier, due to earthing at the motor as well as the amplifier, or between a f.m. tuner and pre-amplifier, in which case the tuner may be efficiently earthed in the normal manner while the earth point of the amplifier is connected to the earth tag of a three-pin power plug. In both cases a common impedance, in which is circulating small alternating currents at mains frequency, is developed between the signal source and the amplifier input. This presents to the amplifier a spurious 50 hz (60 hz in America) signal along with the program signal.
FIG. 4.18. The first-stage heaters can be energized by the rectified
h.t. current to reduce hum injection.
In very high-gain low-level stages the valve heaters are sometimes energized by direct current as a means of reducing the hum level. If the required d.c. is not obtained from a small l.t. rectifier in association with an l.t. winding on the mains transformer, the h.t. current is suitably adjusted and allowed to pass through the heater chain.
The basic idea is illustrated in Fig. 4.18. Instead of the center-tap of the h.t. winding on the mains transformer being connected direct to the chassis, as is usually the case, it is first passed through a variable resistance R and the valve heaters. It is assumed that the total h.t. current is more or less equal to the current required by the heaters, and any small discrepancy is corrected by the variable resistor. If the total amplifier h.t. current is, say, 90 mA, and the pre-amplifier valve heaters are rated at 100 mA, a h.t. bleeder resistor is connected across the h.t. circuit to pass the additional 10 mA, so that the total current flowing from chassis through the heaters into the center-tap matches the 100-mA valve rating.
Capacitor C serves to hold the circuit down to chassis at low frequencies, and also acts as a part of a filter when the negative voltage, relative to chassis, at the center-tap is used as a bias for the output valves.
This arrangement avoids having to run a.c. heater leads into the pre amplifier section where either capacitively or inductively they may inject a hum signal into the grid circuits. Normally, however, if all the basic pre cautions are observed, and the heater leads are twisted together to cancel hum fields, there is little need for d.c. operation these days, particularly with modern valves such as the EF86.
The practice of providing a slightly positive potential on the heater line, either from a decoupled potential-divider across the h.t. circuit or from the cathode of one of the output valves, is frequently adopted in modern equipment. This prevents the a.c.-modulated electrons emitted from the heater, at the point where it enters and leaves the cathode, from reaching the anode and causing hum. Since the heater is made more positive than the cathode, random emission of electrons from the heater section which is outside the cover of the cathode are attracted back to the heater, and thus do not con tribute to the normal electron stream.
The hum and noise levels are usually given as a composite figure in terms of decibels relative to the full output of the amplifier. Figures range from - 60 db to - 90 db; for example, the GEC BCS23 I 7/8 is approximately - 66 db relative to full output ( 12 watts), while the Pye H F25 is given as - 90 db on 25 watts. In neither case can the hum be heard.
The general "hiss" that a high-gain amplifier gives at full volume represents the noise output. As already mentioned, this is often referred to as "white noise", since it is not confined to any particular frequency, and is contributed mainly by the valves and resistors in the low-level stages.