▼ Menu

Wireless World PA Amplifier

W T Cocking, Wireless World, April 3 & 10, 1936.
    
Extras ▼

 

Twelve Watt High-Fidelity Equipment

The Power Amplifier.

There are many cases where an undistorted output of 12 Watts is necessary in sound reproduction, and an amplifier which combines this output with an exceedingly good frequency response is described in this article. Details are also given of a feeder-unit embodying a wide range tone-control circuit of novel design.

Although most domestic requirements are met by an amplifier having an undistorted output of 4-6 watts, there is no doubt that there are many cases where a larger output is needed. When rooms are exceptionally large or it is desired to operate several loud speakers together, an output of about 12 Watts becomes necessary, while an output of this order will often suffice for many public address requirements. The figure needed for PA work naturally varies considerably in different cases, but some 12 Watts will be found sufficient for most purposes, such as dancing in small halls, announcements at local functions, and so on.

The Wireless World PA Amplifier is based on the well-known Push-Pull Quality Amplifier, which has proved itself over a period of years to be reliable and trouble-free while giving practically perfect results from the electrical point of view. The circuit diagram of the amplifier appears in Fig. 1, and it will be seen that the output stage consists of two PP5/400 valves in push-pull operated in accordance with their maker's rating. The maximum undistorted output of 12 Watts is secured with a total load impedance of 6,000 Ω, and the stage requires a total input of about 64 Volts peak. This is too much to obtain safely from a single valve with resistance coupling, so that the penultimate stage is also of the push-pull type.

In order to prevent parasitic oscillation both grid and anode stopping resistances are employed in the output stage. In the anode circuits R13 and R14 are given the usual values of 100 Ω each, but the resistances R11 and R12 in the grid circuits have values of 1,000 Ω only, since it has been found that higher resistances lead to considerable attenuation of the upper frequencies with the output valves employed. This is because the PP5/400 has a much higher input capacity under operating conditions than the smaller PX4, principally because of its higher mutual conductance.

The LF Stages

The valves in the penultimate stage are of the MHL4 type, and the coupling resistances R5 and R6 have values of 25,000 Ω. Decoupling is provided by the 10,000 Ω resistances R7 and R8 in conjunction with the 8 μF. electrolytic capacitors C5 and C6. Grid bias is naturally derived from resistances in the cathode circuits, and R3 and R4 have values of 1,000 Ω and are shunted by the 250 μF capacitors C3 and C4. The grid leaks of the output stage R9 and R10 have values of 0.25 MΩ, and the coupling capacitors associated with them C7 and C8 have capacities of 0.1 μF. Since any leak in these capacitors would have a disastrous effect not only upon the performance, but also on the life of the output valves, these capacitors are of the mica-dielectric type, and are rated for working at 500 Volts. In the case of the coupling to the intermediate stage, a leak would not be attended by such serious results, and paper-type capacitors of 0.1 μF (C1 and C2) are accordingly employed at this point, in conjunction with 0.5 MΩ grid leaks R1 and R2.

The two output valves are operated from separate filament windings on the mains transformer; and grid bias is derived from the voltage drop across resistances connected between negative HT and centre-taps on these windings. These resistances R15 and R16 have values of 500 Ω, and are shunted by the 50 μF by-pass capacitors C9 and C10.

Turning now to the HT system, two entirely separate supplies are used - one for the output stage alone and the other for all early stages and speaker fields. Although at first sight wasteful, this course is not really so, for at most it involves an extra smoothing choke and a slightly more expensive mains transformer. In compensation, however, most of the smoothing capacitors can be of comparatively low voltage rating, no difficulties arise about voltage dropping nor about energising field windings of widely different characteristics, and, furthermore, feed-back effects from the output stage, their usual source, are entirely eliminated.

Fig. 1. - The complete circuit diagram of the amplifier. Two separate rectifiers and smoothing systems are used in the HT supply.

Referring to Fig. 1, it can be seen that the 500-0-500 Volts 120 mA winding on the mains transformer supplies the HT for the output stage in conjunction with the 460BU (UU5) rectifier valve which has its filament heated from one of the 4 Volt 2.5 Amp windings. The reservoir capacitor C12 has a capacity of 4 μF, and is rated for 1,000 Volts working. Smoothing is effected by the 12 H choke Ch1in conjunction with another 1,000 Volts capacitor C11 of 4 μF capacity.

The total voltage required by the output stage is some 400 Volts for other anodes, 32 Volts for grid bias, 6 Volts loss in R13 and R14, and, say, 9 Volts loss in the output transformer, or roughly 450 Volts. The un-smoothed output of the rectifier is some 510 Volts, and the requisite voltage drop is obtained partly in the DC resistance of the smoothing choke (200 Ω) and partly in the 300 Ω resistance R17. The 0.5 MΩ resistance R18 is included to prevent their being any possibility of the capacitors retaining a charge when the amplifier is switched off.

Now the second HT supply is obtained from the 350-0-350 Volts winding and the MU12 indirectly heated rectifier. The reservoir capacitor C15 has a capacity of 4 μF, and is of the electrolytic type. Preliminary smoothing is effected by means of a 12 H choke Ch3 and an 8 μF capacitor C14, and it is completed by the 36 H choke Ch2 and another 8 μF capacitor C13. When a total current of 120 mA. is drawn a potential of some 250 Volts appears across this last capacitor and constitutes the HT supply for the early stages. As will be shown later, speaker fields can be energised by being connected across the HT supply at this point.

The Feeder-Unit

Fig. 2. - A fader-type volume control is used in the feeder-unit, while a wide-range tone control is included. This permits the bass response to be raised and the treble response to be raised or lowered at will.

The amplifier proper, with its mains equipment, can be used with any feeder-unit or receiver which is designed for the Push-Pull Quality Amplifier without alteration to either. A special feeder-unit has been designed for it, however, since in PA work a microphone is commonly employed, and, moreover, a wide range tone-control is often needed. The use of this unit, however, is not confined to this amplifier, and it can equally well be used with the PPQA when the smaller output will suffice.

The circuit appears in Fig. 2, and three valves are used; of these one is a phase-changer, another an amplifier (41MH), and a third a part of the tone-control system. The amplifier and phase-changer are straight-forward; D4 type valves are used and bias is obtained by the 2,000 Ω cathode resistances R26 and R31 shunted by the 50 μF capacitors C23 and C26. In the anode circuit of the amplifier a 50,000 Ω coupling resistance R28 is used with a 50,000 Ω decoupling resistance R27 and 8 μF decoupling capacitor C24. A 0.1 μF coupling capacitor C25 to the phase-changer is employed and a 2 MΩ grid leak R29. This next valve has 50,000 Ω coupling resistances R32 and R30 in anode and cathode circuits respectively, and the AC voltages developed across these are used to feed the push-pull amplifier. Anode circuit decoupling is provided by the 50,000 Ω resistance R33 and the 8 μF capacitor C27.

The phase-changer gives an effective gain of about 1.8 times, and the amplifier preceding it a gain of roughly twenty-five times, or a total of about forty-five times. This is nearly enough, and the amplification given by the first valve and its coupling is only two or three times. This first valve is fitted primarily to make up for the loss introduced by the tone-control circuit.

A view of the feeder-unit showing the controls.

The tone-control normally fitted to receivers and amplifiers merely permits the response at the upper audible frequencies to be reduced when required. This is by no means an ideal arrangement, but it is the simplest method. A better system would undoubtedly be an arrangement whereby the bass and treble could be independently controlled and raised or lowered as required.

A reduction in the bass response is not often needed, however, so in this amplifier arrangements are made for it to be lifted only. The treble can be raised or lowered, however. Referring to Fig. 2, it will be seen that the first valve is conventionally connected in that a 50,000 Ω coupling resistance R21 is used with a 50,000 Ω decoupling resistance R22 and 8 μF decoupling capacitor C18. Grid bias is obtained by means of the 2,000 Ω resistance R20 shunted by the 50 μF capacitor C16, while two input plugs are provided with a fader-volume control R19. This permits a microphone and pick-up to be permanently connected and a rapid change-over from one to the other to be made.

The tone-control circuit comes in the coupling between the two valves. Consider the state of affairs when all switches are in their centre positions. The 3,000 Ω resistance R24 is then connected to the earth line, and it forms a potentiometer with the 50,000 Ω resistance R23, so that only 3/53 rds. of the voltage across R21 is applied to the second valve. This valve is provided with a grid leak R25 of 0.5 MΩ in order to prevent an open grid circuit being obtained at certain settings of the switches. In view of the low effective value of the circuit resistances the coupling capacitor C17 has the large capacity of 0.5 μF.

The feeder unit seen from below.

Now, in these switch positions the frequency response is normal, and the two resistances R23 and R24 merely throw away most of the amplification given by the first valve. When S1 is moved to the next stud, however, a capacitor C20 of 0.25 μF is interposed between R24 and the earth line. As long as the reactance of this capacitor is small compared with 3,000 Ω it has negligible effect and the response at medium and high frequencies remains normal. At low frequencies, however, the reactance is no longer negligible, for it increases in a manner inversely proportional to frequency. The total impedance between the grid of the second valve and the earth line consequently rises, and a greater proportion of the voltage developed across R21 is applied to the second valve. Actually, an increase at 50 Hz of some 10 dB is obtained. On the next stud of the switch a smaller capacitor, C19, of 0.1 μF capacity, is inserted, and a greater rise in bass is secured. The rise at 50 Hz is then about 20 dB.

A similar arrangement is used at high frequencies, but here a choke is inserted. When S2 is fully rotated in a clockwise direction (on the feeder unit) the 0.54 H. coil L is inserted in series with R24, and a rise of some 20 dB at 10,000 Hz is obtained. The next position of the switch connects R24 to the tapping point on L and the inductance in use is 0.18 H, giving a rise of some 10 dB. The switch S3 gives a control reducing the response at high frequencies, again in steps of about 10 dB, by shunting R24 by the capacitors C21 or C22 of 0.015 μF and 0.05 μF respectively.

The two switches S2 and S3 are in practice on a common shaft and form the treble control.

When the control knob is fully rotated in anti-clockwise direction the response at 10,000 Hz is about -20 dB/in the next position it is some -10 dB, while in the centre position a flat characteristic is obtained. The two further positions give responses of +10 dB and +20 dB respectively. In the case of the bass control S1, the first three positions are the same, and the two last (clockwise) give an increased response at 50 Hz of about +10 dB and +20 dB respectively.

If it should be required to obtain a reduced bass response this may readily be done by arranging the unused half of the switch assembly to connect additional capacitors in series with C17. The assembly of the components, wiring and operation of the amplifier and its feeder-unit will be fully dealt with in next week's issue.

Construction and Operation

Fig. 3. - The overall frequency response curve of the amplifier and feeder-unit is shown by the heavy line, while the dotted curves indicate the effect of the tone-control system.

The tone-control circuit fitted to the feeder-unit permits the bass response to be raised and the treble response to be raised or lowered at will. The results obtainable are shown by the curves of Fig. 3, in which the full-line curve is the normal overall response of the amplifier with its feeder unit. At as low at frequency as 20 Hz, the response falls by 2.4 dB only as compared with that at 400 Hz, while at 10,000 Hz it is actually +0.8 dB. The dotted curves show the effect of the tone control, curve A giving the response when both bass and treble controls are set for full lift, and curve B when they are set for reduced lift. At 50 Hz, a maximum lift of 20 dB is obtained and in the intermediate position an increase of 13 dB while at 10,000 Hz the figures are 24 dB. and 14 dB. Curves C and D show the effect of the treble control in reducing the upper-register, with the bass control set for normal response. It should be clearly understood that according to the precise positions in which the switches are set any of the five different curves in the treble region can be obtained in conjunction with any of the three curves in the bass region. Thus no fewer than fifteen different response curves are obtainable at will.

In general, for high-quality reproduction when the apparatus is used in conjunction with a receiver the normal response curve should be used, but if the receiver gives any appreciable degree of sideband cutting the treble should be increased. On gramophone, however, with normal recordings and pick-ups, the bass response should be increased to compensate for the deficiencies of recording; the treble response should be flat or reduced to avoid needle scratch.

When the amplifier is used with recording apparatus, the ability to increase the treble will be found extremely useful. In most cases, the best results will be secured by recording with a flat bass characteristic and the maximum rise in the treble. In reproducing records cut in this way, a rising bass characteristic is needed and a falling response in the treble in order that the overall response shall be flat. The reproduction of the various frequencies is then as good as if a flat characteristic were used both for recording and reproducing, but there is a marked diminution of needle scratch.

With the switches specified, the tone controls will be found to be quite silent in action and to have no appreciable effect on the volume. It is important, however, to note that it is necessary to take great care to avoid amplitude distortion being caused by any apparatus, such as a pick-up, which precedes the feeder unit, if the ability to increase the treble response is to be any advantage. This is because if the input is distorted, the rising characteristic increases the percentage of harmonics. Thus, suppose that the input waveform is at 1,000 Hz and contains 1% of both second and third harmonics, and that the full treble lift is being used. The second harmonic in the output will become 1.7% and the third harmonic 2.45%. If the rising characteristic is used to correct for a falling characteristic in the loud speaker, of course, there will be no increase in the harmonic content of the actual sound output, but if it is used to correct for sideband cutting in a receiver, and the distortion is due to the detector, then it will certainly appear. The indiscriminate use of a rising characteristic is thus to be avoided, but there are many occasions when it can be usefully employed.

Fig. 4. - The connections to the input plugs of a pick-up and a microphone are shown here.

The actual construction of the amplifier is entirely straightforward and needs no description beyond the remark that the high-voltage smoothing capacitors have their terminals protected by a strip of Presspahn or other convenient insulating material. The uses of the equipment are also obvious. The input to the feeder-unit should be by means of screened leads connected as in Fig. 4, a microphone being preferably energised by a battery to avoid hum troubles. The connections between the feeder unit and amplifier can be made with twisted flex and can be up to two feet in length, for the two units should not be immediately adjacent. Should it be found that hum appears when the tone control is set to lift the treble, it is due to pick-up by L from the mains transformer, and it can be avoided by moving the feeder-unit.

Fig. 5. - The connections for single 1,250 Ω and 2,500 Ω field windings are shown at (a) and (b) respectively, while the method of energising four 2,500 Ω fields is indicated at (c).

Turning now to the loud speaker, a single 5,000 Ω 60 mA field or two in parallel can be connected directly to the 'field' terminals provided, as also can a single 2,500 Ω 120 mA field, or two series-connected 1,250 Ω 120 mA fields. A single field winding of this last type must be connected with a 1,250 Ω resistance of at least 20 Watts rating, as shown at Fig. 5 (a), while a single 2,500 Ω 60 mA field should be connected as at (b) with a resistance of not less than 10 Watts rating. As many as four fields of this rating can be energised by connecting them in series-parallel as in It should be noted that 120 mA can only be drawn for field current when the amplifier is used with the feeder-unit. When current for working a receiver is taken from the equipment, it will usually be possible to draw only 60 mA for field windings. It will be noted that even when the feeder-unit only is used, the total current drawn from the rectifier exceeds 120 mA when this current is taken for the field supply; The overload is not great, however, and should not lead to a material reduction in the life of the rectifier.

Fig. 6. - The primary of the output transformer should be joined to the speaker-plug in the way shown here.

The output stage requires a load impedance of 6,000 Ω obtained by means of a transformer connected as in Fig.6. The ratio of the transformer can be calculated by dividing 6,000 by the speech-coil impedance and taking the square root of the result. The transformer should, of course, be of high quality with a good frequency response and capable of handling a power of 12 Watts without introducing distortion. A primary inductance of about 40 H is adequate.

In cases where several loud speakers are used the speech coils can all be connected in parallel if they are all of the same type. The effective speech-coil impedance to be used in calculating the output transformer ratio is then the figure for one coil divided by the number of speakers. Where dissimilar loud speakers are used, however, each must be provided with its own transformer. The figure of load impedance used in calculating the ratios should then be multiplied by the number of speakers.

Suppose, for instance, that two 15 Ω speakers are being used. The speech coils can be paralleled; and the combined impedance is7.5 Ω, so that the transformer should have a ratio of √6,000/7.5 = 28.25:1 instead of the 20:1 ratio which would be needed for a single speaker. If one speaker of 15 Ω is used with another of 5 Ω, then two transformers are needed. The 15 Ω speaker requires one having a ratio of √12,000/15 = 28.25:1, while the 5 Ω one needs a ratio of √12,111/5 = 49:1. The two transformer primaries should be connected in parallel and, strictly speaking, each should have an inductance of twice the normal figure, or 80 H. This is likely to be impracticable, however, on account of the leakage inductance, and a compromise of about 60 H will generally lead to the best results.

The placing of the output transformer is not important. In general, with a speaker having a speech-coil impedance up to about 15-20 Ω, it should be placed with the speaker and any extension leads run in the transformer primary circuit. With a higher impedance speech coil, however, it is perfectly practicable to run the leads in the secondary circuit, provided that heavy conductors are used, and the transformer can then be mounted with the amplifier.

List of components, chassis and wiring diagrams can be found here.

The amplifier with the mains equipment is built on a steel chassis.

Use browser back button to return.
art-481