You are on page 1of 75

SIMPLE AUDIO CIRCUITS

Part 1 Introduction and Power Amplifiers

Constructional Project

RAYMOND HAIGH
A selection of pic-n-mix low-cost audio circuits from preamplifier to speaker!
HOW MUCH POWER?

their particular area of interest, most electronics enthusiasts encounter the need to amplify and reproduce audio signals. The final stages of radio receivers, intercom units, security and surveillance installations, or just a hankering for a big sound from a Walkman or portable CD player, all involve audio amplification and a speaker system. And the amplification usually goes hand-in-hand with some form of signal processing. Music reproduction calls for a wide frequency response and tone-control circuitry. Speech communication, especially
HATEVER

under difficult conditions, is greatly clarified if the frequency response is curtailed. This short series of articles describes simple, but effective, ways of meeting these different requirements. Although the circuits are capable of a good standard of reproduction, they will particularly interest the constructor who looks for plenty of performance per pound or dollar. We begin with the power amplifier. Six alternatives are given and, with the requirement of maximum performance for minimum cost and effort, they are all based on widely available integrated circuits (i.c.s): * TDA7052 * TBA820M

Before considering the various circuits, it is worthwhile to reflect on the amount of power actually needed. Clear reproduction in Walkman type earphones of reasonable sensitivity can be achieved with a miserly milliwatt (0001W). When listening to talk programmes in a quiet domestic setting, the power fed to the speaker will hover around 50mW (005W), and this is ample for the operator of a communications receiver whose ears are closer to the sound source. During the valve era, ten watts was considered adequate for the realistic reproduction of orchestral music, and some experts suggested a figure as low as five. One watt of power delivered to a * TDA2003.

* LM386N-1

* LM380N

Collection of simple i.c. power amplifier modules (left-to-right, top-to-bottom). * Single TDA2003 Amp * Twin TDA2003 Amp * TDA7052 Amp * TBA820M Amp * LM380N Amp * LM386N-1 Amp

340

Everyday Practical Electronics, May 2002

reasonably efficient speaker will produce a loud sound, a continuous five watts is becoming deafening, and ten will rattle windows. This assumes single channel, or mono, reproduction in a normal living room. The impact is, of course, greater with a stereo system. This is at odds with the high power ratings of many of the quality amplifiers currently advertised. The desire for a big reserve of power, the low efficiency of some modern speakers, and different ways of measuring output, may account for the difference. Output is variously rated as music power, sustained music, speech and music, and sinewave power. To add to the confusion, the figures are quoted at different distortion levels. The standard most often used when valves were commonplace, and the one adopted in this article, is the r.m.s. (root mean square) value of a continuous sinewave. This gives the lowest rating and is the most realistic expression of the amplifiers ability to deliver power into a load (speaker). An increase in output power is reflected as much, if not more, in the cost of the power supply as it is in the actual amplifier. Because the theme of this article is good performance at modest cost, the most powerful amplifier described is rated at 125W r.m.s.

Provided a few basic precautions are observed, the amplifiers are all unconditionally stable. Most i.c.s of this kind have a ground connection for the input circuitry and a separate ground pin for the output stage. The printed circuit board (p.c.b.) layouts have been designed to maintain this isolation, and care should be taken to ground the signal inputs and connect the negative power supply lead to the designated points on the board. Failure to do this could result in motor boating (low frequency instability). Input leads should be screened to avoid mains hum and radio frequency (r.f.) signal pick up. Speaker leads should be twisted together to minimise external fields. Input and output leads should be spaced as far apart as possible: this is particularly important when the LM386N and TBA820M are set for high gain. All of the circuits include high and low frequency bypass capacitors across the supply rails. The former minimise the possibility of r.f. oscillation: the latter avoid low frequency instability when long power supply leads are used, or when batteries are ageing.

STABILITY

overloads them, causing distortion and loss of clarity. Indeed, if the amplifier is being used primarily for speech, clarity can be much improved by rolling-off the frequency response below 300Hz, and an even lower value of coupling capacitor, say 100mF or even 47mF, would be of benefit. Readers seeking quality music reproduction at low power, via a speaker of reasonable size, should increase the coupling capacitor to say 1000mF. This relationship between coupling capacitors and frequency response will be considered more fully in the next article.

DEVICE PROTECTION

The integrated circuits covered here are electrically robust but they are by no means indestructible. The TDA7052, LM380, and the TDA2003 incorporate protection against overload and output short ciruits: the other devices do not.

HIGH FREQUENCY RESPONSE

The bandwidth of the amplifiers extends into the r.f. spectrum, and this makes the devices vulnerable

DISTORTION

Manufacturers of power-amplifier integrated circuits and modestly priced hi-fi systems (which invariably incorporate devices of this kind) usually rate the maximum power output at 10 per cent distortion. At this level there is a very noticeable roughness to the sound and clipping of the waveform on loud passages. The power output levels quoted here have been measured just before the onset of clipping or any noticeable distortion of the output waveform. They are somewhat lower than the figures quoted by the i.c. manufacturers, but they do represent the highest output, free from audible distortion, that the device can deliver for a particular supply voltage and load.

r. fie pli m ra However, we o p even with this proge tection, excessive supid br ply voltage will result in 03 0 immediate failure, and shorting A2 TD the output when the amplifier is in being driven hard and/or when the supTw ply voltage is close to the operational maximum will quickly ruin the device.

SUPPLY VOLTAGE
to r.f. interference. Some of the i.c.s provide for the connection of an external capacitor in a negative feedback loop to roll-off the high frequency response. Selecting an appropriate value for this component will help to make the device immune. The problem of r.f. pick up invariably manifests itself when a high value (more than 10 kilohms) input potentiometer (VR1) is used to match the amplifier to the impedance of a signal source. If the potentiometer or volume control must have a high resistance, connecting a 1nF or, at most, 10nF capacitor across its track will shunt unwanted r.f. to ground.

NOISE

Modern power amplifier i.c.s have a very low noise level. Manufacturers usually define this internally generated electrical noise as an equivalent signal voltage at the input, but this doesnt give the average experimenter an immediate impression of its audible effect. Accordingly, the devices described here were tested by disconnecting the signal source, turning the input or volume control to maximum, and then listening to the output on a pair of sensitive, Walkman type earphones. In all cases the noise was no more than barely audible. The two devices which can be configured for high gain (LM386N and TBA820M) did produce a faint, but audible, hiss when the gain was set at maximum. The hiss was also noticeable with a loudspeaker connected. However, when the gain preset was turned back a little, these i.c.s became as silent as the rest. Some constructors may need the highest possible gain, and details will be given later of measures which can be taken to eliminate the noise.

Suitable mains power supplies will be covered in a later article, but it should be mentioned now that, off-load, d.c. output voltages rise to 14 times the a.c. voltage delivered by transformer secondaries. When using unregulated mains power supplies care should, therefore, be taken to ensure that the off-load voltage is always less than the maximum safe working voltage of the amplifier. Never connect a working power supply to an amplifier without first checking its output voltage.

ELECTRICAL CHARACTERISTICS

LOW FREQUENCY RESPONSE

The low frequency response of three of the lower powered amplifiers has been curtailed a little by fitting a 220mF speaker coupling capacitor. Amplifiers of this kind are invariably used with small, inexpensive speakers which are incapable of producing an audible output at frequencies below 150Hz or so. Feeding low frequencies to speakers of this kind only

The electrical characteristics of the various devices are tabulated alongside the circuit diagrams (except one) for easy reference. Power output figures are based on measurements taken on a single, randomly purchased sample. For reasons already given, they are somewhat lower than the figures quoted by the manufacturers. Recommendations are made regarding the speaker impedances to use with various supply voltages in order to keep the dissipation of the devices within reasonable limits. The input resistance, maximum voltage ratings, and frequency response details are those supplied by the manufacturers.

Everyday Practical Electronics, May 2002

341

AMPLIFIER PROJECTS
In use, there is little to distinguish between the four, low powered amplifiers, all perform well. There are, however, differences which make one device more suitable than another for a particular application. Low current consumption is important when equipment is powered from dry batteries. Quiescent current drawn by the small amplifiers is in the region of 6mA (13mA for the LM380).

In the case of the LM386N, TBA820M and LM380, current rises to around 120mA when 500mW is being delivered into an 8 ohm load. Current consumed by the TDA7052 is approximately 220mA, or almost double, under these conditions. In all cases, the signal input pin has been connected to the slider (moving contact) of the Volume control potentiometer (via a blocking capacitor in the case of the TDA2003). This minimises hum and noise and ensures that a more or less constant impedance is presented to the signal

source. Potentiometers of 4700 ohms or 10 kilohms (10k) are usual, but the value can be increased to 100k to raise input impedance. This will, however, make the circuits more vulnerable to mains hum, r.f. interference and instability, and the value should be kept as low as the signal source impedance permits. This applies particularly to the TDA7052, where the value of the Volume control should, if possible, be no more than 10k. Earlier comments regarding stability are of relevance here.

LM386N-1 AMPLIFIER
A circuit diagram for a simple amplifier using the low-voltage LM386N-1 power amplifier i.c. is shown in Fig.1. Also shown are the general performance and electrical characteristics of the circuit. Blocking capacitor C1 prevents any disturbance of the d.c. conditions in the signal source and potentiometer VR1 (the Volume control) sets the input level. The manufacturers of the chip, National Semiconductor, suggest an input network to roll-off high frequencies and resistor R1 and capacitor C2 perform this function. The unused non-inverting input (pin 3) is grounded to avoid instability when gain is set high. Capacitors C3 and C4, connected across the supply rails, prevent low and high frequency instability.

CIRCUIT BOARD

The printed circuit board component layout, wiring details and full-size copper foil master pattern are shown in Fig.2. This board is available from the EPE PCB Service, code 343 (LM386N-1).

+
SIGNAL INPUT

C3 220

+3V TO +12V
C4 100n VR2 10k

+
C1 47
W

+
1
C5 10

2
R1 470 3

FEEDBACK

IC1
LM386N-1

An internal negative feedback path can be accessed via pin 1 and pin 8. Bypass capacitor C5 reduces the feedback and increases the gain of the chip from 23 to 170 times (as measured: samples will vary). Preset potentiometer VR2 (wired as a variable resistor) controls the bypassing effect of C5 and enables the gain to be set within these limits. Bypass capacitor C6 makes the device more immune to supply line ripple, and C8 couples the output to the speaker LS1. The Zobel network, formed by resistor R2 and capacitor C7, ensures that the speaker always presents a resistive load to the amplifier. Without these components there is a risk of high level transients causing damage to the output transistors. Tabulated power output levels for various supply voltages and speaker impedances are included below the circuit diagram. Sustained operation at more than 300mW is not recommended.

VR1 10k

+
4

R2 10
LS1 8 C7 47n

VOLUME

SCREEN

C2 1n

C6 10

C8 220

0V

SET GAIN INPUT INPUT +

1 2 3

8 7 6 5

SET GAIN SUPPLY VOLTAGE RIPPLE REJECTION SUPPLY VOLTAGE +V SIGNAL OUT

GROUND (0V) 4

TOP VIEW OF LM386N-1

Fig.1. Circuit diagram and pinout details for the LM386N-1 Power Amplifier.

LM386N-1 POWER AMPLIFIER


R.M.S. power output just before the onset of waveform clipping
Speaker Impedance ohms 3V
4 8 16 32 60mW 26mW 15mW -

Supply Voltage 45V


150mW 105mW 60mW 35mW

6V
320mW 200mW 110mW 62mW

9V
500mW 560mW 320mW 170mW 6mA 50k ohms

12V
900mW 605mW 330mW

Completed LM386N-1 circuit board.

Quiescent current: Input resistance: Input sensitivity for 560mW output (8 ohm load, 9V supply), (a) VR2 set for maximum resistance: (b) VR2 set for minimum resistance: Absolute maximum supply voltage, beyond which damage will occur: Suggested maximum supply voltage with a 4 ohm speaker Frequency response

90mV r.m.s. (gain 23) 12mV r.m.s. (gain 170) 15V 6V up to 300kHz

342

Everyday Practical Electronics, May 2002

POWER SUPPLY +VE SCREENED INPUT LEAD

VR2 C5 C3

COMPONENTS
LM386N-1 AMPLIFIER
Resistors
R1 470W All 025W 5% carbon film R2 10W

VOLUME

+
W

+
C4

C1

See

R1 C2

IC1
R2

C7

Potentiometers
+
C8

VR1 (FRONT VIEW)

VR1 VR2

10k min. rotary carbon, log. 10k enclosed carbon preset 4m7 radial elect. 25V 1n disc ceramic 220m radial elect. 25V (2 off) 100n disc ceramic 10m radial elect. 25V (2 off) 47n polyester LM386N-1 audio power amp i.c. 4 to 32 ohm loudspeaker (see text)

SHOP TALK
page

C6

Capacitors
(0V) POWER SUPPLY VE 2.0IN (50.8mm) TO SPEAKER (LS1)

C1 C2 C3, C8 C4 C5, C6 C7 IC1


1.6IN (40.6mm)

343

Semiconductor Miscellaneous
LS1 Printed circuit board available from the EPE PCB Service, code 343 (LM386N-1); case (optional), size and type to choice; 8-pin d.i.l. socket; multistrand connecting wire; audio screened cable; solder pins; solder etc.

Approx. Cost Guidance Only Fig.2. Printed circuit board, component layout, full-size copper foil master and interwiring for the LM386N-1 Amp.

excluding case & speaker

10.50

TDA7052 AMPLIFIER
Philips have adopted a bridge arrangement for the TDA7052s output stage. This enables the chip to maintain a good output at low supply voltages and eliminates the need for a speaker coupling capacitor. Gain is fixed internally, no provision is made for ripple rejection, and there is no Zobel network. This reduces the external component count to the d.c. blocking capacitor C1, Volume control VR1 and the supply line bypass capacitors, C2 and C3. The full circuit diagram, together with a specification guide, for the TDA7052 amplifier is shown in Fig.3. Protection against output short circuits is built in and the device shuts down when the dissipation becomes excessive. This explains the small rise in sustainable output when the speaker impedance is increased to 16 ohms with a 9V supply. Although usually costing a little more than the other low-power chips, this is the device of choice when the supply voltage has to be low, a good output is required, and high gain is not important. Current consumption for a given output power is, however, almost twice that of the LM386N and the TBA820M.

CIRCUIT BOARD

The printed circuit board component layout, wiring details and full-size copper foil master pattern are shown in Fig.4. This board is available from the EPE PCB Service, code 344 (TDA7052).

TDA7052 POWER AMPLIFIER


R.M.S. Power output just before the onset of waveform clipping
Speaker Impedance Ohms 3V
4 8 16 32 70mW 60mW 40mW 24mW
SIGNAL INPUT

1 2 5

+3V TO +12V
LS1 8

Supply Voltage 45V


500mW 455mW 235mW 145mW

IC1
+
TDA7052
3 8 6

6V
780mW 640mW 450mW 250mW

9V
1W 112W 600mW 5mA 100k ohms

12V
126W

C1 10

VR1 4k7

C2 100n C3 220

SCREEN

VOLUME

0V

Quiescent current Input resistance Input sensitivity for 1W output (8 ohm load, 9V supply) Absolute maximum supply voltage beyond which damage will occur Suggested maximum supply voltage: with a 4 ohm speaker with 8 or 16 ohm speakers Frequency response at the 3dB points

SUPPLY VOLTAGE +V INPUT INPUT GROUND

1 2 3 4

8 7 6 5

SIGNAL OUT NOT CONNECTED OUTPUT GROUND SIGNAL OUT

40mV r.m.s. (gain 70) 18V 6V 9V 25Hz 20kHz

NOT CONNECTED

TOP VIEW OF TDA7052

Fig.3. Circuit diagram and pinout details for the TDA7052 Amp. See left for performance guide.

Everyday Practical Electronics, May 2002

343

COMPONENTS
TDA7052 AMPLIFIER
Capacitors
C1 C2 C3 10m radial elect. 25V 100n disc ceramic 220m radial elect. 25V

See

SHOP TALK
page

Only four components are mounted on the TDA7052 p.c.b. Fig.4 (below). Component layout, interwiring and full-size copper foil master for the TDA7052 Amp.
1.2IN (30.5mm) SCREENED INPUT LEAD

Potentiometers
VR1 4k7 min. rotary carbon, log. TDA7052 power amp i.c. 4 to 32 ohm loudspeaker (see text)

Semiconductor
IC1 LS1

Miscellaneous

Printed circuit board available from the EPE PCB Service, code 344 (TDA7052); case (optional), size and type to choice; 8pin d.i.l. socket; multistrand connecting wire; audio screened cable; solder pins; solder etc.

VOLUME

C1

+
C2

C3

POWER SUPPLY +VE POWER SUPPLY VE (0V)

IC1
VR1 (FRONT VIEW)

TO SPEAKER (LS1)

Approx. Cost Guidance Only excluding case & speaker

9.50

344

TABA820M AMPLIFIER circuit diagram incorporating the TBA820M audio amp i.c.,
which is manufactured by SGS-Thomson, together with a general performance guide, is given in Fig.5. The input arrangements, supply line bypassing, speaker coupling and Zobel network are conventional, and the relevant components can be identified from previous circuit descriptions. Gain can be controlled by shunting an internal negative feedback loop, which is accessed at pin 2. Preset potentiometer VR2, placed in

+
C2 220

+3V TO +12V
C3 100n

Completed TBA820M amplifier module.

+
C5 47

LS1 8 C8 220

6
SIGNAL INPUT

TBA820M POWER AMPLIFIER


R.M.S. power output just before the onset of waveform clipping
Speaker Impedance Ohms 4 8 16 32 Supply Voltage 3V 10mW 20mW 30mW 20mW 45V 320mW 200mW 115mW 60mW 6V 405mW 300mW 180mW 90mW 9V 980mW 680mW 405mW 225mW
6mA 5M ohms

8
3

7 5 1 C6 390p

IC1
TBA820M 4 2

+
C1 47

VR1 10k

R1 22

VOLUME

VR2 100

R2 1

12V 11W 720mW 390mW

+
SCREEN
C4 100 C7 220n

0V

SET HIGH FREQUENCY RESPONSE SET GAIN INPUT GROUND (0V)

1 2 3 4

8 7 6 5

SUPPLY VOLTAGE RIPPLE REJECTION BOOTSTRAP SUPPLY VOLTAGE +V SIGNAL OUT

TOP VIEW OF TBA820M

Fig.5. Circuit diagram and pinout details for the TBA820M power amplifier. See right for general performance details.

Quiescent current Input resistance Input sensitivity for 680mW output (8 ohm load, 9V supply): (a) VR2 set for maximum resistance (b) VR2 set for minimum resistance Absolute maximum supply voltage beyond which damage will occur Suggested maximum supply voltage: with a 4 ohm speaker with an 8 ohm speaker High frequency response at 3dB point: with C6 220pF with C6 680pF

56mV r.m.s. (gain 40) 10mV r.m.s. (gain 230) 16V 9V 12V 20kHz 7kHz

344

Everyday Practical Electronics, May 2002

1.6IN (40.6mm)

COMPONENTS
TBA820M AMPLIFIER
Resistors
R1 22W R2 1W All 025W 5% carbon film

SCREENED INPUT LEAD

+
VOLUME C1

C6

See

C5 W R 1

IC1
C3

+
R 2

+
C8

SHOP TALK
page

TO SPEAKER (LS1)

VR1 (FRONT VIEW) C4

C2

POWER SUPPLY +VE

Potentiometers
VR1 VR2

+
C7

10k min. rotary carbon, log. 100W enclosed carbon preset 4m7 radial elect. 25V 220m radial elect. 25V (2 off) 100n disc ceramic 100m radial elect. 25V 47m radial elect. 25V 390p ceramic 220n polyester TBA820M audio power amp i.c. 4 to 32 ohm loudspeaker (see text)

POWER SUPPLY VE (0V)

VR2

2.4IN (61.0mm)

Capacitors
C1 C2, C8 C3, C4 C5 C6 C7 IC1

Semiconductor

345

Miscellaneous
LS1

Fig.6. Topside component layout, off-board wiring details and full-size copper foil master for the TBA820M Amplifier.
series with capacitor C4, controls the shunting effect and, with the sample tested, gain could be set between 40 and 230. High frequency response is determined by capacitor C6. The response at the 3dB points for different capacitor values is also listed in the table. If desired, the value of C6 can be increased to reduce the upper frequency response even more. In this application, the speaker LS1 is connected to the positive supply rail as this reduces the component count (a capacitor and resistor are saved).

Printed circuit board available from the EPE PCB Service, code 345 (TBA820M); case (optional), size and type to choice; 8pin d.i.l. socket; multistrand connecting wire; audio screened cable; solder pins; solder etc.

CIRCUIT BOARD

Approx. Cost Guidance Only excluding case & speaker

10

The printed circuit board component layout, wiring details and full-size copper foil master pattern are shown in Fig.6. This board is available from the EPE PCB Service, code 345 (TBA820M).

LM380N AMPLIFIER
An amplifier circuit diagram incorporating the LM380N audio i.c. is shown in Fig.7. A general specification guide is shown below. The manufacturers, National Semiconductor, have made provision for optional heatsinking via some of the pins, and this makes the package larger (14-pin). The output is short-circuit proofed and there is dissipation limiting. Gain is fixed. Again, the purpose of the various components will be evident from earlier descriptions. In this version of the circuit, the signal

LM380N POWER AMPLIFIER


R.M.S. power output just before the onset of waveform clipping
Speaker Impedance Ohms
4 8 16 32

Supply Voltage 9V
400mW 275mW 137mW 68mW

12V
112W 720mW 405mW 202mW

15V
162W 132W 720mW 360mW

18V
225W 132W 765mW

Finished LM380N circuit board. A twin heatsink may be required for this chip
A heatsink should be fitted if the device is to be operated, other than intermittently, at output levels in excess of 1W. Without a heatsink, the suggested maximum supply voltages are: with a 4 ohm speaker 12V with an 8 ohm speaker 15V Frequency response up to 100kHz

Quiescent current Input resistance Input sensitivity for 720mW output (8 ohm load, 12V supply) Absolute maximum supply voltage beyond which damage will occur

13mA 150k ohms 50mV r.m.s. (gain 50) 22V

Everyday Practical Electronics, May 2002

1.6IN (40.6mm)

345

+
C2 220 C3 100n

+9V TO +18V

COMPONENTS
LM380N AMPLIFIER
Resistors
R1 2W7 025W 5% carbon film 10k rotary carbon, log. 4m7 radial elect. 50V 220m radial elect. 50V (2 off) 100n disc ceramic (2 off) 10m radial elect. 50V LM380N audio power amp i.c. 4 to 32 ohm loudspeaker (see text)

SIGNAL INPUT

14

C5 220

C1 47

+
2

IC1
LM380N

8
3,4,5, 10,11,12

See

+
7

R1 2 7
LS1 8 C6 100n

VR1 10k

Potentiometers
VR1

SHOP TALK
page

+
VOLUME

C4 10

SCREEN

0V

Capacitors
C1 C2, C5 C3, C6 C4 IC1 LS1

SUPPLY VOLTAGE RIPPLE REJECT INPUT +

1 2 3

14 13 12 11 10 9 8

SUPPLY VOLTAGE +V NOT CONNECTED

Semiconductor
OUTPUT, GROUND (0V) AND HEATSINK

OUTPUT, GROUND (0V) AND HEATSINK

4 5

Miscellaneous
NOT CONNECTED SIGNAL OUT

INPUT INPUT GROUND (0V)

6 7

TOP VIEW OF LM380N

Printed circuit board available from the EPE PCB Service, code 346 (LM380N); case (optional), size and type to choice; 14-pin d.i.l. socket; heatsink (see text); multistrand connecting wire; audio screened cable; solder pins; solder etc.

Fig.7. Circuit diagram for the LM380N Amplifier.

Approx. Cost Guidance Only

excluding case & speaker


is applied to the inverting input (pin 6) and, to avoid instability, the non-inverting input (pin 2) is grounded (0V). The manufacturers quote a minimum supply voltage of 10V. The sample tested worked with a 9V supply, but performance became erratic at lower voltages. Quiescent current, although modest, is double that of the other low-power devices, and this, together with the higher operating voltage, makes the i.c. more suitable for mains-powered equipment.

11

CIRCUIT BOARD

Details of the printed circuit board

component layout, wiring and copper foil master are given in Fig.8. This board is also available from the EPE PCB Service, code 346 (LM380N). Although the board has been kept small, as much copper as possible has been retained to afford some heatsinking.

TDA2003 AMPLIFIER
Produced by SGS-Thomson, the TDA2003 low-cost i.c. is mainly for use in car radios. Although chips designed specifically for hi-fi amplifiers are available, they usually require higher voltage and/or split rail power supplies. This makes them less easy and more expensive to use.
+
C3 220 C4 100n

+6V TO +15V

Component layout on the TDA2003 p.c.b.

5 1

SIGNAL INPUT

+
C2 47

IC1
TDA2003 2 3 C5 39n

C7 1000

TDA2003 POWER AMPLIFIER


R.M.S. power output just before the onset of waveform clipping
Speaker Impedance Ohms
2 4 8
0V

+
C1 47 R1 39

R2 220 C6 470

R4 1 LS1 4

Supply Voltage 9V
225W 128W 720mW

12V
4W 26W 144W

15V
575W 39W 21W

VR1 10k

C8 100n

VOLUME

R3 2 2

SCREEN

TYPE No. METAL TAG CONNECTED TO PIN 3 (GROUND) 5 4 3 2 1 FRONT VIEW OF TDA2003 SUPPLY VOLTAGE +V SIGNAL OUT GROUND (0V) INPUT INPUT +

Fig.9. Circuit diagram for a single TDA2003 Amp.

Quiescent current 45mA Input sensitivity for 2.6W output (4 ohm load, 12V supply) 42mV r.m.s. (gain 80) Absolute maximum supply voltage beyond which damage will occur 28V Absolute maximum operating voltage 18V Frequency response: 40Hz to 15kHz at the 3dB points. The upper frequency limit can be extended by reducing the value of C5.

346

Everyday Practical Electronics, May 2002

SCREENED INPUT LEAD

COMPONENTS
+
C4 C3

+
C2

POWER SUPPLY +VE

TDA2003 AUDIO AMPLIFIER


POWER SUPPLY VE (0V)

VOLUME

+
C1

Resistors
R1 R2 R3 R4 All 025W 5% VR1

See
39W 220W 2W2 1W page carbon film

IC1 + C5

C6

R 1

TO SPEAKER (LS1)

SHOP TALK

VR1 (FRONT VIEW)

Potentiometers
10k rotary carbon, log.
HEATSINK 2.5IN (63.5mm)

Capacitors
C1, C2 C3 C4 C5 C6 C7 C8 IC1 4m7 radial elect. 50V (2 off) 220m radial elect. 50V 100n disc ceramic 39n polyester 470m radial elect. 50V 1000m radial elect. 50V 100n polyester TDA2003 audio power amp i.c. 4 to 32 ohm loudspeaker (see text)

346
1.8IN (45.7mm)

Semiconductor

Miscellaneous
LS1

Fig.8. The LM380N printed circuit board component layout, off-board interwiring and full-size copper foil master pattern. Note the heatsinks.

HEATSINKS
For those readers who wish to get the most out of the chip, a suitable heatsink for the LM380 (Fig.7 and Fig.8) can be formed from two, 40mm (15/8in.) lengths

of 25mm 04mm (1in. 1/64in.) brass strip. Make two shallow cuts, 5mm (3/16in.) apart, close to the centre, and bend out a tag which can be soldered to the relevant pins of the i.c. Thin brass strip can be purchased from almost all model shops.

Printed circuit board available from the EPE PCB Service, code 347 (TDA2003); case (optional), size and type to choice; heatsink (see text); audio screened cable; multistrand connecting wire; solder pins; solder etc.

Approx. Cost Guidance Only excluding case & speaker

11

The TDA2003 incorporates short circuit and overload protection, and is extremely rugged. It will deliver a worthwhile output at modest supply voltages, and the suitability of car batteries as a power source may make it of particular interest to some readers. The

circuit diagram of a single chip TDA2003 audio amplifier is given in Fig.9. Grounding the input (pin 1) of this device would upset the internal biasing arrangements, so a second blocking capacitor C2 must be provided. The high

POWER SUPPLY +VE POWER SUPPLY VE SCREENED INPUT LEAD C5

VOLUME

C1

R 2 R1 1

2 3

4 5

IC1
C4

+
C7 R 4

+
C2

C6

+
R3

+
C3 C8

VR1 (FRONT VIEW) TO SPEAKER (LS1) 3.4IN (86.4mm)

frequency response is set by capacitor C5 in conjunction with resistor R1. The response can be extended by reducing the value of C5. Supply line ripple rejection is afforded by capacitor C6. The outputs which can be delivered at various supply voltages are tabulated in the accompanying table. The current drawn from a 15V supply when 4W are dissipated into a 4 ohm load is around 500mA. The 2 ohm load is obtained by connecting two 4 ohm speakers in parallel. At these power levels, the device must, of course, be connected to an adequate heatsink, and this is discussed later. The printed circuit board component layout, wiring and full-size copper foil master pattern for the single chip TDA2003 amplifier are shown in Fig.10. This board is available from the EPE PCB Service, code 347 (TDA2003).

TWIN TDA2003 AMPLIFIER


1.2IN (61.0mm)

347
Fig.10. Printed circuit board component layout, full-size foil master and off-board wiring for the single TDA2003 Amplifier.

A circuit diagram using two TDA2003 chips in a bridge configuration is shown in Fig.11, together with a general performance guide. Drawing around 17A from a 15V supply, this combination will deliver a clean 125W into a 4 ohm load. The case for this being adequate for domestic listening has already been argued, but individual constructors will, of course, decide whether or not it will meet their needs.

Everyday Practical Electronics, May 2002

347

TWIN TDA2003 AMPLIFIER

COMPONENTS
TWIN TDA2003 POWER AMP
Resistors
R1, R4 R2 R3, R5 R6 All 025W 5% VR1

See
1W (2 off) 220W 10W (2 off) 470W carbon film page

SHOP TALK
C3 100n

Twin TDA2003 Amplifier circuit board component layout.


+6V TO +15V

Potentiometers
10k rotary carbon

R4 1

C6 100n

+
C8 100n
C10 220

Capacitors
C1, C2 C9 C3, C8 C4 C5, C6 C7 C10 IC1, IC2 4m7 radial elect. 50V (3 off) 100n disc ceramic (2 off) 22m radial elect. 50V 100n polyester (2 off) 10m radial elect. 50V 220m radial elect. 50V TDA2003 audio power amp i.c. (2 off)
SCREEN
5 1 LS1 4 5 4 4

SIGNAL INPUT

+
C2 47

IC1
TDA2003 2 3 R1 1 C4 22

IC1
TDA2003 2 3

+
C1 47

R6 470

R2 220

R5 10

C7 10

Semiconductor

VR1 10k

+
VOLUME C5 100n R3 10 C9 47

Miscellaneous
LS1 2 to 8 ohm loudspeaker (see text)

0V

Printed circuit board available from the EPE PCB Service, code 348 (TDA2003); case (optional), size and type to choice; heatsink (see text); audio screened cable; multistrand connecting wire; solder pins;

Fig.11 (above). Circuit diagram for the Twin TDA2003 Power Amplifier.

TWO TDA2003 BRIDGE CONFIGURATION POWER AMP


R.M.S. power output just before the onset of waveform clipping
Speaker Impedance Ohms
2 4 8

Supply Voltage 9V
625W 378W 2W

Approx. Cost Guidance Only excluding case & speaker

14

12V
105W 8W 5W

15V
125W 82W

Fig.12 (below). Component layout, off-board interwiring and full-size copper foil master for the Twin TDA2003 Amp. You will need a heatsink for these devices.

Quiescent current 80mA Input sensitivity for 8W output (4 ohm load, 12V supply) 70mV r.m.s. (gain 40) See single TDA2003 for details of absolute maximum ratings.

TO SPEAKER (LS1) SCREENED INPUT LEAD

VOLUME

IC1 + C1
1

C4 2 3 4 5 C3

+
C6 R2 R 1 R4 R5 R 6 C7

IC2 +
C9 2 4 3 5 C8 POWER SUPPLY +VE 1

+
C2

R 3 C5 C10

POWER SUPPLY VE (0V)

VR1 (FRONT VIEW)

4.0IN (101.6mm)

348

348

Everyday Practical Electronics, May 2002

1.2IN (61.0mm)

The printed circuit board component layout, wiring and copper foil master are detailed in Fig.12. Again, combined or separate heatsinks must be fitted to the integrated circuits metal tabs. The p.c.b. is obtainable from the EPE PCB Service, code 348 (Twin TDA22003).

HEATSINKS

A large area metal heatsink is required for the TDA2003 (Fig.9 and Fig.11). Because the device incorporates overload protection, the actual size is not too critical (the i.c. will shut down when it begins to overheat), but sustained high output will only be developed if the heatsink is adequate. At the very least use 40sq. cm (65sq. in.) of 16s.w.g. aluminium per chip, or fit a proprietary heatsink with a thermal resistance not greater than 7C per watt. The i.c.s are arranged on the p.c.b. (see Fig.10 and Fig.12) so that they can be bolted to the back of a metal case by their metal tabs. A 50mm 150mm 200mm (2in. 6in. 8in.) aluminium box would be more than adequate as a heatsink. Insulating washers are not required, but a smear of heat transfer compound should be applied.

LM386 are usually referring to the N-1 version. The TDA7052 is sometimes given the suffix A. This indicates that the chip contains a d.c. volume control and is not suitable for the circuit described here. Some suppliers give the LM380 the suffix 14 to indicate the 25W 14-pin version, and the suffix 8 for the 8-pin 600mW alternative. When ordering, make it clear that the 14-pin chip is required. The suffix P or V is sometimes added by suppliers to the TDA2003 to indicate that it is for vertical, and H for horizontal, mounting. There is no electrical difference, but the p.c.b.s illustrated here have been designed for vertical chips.

CONSTRUCTION

COMPONENTS

Slight differences in the i.c. type numbers can cause confusion. The LM386N-1 has the lowest power rating of this group of devices. The suffixes N-3 and N-4 indicate devices rated at 700mW and 1W respectively. The suffix M indicates surface mounting. Suppliers offering the

All the amplifiers covered in this part are assembled on printed circuit boards and construction is reasonably Next Month: Transistor preamplifiers straightforward. The use of an i.c. holder will permit the substitution and checking of the low power amplifiers. However, if reliance is to be placed on the p.c.b. foil for minimal heatsinking of the LM380, the device should be soldered directly in place. Solder pins, inserted at the lead-out points, will simplify off-board wiring. It may help to start construction of the chosen circuit board by first placing and soldering the i.c. holder

on the p.c.b. to act as an orientation guide. This should be followed by the leadoff solder pins, and then the smallest components (resistors) working up to the largest, electrolytic capacitors and presets. Finally, the lead-off wires (including the screened input cable), off-board Volume control and loudspeaker should be attached to the p.c.b. On completion, check the board for poor soldered joints or bridged tracks. Check the orientation of the electrolytic capacitors and the i.c.(s). If using a mains power supply, make sure the voltage delivered does not exceed the safe working voltage of the amplifier for the load impedance being used. If all is in order, connect the power supply and check the quiescent current consumption. Inject a signal and re-check the current drain and supply voltage.

A Complete range of regulated inverters to power 220V and 240V AC equipment via a car, lorry or boat battery. Due to their high performance (>90%) the inverters generate very little heat. The high stability of the 150W TO 2500W - 12V & 24V output frequency (+/-1%) makes them equally suitable to power sensitive devices. These inverters generate a modified sine wave, which are considerably superior to the square waves which are produced by most other inverters. Due to this superior feature they are capable of powering electrical equipment such as TV,s, videos, desktop & notepad computers, microwave ovens, electrical lamps, pumps, battery chargers, etc. Low Battery Alarm The inverters give an audible warning signal when the battery voltage is lower than 10.5V (21V for the 24V version). The inverter automatically shuts off when the battery voltage drops below 10V (20V for the 24V version). Fuse protected input circuitry. Order Code Price Voltage Power

A COMPLETE RANGE OF

INVERTERS

651.581 651.578 651.582 651.585 651.583 651.593 651.587 651.597 651.602 651.605 651.589 651.599

REF D4

150W Continuous 150W Continuous 300W Continuous 300W Continuous 600W Continuous 600W Continuous 1000W Continuous 1000W Continuous 1500W Continuous 1500W Continuous 2500W Continuous 2500W Continuous

12V 24V 12V 24V 12V 24V 12V 24V 12V 24V 12V 24V

36.39 36.39 50.64 50.64 101.59 101.59 177.18 177.18 314.52 314.52 490.54 490.54

ILLUSTRATION SHOWN IS 651.583 600W VERSION DELIVERY CHARGES ARE 6-00 PER ORDER. OFFICIAL ORDERS FROM SCHOOLS, COLLEGES, GOVT. BODIES, PLC,S ETC. PRICES ARE INCLUSIVE OF V.A.T. SALES COUNTER. VISA AND ACCESS ACCEPTED BY POST, PHONE OR FAX, OR EMAIL US AT SALES@BKELEC.COM ALTERNATIVELY SEND CHEQUE OR POSTAL ORDERS MADE PAYABLE TO BK ELECTRONICS.

All prices are inclusive of V.A.T. Carriage 6.00 Per Order


Many uses include:- * Fetes * Fairgrounds * Airshows * Picnics * Camping * Caravans * Boats * Carnivals * Field Research and * Amateur Radio field days * Powering Desktop & Notepad Computers.

B.K. ELECTRONICS
Everyday Practical Electronics, May 2002

UNIT 1, COMET WAY, SOUTHEND-ON-SEA, ESSEX. SS2 6TR TEL.: +44(0)1702-527572 FAX.:+44(0)1702-420243 W W W . B K E L E C . C O M / I N V E R T E R S . H T M 349

For Full Specifications View our web site at:-

SIMPLE AUDIO CIRCUITS


Part 3 Power Supplies, Loudspeakers, Crossover Networks and Filters

Constructional Project

RAYMOND HAIGH
A selection of pic-n-mix low-cost audio circuits from preamplifier to speaker!

a modest output from one of the smaller power amplifiers (May 02) is all that is required, dry batteries represent a suitable power supply. However, when the output is expected to exceed the half-watt level for sustained periods, a mains power unit is more appropriate. Savings in the cost of batteries will quickly cover expenditure on components. Compromises, inherent in the design of loudspeakers, give rise to limitations which are normally overcome by the use of two or more units and a crossover. Power supplies, loudspeakers and associated networks are the topics to be covered this month.
F

A simple mains power supply comprising a full-wave rectifier and capacitor input filter will deliver an off-load voltage of around 14 times the transformer secondary voltage. With a secondary rated at 12V a.c., the off-load d.c. output voltage will, therefore,

SUPPLY REGULATION

be almost 17V. If the power supply output is close to the maximum safe operating voltage of the amplifier i.c., there is a danger that, under no-signal conditions, the device will be ruined. When fully loaded, the d.c. output voltage will fall to around 14V with an adequately rated transformer; lower when the transformer specification has been skimped. Voltage will, therefore, be low at the very moments when the power amplifier is being called upon to deliver a high output. These voltage variations are a cause of distortion and impair the performance of the power amplifier. Moreover, when highgain preamplifiers or radio tuners are fed from the same supply, the variations can also result in instability, even when substantial decoupling is provided.

POWER SUPPLY

These problems can be avoided by regulating the output of the power supply, and a versatile circuit, which can be adapted

for single or stereo pairs of any of the amplifiers described in Part 1 (May 02), is given in Fig.1. The mains voltage is stepped down by transformer T1, and a full-wave bridge rectifier arrangement, D1 to D4, produces the d.c. output. Reservoir capacitor C5 reduces supply ripple. Voltage regulators IC1 and IC2 virtually eliminate any voltage swings caused by load variations. The regulators also remove any residual 100Hz ripple on the supply voltage rails and permit the use of a lower value reservoir capacitor (C5). Low level electrical noise, extending into the r.f. spectrum, is present in the output of the i.c.s, and bypass capacitors, C6, C7, C8 and C9, shunt this to the 0V rail. The voltages required by amplifiers, preamplifiers and auxiliary equipment are often different, and provision is made for two regulated outputs. Alternatively, each output can supply a separate channel of a stereo system in order to double the current rating. The switching action of the rectifier diodes (D1 to D4) modulates any r.f. (radio frequencies) present in the mains input. This modulated r.f. can be picked up by radio receivers connected to the supply and it manifests itself as a 100Hz hum which only appears when a station is tuned in. Capacitors C1 to C4, connected across the diodes, suppress this interference, which is known as modulation hum. If radio tuners are to be powered from this circuit, these capacitors must be fitted.

COMPONENT RATINGS

Simple i.c. Power Amp. modules (left-to-right, topto-bottom) from May 02 issue. B Single TDA2003 Amp B Twin TDA2003 Amp B TDA7052 Amp B TBA820M Amp B LM380N Amp B LM386N-1 Amp

Fuse It is good practice to protect the equipment with an internal fuse of the lowest possible rating. Because of the nature of the load, this should be of the anti-surge or slow-blow type, and a component rated at one amp (1A) would be suitable for power supplies serving the amplifiers described in this series of articles. Transformer The rectified d.c. voltage across the reservoir capacitor (C5) must be at least 3V more than the regulator output when

500

Everyday Practical Electronics, July 2002

maximum current is being drawn from the supply. Further, the maximum input voltage to the regulator i.c., which is usually 35V for devices with a 2A rating, must not be exceeded. It is also desirable for the voltage drop across it to be no more than 10V or so, or power dissipation within the chip will be increased and more elaborate heatsinking will be required. These requirements can best be met if the mains transformer secondary voltage is 3V more than the regulated d.c. output. To determine the required current rating of the secondary winding, add together the demands of the amplifiers and ancillary equipment to be connected to the power supply, and increase this by at least 25 per cent to allow for the reactive load presented by the reservoir capacitor (C5). The current requirements of the power amplifiers were given in Part 1. For convenience, they are repeated here in Table 2. Manufacturers usually indicate the current delivering capacity of their mains transformers by quoting a VA rating. This is, of course, the secondary output voltage multiplied by the maximum current which the transformer can supply. In Europe, mains transformers often have two 115V primary windings and two identical secondary windings. The primary windings must be series or parallel connected to suit the local supply voltage, and the secondary connected to deliver the desired output. Parallel connecting the secondary will, of course, double the current available. Connect the windings in phase or the transformer will be short circuited.

MAINS TRANSFORMER PRIMARY TO SUIT SUPPLY VOLTAGE. SEE TABLE 1 FOR DETAILS OF SECONDARY T1
230V C1 100n

SEE TABLE 1 FOR DETAILS OF VOLTAGE REGULATORS IC1 AND IC2


C3 IN 100n

REGULATED OUTPUT 1

IC1
COM

OUT

+V
C6 100n

D1 1N4002 k
a

D3 a 1N4002

+
C7 470

1A TIME DELAY (SLOW BLOW) FUSE

FS1

PL1

k SEE + COMPONENT LIST a k D2 k D4 a 1N4002 1N4002

REGULATED OUTPUT 2

L
N

0V

IN
C4 100n

IC2
COM

OUT

C2 100n

+V +
C9 470
0V

E
EURO STYLE MAINS INLET PLUG

C5 2200 SEE NOTE

C8 100n

Fig.1. Circuit diagram for a Dual Output Regulated Power Supply.


Table 1: Component Ratings
Regulated Output V d.c.
6 9 12 15

Transformer Sec. V r.m.s.


9 12 15 18

Regulator I.C. (1A max output)


L7806 L7809 L7812 L7815

C5 Working Voltage
25 25 35 35

NOTES: (1) To determine the transformer current rating, add together the current demands of pre and power amplifiers and any ancillary equipment, then increase the total by at least 25% to allow for the reactive load presented by C5. (2) A bridge-connected pair of TDA2003 i.c.s with a 4 ohm load will draw 17A from a 15V supply and the ratings of the rectifiers, regulator and reservoir capacitor must be increased. Use 1N5401 rectifiers, an L78S15 regulator and a 4700mF capacitor for C5 (35V working). (3) For two, bridge-connected pairs of TDA2003 i.c.s in a stereo combination, fit a 10000mF (or two 4700mF) 35V reservoir capacitor, two L78S15 regulators, (one for each stereo channel) and use P600D rectifiers.

Rectifiers With a capacitor input filter, the rectifiers (D1 to D4) must have a p.i.v. (peak inverse voltage) rating at least three times the secondary voltage of the mains transformer. Their current rating should be at least 50 per cent greater than the maximum load on the power supply. Reservoir Capacitor The value of the reservoir capacitor, in microfarads (mF), should be at least 2500 times the maximum load current in amps when the supply is regulated, and double this value when unregulated. The working voltage should be at least double the secondary voltage of the mains transformer. Regulators The current rating of the voltage regulators (IC1 and IC2) must, of course, be equal to or greater than the maximum current demand on the power supply. The maximum input voltage rating (usually 30V to 35V) must be at least 15 times the secondary voltage of the mains transformer. Regulator i.c.s are available in a range of output voltages suitable for the audio amplifiers (May02) and preamplifiers (June02) described in this Table 2: Power Amplifier Current Requirements
Power Amp I.C.
LM386N-1 LM386N-1 TDA7052 TDA7052 TBA820M TBA820M LM380N LM380N TDA2003 TDA2003 TDA2003 x 2 TDA2003 x 2

Completed power supply board.


series. Maximum current ratings are 5A for 12V and 3A for 15V units, but chips rated at more than 2A can be difficult to obtain. When the current demand exceeds 2A; e.g. when two, bridge-connected, pairs of TDA2003 audio power amplifier modules are used in a stereo combination, fit a 2A regulator to each output of the power supply and use one for each stereo channel. Suppressor Capacitors The working voltage of capacitors C1 to C4, connected across the rectifier diodes, should be at least four times the secondary voltage of the mains transformer. Bypass capacitors C6, C7, C8 and C9, should have a working voltage at least 15 times the transformer secondary voltage to protect them in the event of regulator failure.

Speaker Imp Ohms


4 8 4 8 4 8 4 8 4 8 4 8

Supply volts V d.c.


6 9 6 9 9 12 12 15 15 15 15 15

Current drain A
013 012 042 039 023 017 023 019 05 027 17 096

Power output W
032 056 078 1 098 11 112 132 392 21 125 82

Current drain and power output measured just before the onset of clipping.

Everyday Practical Electronics, July 2002

501

DUAL OUTPUT REGULATED POWER SUPPLY

CONSTRUCTION

COMPONENTS
POWER SUPPLY Capacitors C1 to C4 C5 C6, C8 C7, C9 D1 to D4 100n ceramic, 100V (4 off) 2200m radial elect. (see Table 1) 100n ceramic, 50V (2 off) 470m radial elect. 50V (2 off)

See

Any readers who have no experience of building or commissioning mains-powered equipment are reminded that the voltages involved can kill! Anyone who feels unsure of his or her ability to complete a project of this kind MUST seek help and guidance from an experienced constructor. The small components are assembled on the printed circuit board (p.c.b.) as illustrated in Fig.2, together with a full-size copper foil master and the interwiring to off-board components. This board is available from the EPE PCB Service, code 356. Commence construction by first soldering in position on the p.c.b. the rectifier diodes and non-electrolytic capacitors. This can be

SHOP TALK
page

Semiconductors
1N4002 rect. diode for 1A max. output (4 off) 1N5401 rect. diode for 3A max. output (4 off) P60D rect. diode for 4A max. output, limited by regulators (4 off) 78 series for 1A; 78S series for 2A maximum output. See Table 1 (2 off) mains transformer see text and Table 1 1A 20mm slow-blow fuse to suit holder Euro fused mains inlet, chassis mounting, plug with line socket

IC1, IC2

Miscellaneous
T1 FS1 PL1

Printed circuit board available from the EPE PCB Service, code 356 (PSU); metal case, size and type to choice; multistrand connecting wire; mains cable; aluminium sheet or proprietary heatsink and heatsink compound; solder pins; nuts, bolts and washers; stand-off pillars (4 off); solder etc.

Approx. Cost Guidance Only

excluding case

20

FS1

a D3

C 3

OUT
k

C 6

+
C7

OUTPUT 1 REGULATED +V

230V

D1 a

C 1

COM IN

IC1
T1
0V
k a D2

C5

IC2
OUT C 2 D4 a C 4
k

C 8 C9

SEPARATE EARTH (0V RAIL) RETURNS TO PREAMP, POWER AMP, ETC.

+
REGULATED +V OUTPUT 2

COM IN

E N
SOLDER TAG HEATSINK 3.4IN (86.4mm)

PL1

356
TYPE NUMBER 1.85IN (47mm)

IC1 AND IC2 CONNECTION DETAILS

VIN
1 2 3

VOUT
COMMON

Fig.2. Power Supply printed circuit board, full-size copper master and suggested mains transformer and separate panel fuseholder interwiring. The 16 s.w.g. aluminium heatsink measures 45mm x 45mm.

502

Everyday Practical Electronics, July 2002

followed by the larger electrolytic types and the voltage regulators IC1 and IC2. Finally, you will need to bolt a heatsink to the regulators and details of choosing a suitable heatsink will be given shortly. Solder pins, inserted at the lead-out points, simplify the task of off-board wiring. Diodes D1 to D4, the reservoir capacitor, C5, and the regulators, IC1 and IC2, have to be chosen to suit the voltage and current to be delivered by the power supply. The requirements are summarised in Table 1 and the associated notes. Details of the modest current needs of the various preamplifiers were given in Part 2 of the series, and the current demands of the power amplifiers are scheduled in Table 2. Dimensions and fixing arrangements for mains transformers vary and this heavy component should be mounted directly into or on the metal equipment case bottom or chassis panel. A Euro-style mains inlet plug, with a built-in fuseholder for FS1, is strongly recommended. You can, of course, use a separate panel-mounting fuseholder if you wish, see Fig.2. Mains Earth should be connected to any metal case and to the core and cladding of the transformer. (A solder tag bolted under one of the mains transformer mounting lugs makes a good earthing point for the mains Earth lead.) Interwiring details to off-board components are also shown in Fig.2. Leads connecting the mains transformer to the inlet plug and the p.c.b., and any mains switch wiring, should be tightly twisted to minimise external fields. Keep the transformer at least 150mm (6in.) away from signal input wiring. Toroidal transformers have a smaller external field than units with conventional cores. They are the component of choice when the equipment is particularly compact and/or high gain preamplifiers are used.

LOUDSPEAKERS
Loudspeaker (speaker) designers have to make compromises. Sensitivity, good transient and good high frequency response call for a lightweight cone and speech coil assembly. Power handling and an extended low frequency response require a large, strong (and heavy) cone and coil. For good sensitivity, the magnetic field cutting the voice coil must be intense. Unfortunately, this increases the impedance at the cones resonant frequency. However, this impedance rise can be controlled by the speaker enclosure, and a powerful magnet is always preferable. The reproduction of low frequencies involves large cone excursions and the suspension must be highly compliant. High compliance also lowers the cones resonant frequency, and this extends the speakers low frequency response. However, the need to maintain control of the position of the voice coil in the magnet gap imposes limits on how free the suspension can be. Cone movement for a given sound output reduces with increasing speaker size but, as we have seen, greater diaphragm mass impairs transient and high-frequency response. to be preferred. Clarity will be impaired if low frequencies are allowed to excite the cone of a speaker of this kind, and measures to prevent this were discussed in Part 1 (May 02).

IMPEDANCE

HORSES FOR COURSES

HEATSINKING

Unless the current drain is to be very low (say 20mA or less), the regulator i.c.s must be bolted to a heatsink. The 45mm 45mm sheet of 16s.w.g. aluminium shown on the drawing (Fig.2) is sufficient for current drains up to 1A when the voltage drop across the regulators is not too extreme. For larger current loads it is suggested that the heatsink be extended and bolted to the metal case or chassis of the unit to ensure adequate heat transfer. Failure to properly dissipate heat from the regulators will result in the devices shutting down.

COMMISSIONING

Once construction has been completed, check the p.c.b. for poor soldered joints and bridged tracks. Check the orientation of electrolytic capacitors, diodes and regulators. Make sure that the primary windings of the mains transformer are connected to suit the local supply voltage, and that the secondary windings are connected, in phase, to deliver the correct voltage to the power supply p.c.b. It is a good idea to connect the transformer to the mains and check the secondary voltage with a test meter before linking it to the p.c.b. Extra care must be taken when carrying out this last task. Check the voltage across the reservoir capacitor C5, and that the voltages delivered by regulators (IC1 and IC2) are correct before using the supply to power any equipment.

To avoid performance being excessively degraded by these conflicting requirements, domestic hi-fi systems usually combine two or more speakers, each being designed to reproduce part of the audio frequency spectrum. The low frequency unit, or bass speaker, has a comparatively heavy cone and voice coil with a highly compliant suspension. Clever designers have managed to obtain reasonable results with small speakers, but an extended low frequency response and good power handling are more easy to achieve with speakers of 200mm (8in.) or more in diameter. Mid-range units are sometimes provided when the low frequency speaker is large (300mm to 450mm or 12in. to 18in. diameter). As one would expect, cones are lighter, the compliance is often stiffer, and the chassis can form a sealed enclosure. High-frequency units, or tweeters, have a very small diaphragm, which is commonly dome shaped to improve sound dispersal. Units of this kind always have sealed backs. Whilst moving coil tweeters are the preferred option for hi-fi applications, hornloaded piezoelectric units are often fitted in the high power speaker systems used by musicians. The impedance of these devices rises, and their power consumption falls almost to zero, as the applied frequency is lowered. They do not, therefore, require a crossover unit, and are easy to connect into multiple speaker systems.

Speech coil impedance is usually measured at around 400Hz. At this frequency, the inductance of the coil has a minimal effect, and its impedance is only one or two ohms more than its d.c. resistance. As frequency rises, the inductance of the speech coil has a growing impact and impedance mounts steadily. The movement of the speech coil in the magnetic field induces in it a voltage which opposes the signal voltage. At the cones resonant frequency, very little energy is needed to sustain it in motion, and it vibrates readily, over larger distances, for a comparatively small power input. These larger cone excursions generate a greater opposing voltage, or back-e.m.f., and speech coil impedance, at resonance, increases by as much as a factor of ten over its nominal value. The more powerful the magnetic field, the more dramatic the rise in impedance. Impedance peaking at cone resonance (between 30Hz and 100Hz for low frequency speakers), and the gradual rise in impedance with increasing frequency, makes the response of the speaker non-linear. (The power which can be fed to a speaker system falls as its impedance rises). Fortunately, the former can be tamed by good enclosure design, and the latter can be overcome by the use of filter networks and the addition of a tweeter. Care must always be taken to ensure that the rated impedance of a speaker system is not too low for the power amplifier. Too low an impedance will cause excessive dissipation in the output transistors and, if there is no overload protection circuitry, the power amplifier will be ruined.

COMMUNICATIONS

Loudspeakers intended primarily for speech reproduction in communications equipment have to perform well over a restricted frequency range, usually around 300Hz to 3000Hz. Inexpensive speakers of the type manufactured for portable receivers are better suited for this purpose, and, if space is available, a 102mm (4in.) diameter unit is

Everyday Practical Electronics, July 2002

503

Safe supply voltage and speaker impedance combinations for the various i.c. power amplifiers were given in Part 1. They are summarised here in Table 2.

Table 3: Crossover Network Inductor and Capacitor Values


Crossover frequency Hertz
4 ohm Speaker 1st Order Filter 4 ohm Speaker 2nd Order Filter 8 ohm Speaker 1st Order Filter 8 ohm Speaker 2nd Order Filter L C L C L C L C

500
13 80 18 56 26 40 36 28

1000
063 40 09 28 126 20 18 14

1500
042 26 06 18 084 13 12 9

2000
032 20 05 14 064 10 1 7

2500
025 16 035 11 05 8 07 6

3000
021 13 03 9 042 65 06 45

3500
018 11 025 8 036 55 05 4

4000 4500
016 10 022 7 032 5 044 35 014 8 02 6 028 4 04 3

CROSSOVERS

When two or more speakers are used to improve performance, arrangements must be made to allocate the audio spectrum between them. The resistance presented by capacitors to the flow of alternating current decreases as frequency rises. With inductors, resistance increases with rising frequency. This frequency-dependant opposition to current flow is known as reactance. Capacitors and inductors can be combined in simple networks which utilise this phenomenon to allocate frequency bands to different speakers. Circuits and design data are given in Fig.3 and inductor and capacitor values for common speaker impedances, and a range of crossover frequencies, are set out in Table 3. The reactances of standard value capacitors, at

Inductance values, L, are given in mH (millihenries). Capacitor values, C, are given in mF (microfarads). See text for guidance on rounding figures up or down to nearest standard value.

various audio frequencies, were tabulated in Part Two.

FILTER ORDERS

The simple first order filters shown in Fig.3a and Fig.3d are perfectly suitable for domestic systems rated at up to 15W.

Low frequency roll-off above the crossover frequency is 6dB per octave and this may not be sufficient to protect some tweeters when higher powered amplifiers are used. In these cases, the second order filters, shown in Fig.3b and Fig.3e, which produce a 12dB roll-off, are safer options.

L
L1

INPUT

BASS

INPUT

MID RANGE

INPUT

C1

C1

C)
BASS
TREBLE L1

A)

FIRST ORDER NETWORK MID-PASS ARRANGEMENT


L2

FIRST ORDER NETWORK TWO SPEAKER SYSTEM Make reactance of inductor, L, and capacitor, C, at the crossover frequency, equal to the rated speaker impedance.

Make the reactance of inductor, L, and capacitor, C, at the centre frequency of the pass band, equal to the rated speaker impedance. Assume a band centre frequency of 1000Hz when the circuit is being used as a band-pass filter for speech frequencies.

C2

C2

MID RANGE

L2

TREBLE

E)

L1

INPUT

C1

BASS

SECOND ORDER NETWORK THREE SPEAKER SYSTEM (1) Calculate inductor and capacitor values as for the first order three speaker network.

INPUT

BASS

L2

C2

MID RANGE

TREBLE

B) D)

TREBLE

(2) Divide the capacitor values by 14 and multiply the inductor values by 14 to obtain the correct values for the second order three speaker network.

SECOND ORDER NETWORK TWO SPEAKER SYSTEM (1) Calculate inductor and capacitor values, as for the first order network. (2) Divide the capacitor values by 14 and multiply the inductor values by 14 to obtain the correct values for the second order two speaker network. Inductors, L, are identical. Capacitors, C, are identical.

FIRST ORDER NETWORK THREE SPEAKER SYSTEM (1) Make inductor, L1, and capacitor, C1, reactances, at the bass/midrange cross-over frequency, equal to the rated speaker impedance. (2) Make inductor, L2, and capacitor, C2, reactances, at the mid-range/treble cross-over frequency, equal to the rated speaker impedance.

Fig.3. Circuit and design data for loudspeaker crossover networks. Inductor and capacitor values for common speaker impedances and crossover frequencies are given in Table 3.
Components with the same reference numbers have identical values i.e. L1 and L2 are two inductors of the same value; C1 and C2 are capacitors of the same value.

504

Everyday Practical Electronics, July 2002

CROSSOVER FREQUENCY

With two-speaker systems the crossover frequency is usually between 1kHz and 45kHz, and the tweeter manufacturers recommendations should be followed. If the unit is of uncertain origin, adopt a crossover frequency of around 25kHz: this will normally be satisfactory. When the bass speaker is large (12 inches diameter or more), a crossover at 1kHz or even lower can produce a more even frequency response. Suitable tweeters tend to be rather costly, but an inexpensive alternative will be described later.

CUT BOBBIN ENDS FROM 3mm HARDBOARD (MASONITE IN USA) USING A HOLE SAW IN AN ELECTRIC DRILL

CORE 60mm LENGTH OF 9.5mm DIA FERRITE ROD 50mm

13mm

25mm

GLUE BOBBIN END TO PAPER TUBE

ROLLED AND GLUED PAPER TUBE 13mm O.D.

THREE SPEAKERS

Fig.4. Inductor bobbin construction details.


Table 4: Inductance of Ferrite-cored Coils
Induct 01 mH No. of 45 turns 02 60 03 75 04 90 05 100 075 125 1 150 15 175 2 200 25 225 3 250 35 275

Another way of ensuring a more even response when a large bass speaker is used is to install a third, mid-range unit. Suitable circuits are given in Fig.3d and Fig.3e. The bass/mid-range crossover point is usually around 500Hz with open chassis mid-range speakers, and 1000Hz with sealed back units. The mid-range/treble crossover is generally between 45kHz and 6kHz. Again, the recommendations of the speaker manufacturer should be followed.

Use 20 s.w.g. (19 a.w.g.) enamelled copper wire for coils up to 2mH. Use 22 s.w.g. (21 a.w.g.) enamelled copper wire for 25mH to 35mH coils. See illustration for details of bobbin and core.

PHASING

Parallel connected bass speakers must be wired in phase to avoid cancellation of the lower audio frequencies. Use a 15V dry cell to test for phasing on unmarked speakers by noting the battery positive connection for the outward movement of the cone. Crossover networks introduce phase shift, but, as frequency increases, phasing becomes less important. Readers can try reversing the connections to mid-range units. However, unless they have a very refined ear, they are not likely to detect any difference.

will be suitable for all of the power amplifiers described in Part 1. The performance of electrolytic capacitors can become uncertain at high audio frequencies, and the best crossover networks use components with a paper, polyester or polypropylene dielectric. Tolerances Variations in the composition of ferrite rod will affect the tabulated inductor values shown in Table 4 by plus or minus 10 per cent or so. Bipolar electrolytics, whether purchased or homemade, have a tolerance, at best, of plus or minus 20 per cent. Fortunately, loudspeaker crossover networks are very forgiving, and component spreads even greater than this produce no audible difference. When calculated values are being rounded up or down, it is prudent to err on the high side with inductors and on the low side with capacitors.

FORMULAE FOR THE REACTANCE OF INDUCTORS AND CAPACITORS XL = 000628 f L ohms XC = 159000 ohms fc

where f is in Hertz L is in millihenries (mH) m and C is in microfarads (mF)

C1

C2

CROSSOVER COMPONENTS

Inductors Inductors for home-made crossovers have to be hand wound. The amount of wire, and the resistive losses, can be greatly reduced by winding the coils on short lengths of ferrite aerial rod. Core saturation problems should not arise at the power levels encountered in domestic installations. Bobbin construction is illustrated in Fig.4. Winding details for the inductor values likely to be encountered are given in Table 4. The wire should be wound on evenly, and masking tape, applied over each layer, will make the task a little easier. Constructors who have difficulty producing neat windings should increase the diameter of the bobbin ends for the larger inductance coils. Capacitors The bipolar electrolytic capacitors used in crossover networks are available in a limited range of values. Capacitors of this kind can be formed by connecting two ordinary electrolytics back-to-back, and this makes possible the production of nonstandard values. The details are given in Fig.5. Capacitors rated at 50V working

C1

C2

BANDPASS FILTERS

Mention has already been made of the desirability of restricting the audio bandwidth of speakers used primarily for speech communication. An inductor and capacitor can be combined to produce a bandpass effect, and a typical circuit is given in Fig.3c. As a starting point, select the inductor and capacitor values for a centre frequency of 1000Hz (1kHz). If a more severe attenuation of frequencies below 300Hz and above 3000Hz (3kHz) is required, reduce the capacitor and increase the inductor value. When using this network with earphones, connect both earpieces in parallel to produce an impedance of 16 ohms, and perform the calculations on this basis. Although extremely simple, this measure will greatly improve the clarity of speech, especially when signals are overlaid by received or generated noise within the amplifiers.

Fig.5. Creating a bipolar electrolytic from two capacitors.

FORMULAE FOR COMBINING CAPACITORS Two capacitors in series: Cx = C1 C2 C1 + C2

Capacitors in parallel: Cx = C1 + C2 + C3 . . . . . The working voltage of each capacitor should be at least 15 times the peak-to-peak signal voltage developed across the loudspeaker at maximum input.

CROSSOVER UNIT

The circuit diagram for an inexpensive 8ohm Crossover/Filter unit suitable for a multi-purpose workshop speaker is shown

Everyday Practical Electronics, July 2002

505

CROSSOVER/AUDIO FILTER
4

S1a
P

2
C1 10m

TAP 100 TURNS START

COMPONENTS
BASS SPEAKER

1
FINISH 200 TURNS 8

BIPOLAR ELECTROLYTIC S1b

L1

P
INPUT FROM AMPLIFIER 5

LS1 8

CROSSOVER/AUDIO FILTER Capacitors C1 10m bipolar radial elect. 50V (Alternatively, two 22m standard elect. connected back-to-back see text and Fig.5)

S1 POSITIONS 1) WIDE RANGE 2) LOW PASS 3) HIGH PASS 4) MID RANGE

Inductor
LS2 8 TREBLE SPEAKER

See

L1

Fig.6. Circuit diagram for the Loudspeaker Crossover/Audio Filter.

INPUT TERMINALS

TAP S1

95mm (3/8in.) dia. ferrite rod, length page 63mm (2in.); card, hardboard and glue for bobbin. Enamelled copper wire: Crossover only 100 turns 20 s.w.g. (19 a.w.g.). Crossover and Filter 200 turns 22 s.w.g. (21 a.w.g.). See Fig.4 and text

SHOP TALK

SK1

FINISH 3 C1 4 5

2 A

1 12 11 C 10 9 7 8

Miscellaneous
S1 3-pole 4-way rotary switch (only two poles used) SK1, SK2 4mm screw terminal post/socket (2 off) Printed circuit board available from the EPE PCB Service , code 357 (Crossover/Filter); multistrand and connecting wire; plastic control knob; speaker terminals; solder pins; solder etc.

SK2

L1

TREBLE

BASS

Approx. Cost Guidance Only

START

excluding speakers

SPEAKERS

1.95IN (49.5mm)

295IN (749mm)

357

Fig.7. Crossover/Audio Filter printed circuit board component layout, interwiring to off-board components and full-size copper foil master. The completed crossover is shown in the above photograph.

506

Everyday Practical Electronics, July 2002

in Fig.6. The first order filter serves as a basic crossover when the speaker is being used for testing or listening to hi-fi equipment. Switching out the Treble speaker and connecting the inductor in series with the Bass speaker gives a low-pass (top cut) effect. Connecting the capacitor in series with the speaker provides a high-pass (bass cut) arrangement. With the inductor and capacitor in series with the speaker, response to speech frequencies is emphasised, making the unit suitable for use with a communications receiver or for surveillance work. Rotary switch S1 selects the required function, and the inductor is tapped to provide appropriate values for the crossover and speech filter.

if capacitor C1 is a single, bipolar electrolytic. Constructors interested only in hi-fi applications can ignore the switching arrangements and simply connect a 100turn (05mH) inductor and the capacitor as shown in Fig.3a. Next Month: The final part will deal with speaker enclosures and include a low-cost, high-performance design

which incorporates this months Crossover/Filter unit. The construction of a simple and inexpensive oscillator and resonance detector, which can be used to match any speaker to an enclosure and optimise performance, will also be described.

CONSTRUCTION

Construction of the Crossover/Filter Unit is based on a small single-sided printed circuit board (p.c.b.). This board is available from the EPE PCB Service, code 357. The topside component layout, full-size copper foil master and off-board wiring details are illustrated in Fig.7. Again, solder pins at the lead-out points will simplify off-board wiring. The p.c.b. makes provision for series and parallel combinations of capacitors, and a wire link must be inserted

Low Frequency Oscillator for loudspeaker resonance checking.


them direct from RS (credit card only) on 2 01536 444079 or on the web at rswww.com. A post and handling charge will be levied. The above company supplied the Texas TLC7524CN 8-bit digital-to-analogue converter chip, code 650-087. It is also currently listed by Rapid (2 01206 751166 or www.rapidelectronics.co.uk), code 82-0764, but double check it is the 16-pin device being supplied. For those readers unable to program their own PICs, a ready-programmed PIC16F877-20 microcontroller can be purchased from Magenta Electronics (2 01283 565435 or www.magenta2000.co.uk) for the inclusive price of 10 each (overseas add 1 p&p). It is the 20MHz version you require. The software is available on a 35in. PC-compatible disk (EPE Disk 5) from the EPE Editorial Office for the sum of 3 each (UK), to cover admin costs (for overseas charges see page 539). It is also available Free from the EPE web site: ftp://ftp.epemag.wimborne.co.uk/pub/PIC/StyloPIC. The printed circuit board/keyboard is available from the EPE PCB Service, code 359 (see page 539). Simple Audio Circuits 3 Most of our components advertisers should be able to supply all the parts needed to construct the circuits in this months instalment of the Simple Audio Circuits. A suitable Bulgin fused Euro-style mains inlet, chassis mounting, plug (code MK18U or FT37S) together with an insulation, rear tag, protective cover (code JK67X) and line socket (UL16S) is listed by Maplin (2 0870 264 6000 or www.maplin.co.uk). They also list the 6A 200V P600D rectifier diode for one version of the Power Supply Unit, code UK60Q. If problems are experienced in obtaining a ferrite rod for the Crossover unit, we understand, from the author, that one is obtainable from JAB, PO Box 5774, Birmingham, B44 8PJ (mail order only), and J. Birkett (2 01522 520767). You will need to cut the rod down to size (take care, it is brittle!). These two firms can also supply 50g (2oz) reels of enamelled copper wire for the Crossover. The two printed circuit boards are available from the EPE PCB Service, codes 356 (PSU) and 357 (Crossover) see page 539. Rotary Combination Lock Probably the most expensive item when purchasing components for the Rotary Combination Lock project is likely to be the heavy-duty power solenoid. The one in the model cost about 15 and came from RS (2 01536 444079 or rswww.com) and is their 12V d.c. standard pull action, spring return type, code 250-1303. They also supplied the Omron 12V d.c. ultramin., p.c.b. mounting relay, code 369-359. The two printed circuit boards are available from the EPE PCB Service, code 260 (Lock) and 361 (Interface).

Infra-Red Autoswitch As the Infra-Red Autoswitch project is mains powered, all the components have been specially selected to fit directly on the small printed circuit board (p.c.b.). If alternative, non-board mounting components, such as the mains transformer and relay, are used you must take extra care when building and testing this unit. In this case, it is very important that the p.c.b. and any offboard parts be mounted in its case before testing and that a separate battery supply is used for checking its operation, prior to mains connection. The special Sharp IS471F infra-red sensor/detector came from RS Components and carries the order code 564-396. They also supplied the p.c.b. mounting, short-circuit proof, mains transformer with twin 9V 0027A (05VA total) secondaries, code 310-1263. These components can be ordered from any bona-fide RS stockists, including some of our advertisers. You can order direct (credit card only) on 2 01536 444079 or on the web at rswww.com. A post and handling charge will be made. The 12V d.c. low-profile relay, with 12A 250V a.c. rated single-pole changeover contacts, used in the model was purchased from Rapid Electronics (2 01206 751166 or www.rapid electronics.co.uk), code 604630. We understand that RS (see above) also stock a similar relay, code 198-6933. The specified low-profile case came from CPC (credit card only), 2 08701 202530, code EN55028. A post and packing charge is made on all orders under 30. The Autoswitch printed circuit board is available from the EPE PCB Service, code 358 (see page 539). Teach-In 2002 Lab 9 Once again, its only the sensor and semiconductor devices called for in this months Teach-In 2002 Lab Work that will give some readers sourcing grief. Starting with the Nemoto NAP-7AU gas sensor/compensator pair, these were obtained from Maplin (2 0870 264 6000 or www.maplin.co.uk), code FM87U and are sold as a pair. We have found two listings for the precision low off-set op.amp type OP177 and it can be ordered from Rapid Electronics (2 01206 751166 or www.rapidelectronics.co.uk), code 82-0092, or RS Components (2 01536 444079 or on the web at rswww.com), code 127-2868. Expect to pay a handling and postage charge. If readers experience any difficulty in finding a local source for the 4093 quad 2-input NAND Schmitt trigger (Rapid 83-0420) and the ADC0804 8-bit analogue-to-digital chip (Maplin QQ00A or RS 411-674) they should contact the above mentioned companies. The relevant code numbers are shown in brackets. The Linear Technology LTC1062CN8 5th order switched capacitor lowpass filter i.c., used in the Anti-aliasing Filter (Lab 9.5), appears to be listed only by RS (see above), code 633-880. EPE Stylopic A couple of items proved hard to find when tracking down parts for the EPE StyloPIC project. The National Semiconductor LM13600 transconductance amplifier i.c. and the SGS-Thompson L272 dual power op.amp i.c. only appear to be listed by RS, codes 304-453 and 635-167 respectively. You can order

PLEASE TAKE NOTE


L.E.D. Sequencer (Ingenuity Unlimited) June 02 Page 406. To prevent the i.c. outputs (IC2, IC3) from adversely affecting each other, 1N4148 signal diodes should be inserted between each i.c. pin and the respective l.e.d. The anode on the pin and cathode on the l.e.d. June 02 World Lamp Where it is said that VR1 should be turned clockwise, this should read anti-clockwise, and where anti-clockwise, clockwise. Toolkit TK3 Updated files for V1.2 are now on our FTP site. Only files Disk 1 and Disk 3 are affected.

Everyday Practical Electronics, July 2002

507

SIMPLE AUDIO CIRCUITS


Part 2 Preamplifiers, Tone Controls and Filters

Constructional Project

RAYMOND HAIGH
A selection of pic-n-mix low-cost audio circuits from preamplifier to speaker!

the power amplifiers described last month have a respectable amount of gain, some signals may be too weak to produce an adequate loudspeaker output without additional amplification. They can also be further weakened by an excessive mismatch between signal source and amplifier. Tone controls are usually required when music is being reproduced, and restricting the bandwidth will clarify speech signals, especially under noisy conditions. These three issues: preamplification, impedance matching and tailoring the frequency response, are covered in this article.
LTHOUGH

sufficient to fully drive the power amplifiers described last month. Experimenters who require the stages to have the highest possible signal-handling capability for a given supply voltage may have to adjust the bias resistors. Guidance on this is given later. Cascading The various preamplifiers, tone controls and filters can be combined to suit individual requirements. Blocking capacitors have been provided at the inputs and outputs, and the units can be used safely with any equipment. Cascading makes one of these capacitors redundant. Similarly, when they are connected to the power amp described last month, the output blocking capacitor can be omitted (C1 on the power amplifier p.c.b. duplicates this component).

TRANSISTOR AMPLIFIERS

Decoupling All of the preamplifier circuits are decoupled from the power supply by a resistor and capacitor. Failure to include these components will almost certainly result in motor boating (low frequency instability). The main cause of this instability is the wide swing in power amplifier current drain: even with small units this can range from 10mA to 150mA. These signalinduced current swings cause variations in the voltage of dry batteries or badly regulated mains power supplies. When high gain preamplifiers share the same supply rail, the resulting feedback causes low-frequency oscillation. If problems are encountered, increase the value of the decoupling resistor, or capacitor, or both, by a factor of ten. A capacitor of 2000mF or more, connected across a dry battery power supply, will also help to eliminate instability at high volume levels. R.F. Interference The single transistor preamplifiers described here have an extended high

Impedances The impedances presented by the input and output ports of transistor amplifier stages are extremely variable. Load and bias resistors exert a major influence, as do the gain of the transistor and its emitter current. Negative feedback can either raise or lower impedance and, to further confuse the issue, the load connected across one port influences the impedance presented by the other. The impedance figures quoted are, therefore, intended as no more than a guide when selecting the best circuit for a particular application. Biasing Transistor amplifier stages are usually biased so that the output (collector or emitter; drain or source) rests at half the supply voltage under no-signal conditions. This enables the stage to deliver the greatest possible signal swing; i.e. the highest output, before the onset of clipping. Transistor gain (hfe), and supply voltage, affect the biasing. However, over a wide range of hfe values (at least 200 to 600), and supply voltages from +9V to +12V, the circuits described here will deliver a low distortion output that is more than

Four single-transistor preamplifiers (left-to-right). B Low Impedance B Medium Impedance B High Impedance B F.E.T. High Impedance.

418

Everyday Practical Electronics, June 2002

frequency response, and problems with r.f. interference may be encountered. Connecting a low value ceramic capacitor between the input (emitter or base) and the 0V rail will cure the problem, and the accompanying printed circuit board (p.c.b.) makes provision for this.

SINGLE TRANSISTOR CIRCUITS

In many cases, all that is required is the additional gain and/or impedance matching afforded by a single transistor stage. Four circuits will now be considered.

Low Input Impedance Preamplifier

It is convenient, with simple intercom units, to make the speaker double up as a microphone. Voice coil impedance and output are very low: a few ohms and less than 1mV at a close speaking distance. Transformers are often used to increase the impedance and voltage of this signal

source, but a transistor can be made to do the job just as well. The grounded base stage illustrated in Fig.1 has an input impedance of around 50 ohms, an output impedance roughly equal to the collector load resistance (R2) of 10 kilohms, and a voltage gain of around 100. Although more commonly encountered at the front-end of a radio receiver, this configuration is suitable for matching low source impedances to the power amplifier and, at the same time, providing a useful amount of voltage gain. In the circuit diagram for the Low Input Impedance Preamplifier shown in Fig.1, C1 is a d.c. blocking capacitor, R1 and R2 are the input and output load resistors, and resistors R3 and R4 bias the transistor. The base (b) is grounded at audio frequencies by capacitor C3. Supply line decoupling is effected by C4 and R5, and C2 is the output coupling and d.c. blocking capacitor.

CIRCUIT BOARD

The printed circuit board component layout, wiring details and full-size copper foil master pattern are shown in Fig.2. This board is available from the EPE PCB Service, code 349 (Single Trans.). Before commencing assembly, check the component, construction and interconnection notes at the end of the article.

VARIATIONS

Readers wishing to operate the stage from lower supply voltages should check the voltage on the collector (c) of transistor TR1 under no-signal conditions. If it is much more than half the supply voltage, reduce the value of resistor R3 to increase the bias current. With 3V on the supply rail, R3 will need reducing to around 68 kilohms and, with a 6V supply, its value will be in the region of 12k. Because of its very low input impedance, the circuit of Fig.1 is not prone to capacitative hum pick up, and the input lead can be

R5 100

R2 10k

R3 18k

+9V TO +12V

LOW INPUT IMPEDANCE PREAMPLIFIER

BC549C c C1 100

TR1

C2 47

e C3 47

SIGNAL OUTPUT

C4 100 R4 2k2

SIGNAL INPUT

R1 1k

SCREEN

SCREEN

0V

VOLTAGE GAIN 100 OVER AN hfe SPREAD OF 110 TO 600. CURRENT DRAIN AT 9V SUPPLY 075mA.

Low Input Impedance Preamplifier components mounted on the single p.c.b.

Fig.1. Circuit diagram for the single-transistor Low Input Impedance Preamplifier.
+9V TO +12V
R 3 c b C1 INPUT C4 R2 R5 SCREENED LEAD TO POWER AMPLIFIER

COMPONENTS
OUTPUT

SCREENED SIGNAL INPUT LEAD, OR USE TIGHTLY TWISTED PAIR. (SEE TEXT)

+
C2

+ +
C3 R 4

TR1
R 1

LOW INPUT IMPEDANCE Resistors R1 1k See R2 10k R3 18k (see text) R4 2k2 page R5 100W All 025W 5% carbon film

SHOP TALK

TO COMMON 0V POINT ON POWER SUPPLY P.C.B.

Capacitors
C1, C4 C2 C3 TR1 100m radial elect. 25V (2 off) 4m7 radial elect. 25V 47m radial elect. 25V BC549C npn transistor (or similar see text)

1.7IN (43.2mm)

Fig.2. Printed circuit board component layout, wiring and full-size copper foil master pattern for the Low Input Impedance Preamplifier.

2.35IN (59.7mm)

Semiconductors

Miscellaneous
Printed circuit board available from the EPE PCB Service, code 349 (Single Trans); audio screened cable; multistrand connecting wire; input and output sockets, type to choice; solder pins; solder etc.

349
Everyday Practical Electronics, June 2002

Approx. Cost Guidance Only

7
419

tightly twisted flex rather than screened cable. If r.f. interference problems are encountered, connect a 100nF capacitor between the emitter (e) of TR1 and the 0V rail: provision is made for this on the p.c.b. Combining this low impedance circuit (Fig.1) with the LM386N-1 or the TBA820M power amplifiers (fully described in Part 1, last month) will produce a decent intercom unit, but more amplification is needed for surveillance purposes. Cascading the grounded base stage with the medium impedance preamplifier described next (Fig.3) is one possible answer.

Medium Input Impedance Preamplifier

The input impedance of the single transistor, common emitter preamplifier

illustrated in Fig.3 is approximately 1500 ohms (15k), and the output impedance roughly equal to the value of the load resistor, R2; i.e. 4700 ohms (47k). Base bias resistor R1 is connected to transistor TR1 collector (c) rather than the supply rail. The resulting d.c. negative feedback makes the biasing more immune to transistor gain spreads and variations in supply voltage. Preset potentiometer VR1 acts as the emitter bias resistor. Connecting capacitor C2 to the slider (moving contact) enables part of it to be left un-bypassed. This introduces varying levels of negative feedback and, with the specified transistor, the gain of the stage can be set between 10 and 160 times to suit different applications.

Comment has already been made about supply rail decouplers, R3 and C4, and blocking capacitors, C1 and C3.

CIRCUIT BOARD

The printed circuit board component layout, wiring details and full-size copper foil master pattern are shown in Fig.4. This board is available from the EPE PCB Service, code 349 (Single Trans.). Before undertaking assembly work, see the component, construction and interconnection details at the end of the article. Provision is made for connecting an r.f. bypass capacitor across the input. A 1nF or 10nF ceramic component should be adequate if problems arise.

R3 100

R1 1M

R2 4k7 C3 47 C4 100

+9V TO +12V +

MEDIUM INPUT IMPEDANCE PREAMPLIFIER

C1 47

TR1
BC549C b e c

+
SIGNAL OUTPUT

SIGNAL INPUT

C2 47
SCREEN

VR1 470

SCREEN

0V

VOLTAGE GAIN WITH VR1 SLIDER AT 0V RAIL, 8 TO 10 OVER AN hfe SPREAD OF 110 TO 600. VOLTAGE GAIN WITH SLIDER AT TR1 EMITTER, 80 TO 600 OVER AN hfe SPREAD OF 110 TO 600. CURRENT DRAIN AT 9V SUPPLY: 125mA.

Fig.3. Circuit diagram for the Medium Input Impedance Preamplifier.

Medium Input Impedance preamplifier components mounted on the single p.c.b.

COMPONENTS
MEDIUM INPUT IMPEDANCE Resistors See R1 1M R2 4k7 R3 100W All 025W 5% carbon film page

+9V TO +12V
SCREENED SIGNAL INPUT LEAD R2 c R 1 b e C4 R3 SCREENED LEAD TO POWER AMPLIFIER

+
C3

INPUT

SHOP TALK

C1

TR1 +
VR1 C2

OUTPUT

Potentiometers
VR1 470W enclosed carbon preset 4m7 radial elect. 25V (2 off) 47m radial elect. 25V 100m radial elect. 25V BC549C npn transistor (or similar see text)
2.35IN (59.7mm)

TO COMMON 0V POINT ON POWER SUPPLY P.C.B.

Capacitors
C1, C3 C2 C4 TR1

Semiconductors
1.7IN (43.2mm)

Miscellaneous
Printed circuit board available from the EPE PCB Service, code 349 (Single Trans); audio screened cable; multistrand connecting wire; input and output sockets, type and size to choice; solder pins; solder etc.

349
Fig.4. Medium Input Impedance Preamplifier printed circuit board component layout, wiring and full-size copper foil master.

Approx. Cost Guidance Only

420

Everyday Practical Electronics, June 2002

High Input Impedance Preamplifier

Crystal microphones and ceramic gramophone pick-ups (there are still a few in use) require an amplifier with a high input impedance, and a stage of this kind is useful when the damping on a signal source has to be kept low. Configuring a bipolar transistor in the emitter-follower (common collector) mode results in a high input and low output impedance, and a typical High Input Impedance Preamplifier circuit diagram is shown in Fig.5. The input impedance is roughly equal to the gain of the transistor (hfe) multiplied by the value of the emitter load resistor R2. This is, however, limited by the bias resistor R1, and the output load, which shunts the emitter resistor. Nevertheless, a high gain transistor will still produce an input impedance of about 100 kilohms. Often the low output impedance is the sought after feature, either for matching purposes or for avoiding high-frequency losses and hum pick-up when long screened cables have to be used. Output impedance is

directly related to the impedance presented by the signal source, and is usually in the region of 1000 ohms. The voltage gain of the circuit is a little less than unity.

input impedances are required, is to use a field effect transistor (f.e.t.); a device which tends to introduce less noise at audio frequencies.

CIRCUIT BOARD

The printed circuit board component layout, wiring details and full-size copper foil master pattern for the High Input Impedance Preamplifier are shown in Fig.6. This board is the same one used for all the single transistor preamplifiers, and is available from the EPE PCB Service, code 349 (Single Trans.). See the component, construction and interconnection notes at the end of the article. High input impedance makes the stage very vulnerable to hum pick up. Careful attention must, therefore, be paid to screening the input leads and, possibly, the entire unit.

USING A F.E.T.

VARIATIONS

It is possible to obtain higher input impedances with a bipolar transistor by applying positive feedback from the emitter to the base bias network. This involves an extra pair of resistors and a capacitor, and an alternative solution, if very high

A circuit diagram for a F.E.T. High Input Impedance Preamplifier is given in Fig.7. The gate resistor R1 is tapped down to the source resistors R2/R3 in order to improve biasing and, hence, signal handling. By this means the f.e.t. develops its gate bias across R2, and R3 drops an additional 3V or so to fix the voltage on the source at around half the supply voltage. Connecting the gate resistor R1 in this way applies a proportion of the in-phase output signal to its lower end, and the resulting positive feedback, or bootstrapping, increases its effective resistance, and the input impedance of the circuit, to around 6 megohms (6M). Output impedance is independent of signal source impedance. It is governed by the transconductance (gain) of the device, and is usually of the order of 500 ohms.

HIGH INPUT IMPEDANCE PREAMPLIFIER


R3 100

R1 1M C1 100n b

+9V TO +12V

TR1
BC549C c C3 100

C2 10

SIGNAL INPUT

+
R2 4k7 SCREEN
SIGNAL OUTPUT

SCREEN

0V

VOLTAGE GAIN: UNITY CURRENT DRAIN AT 9V SUPPLY: 125mA.

High Input Impedance Preamplfier circuit board.

Fig.5. High Input Impedance Preamplifier circuit diagram.

COMPONENTS
HIGH INPUT IMPEDANCE Resistors See R1 1M R2 4k7 R3 100W All 025W 5% carbon film page

+9V TO +12V
SCREENED SIGNAL INPUT LEAD R3 c R 1 C1 R 2 b e C3 SCREENED LEAD TO POWER AMPLIFIER

INPUT

TR1 +
C2

OUTPUT

SHOP TALK

TO COMMON 0V POINT ON POWER SUPPLY P.C.B.

Capacitors
C1 C2 C3 TR1 100n polyester 10m radial elect. 25V 100m radial elect. 25V BC549C npn transistor (or similar see text)
2.35IN (59.7mm)

Semiconductors
1.7IN (43.2mm)

Miscellaneous
Printed circuit board available from the EPE PCB Service, code 349 (Single Trans); audio screened cable; multistrand connecting wire; input and output sockets, type to choice; solder pins; solder etc.

349
Fig.6. Printed circuit board component layout, wiring and full-size copper foil master for the High Input Impedance Preamplifier.

Approx. Cost Guidance Only

Everyday Practical Electronics, June 2002

421

This is the circuit of choice when a high impedance source has to be connected to a long screened cable; e.g., a capacitor or crystal microphone. However, f.e.t. characteristics vary widely, and readers wishing to use the circuit of Fig.7 should be prepared to adjust the value of resistor R3, over the range of 1500 to 4700 ohms, especially when low supply voltages are used, in order to optimise signal handling capability.

LOW-NOISE PREAMPLIFIER
Amplifiers introduce unwanted noise and, as gain increases, more care has to be taken to prevent the noise becoming too intrusive. The noise generated by a bipolar transistor can be reduced by operating it at a low collector current, typically between 10mA and 50mA. This technique has been adopted for the first stage of the directlycoupled, two transistor, Low-Noise Preamplifier shown in Fig.9. Overall gain is stabilised by negative feedback applied via preset VR2. With the value shown, gain is approximately 300. If a 47k potentiometer is used instead, gain will be reduced to around 150, and it can be taken down to 70 or so with a 22k component. Rotating the slider (moving contact) of preset VR2 causes it to be progressively bypassed by capacitor C6, increasing the

CIRCUIT BOARD

Details of the printed circuit board component layout, wiring and copper foil master pattern are given in Fig.8. The board is the single transistor version and is available from the EPE PCB Service, code 349 (Single Trans). Before assembly, check the component, construction and interconnection details at the end of the article.

negative feedback, and reducing gain, at high frequencies. This feature is useful for reducing noise and for correcting the recording characteristic of long playing records. It is usual to incorporate more complicated RC networks in the VR2 position for the latter purpose but, unless the listener has a very refined ear, there will be little or no discernible difference. Operating conditions are stabilised by d.c. negative feedback applied via resistor R5. This, together with the high value collector load, R3, fixes the collector current of transistor TR1 at around 50mA with a 12V supply. Input impedance is around 50k, but the optimum signal source resistance for lowest noise is between 5k and 10k. This has influenced the value of the input potentiometer, VR1. The purpose of the remaining components will be evident from earlier circuit descriptions. However, because of the

F.E.T. HIGH INPUT IMPEDANCE PREAMPLIFIER


R4 100

TR1
2N3819 C1 100n d g s R1 2M2 R2 1k
SIGNAL INPUT
SIGNAL OUTPUT

C3 100 C2 10

+9V TO +12V

R3 1k8 SCREEN SCREEN 0V

VOLTAGE GAIN: UNITY CURRENT DRAIN AT 9V SUPPLY: 175mA.

F.E.T. High Input Impedance Preamplifier p.c.b.

Fig.7. Alternative circuit diagram for a High Input Impedance Preamplifier using a field effect transistor (f.e.t.).
SCREENED SIGNAL INPUT LEAD d g INPUT C1 R 1 R 2 R3 s

+9V TO +12V

+
C3

R4

SCREENED LEAD TO POWER AMPLIFIER

COMPONENTS
HIGH INPUT IMPEDANCE (F.E.T.) Resistors R1 2M2 See R2 1k R3 1k8 (see text) R4 100W All 025W 5% carbon film page

TR1 +
C2

OUTPUT

TO COMMON 0V POINT ON POWER SUPPLY P.C.B.

SHOP TALK
2.35IN (59.7mm)

Capacitors
C1 C2 C3 TR1 100n polyester 10m radial elect. 25V 100m radial elect. 25V

Semiconductors
1.7IN (43.2mm)

2N3819 n-channel field effect transistor (f.e.t.)

Miscellaneous
Printed circuit board available from the EPE PCB Service, code 349 (Single Trans); audio screened cable; multistrand connecting wire; input and output sockets, type to choice; solder pins; solder etc.

349
Fig.8. Printed circuit board component layout, wiring and full-size copper foil master for the F.E.T. High Input Impedance Preamplifier.

Approx. Cost Guidance Only

422

Everyday Practical Electronics, June 2002

higher gain, the supply line decoupling capacitor C7 has been increased in value to ensure stability.

LOW-NOISE PREAMPLIFIER

CIRCUIT BOARD

The printed circuit board component layout, wiring details and full-size copper foil master pattern for the Low-Noise Preamplifier are shown in Fig.10. This board is available from the EPE PCB Service, code 350 (Dual Trans.). See the general construction, component and interconnection guide-lines on the last page.

VARIATIONS

Some readers may wish to use this circuit with electret microphones which contain an internal line-powered f.e.t. amplifier. The load for this remote device is provided by resistor R1, and the supply voltage is reduced to around 45V, which is optimum for most microphones of this kind, by resistor R2. Decoupling is by means of capacitor C1. These components (R1, R2 and C2) should only be fitted if an electret microphone is used, as the circuit maintains a

Completed p.c.b. for the Low-Noise Preamplifier.


*R2
10k C8 10 R8 100

+9V TO +12V

*R1
1k

*SEE TEXT *C1


100 C2 47 C3 100n

R3 220k

R6 6k8

C5 10

TR2
BC549C

+
c C6 10n VR2 100k
SIGNAL OUTPUT

TR1
BC549C b e R5 220k VR1 10k R4 270 C4 100 c

b e

COMPONENTS
LOW-NOISE PREAMPLIFIER Resistors See R1* 1k R2* 10k R3, R5 220k (lownoise metal film page preferred) (2 off) R4 270W R6 6k8 R7 560W R8 100W All 025W 5% carbon film, except R3 and R5. *Only required if electret mic. used
SIGNAL INPUT

SHOP TALK

R7 560 C7 1000

SCREEN

SCREEN

0V

VOLTAGE GAIN 300 OVER hfe SPREAD OF 450 TO 600.

CURRENT DRAIN AT 9V SUPPLY: 1mA.

Fig.9. Circuit diagram for the Low-Noise Preamplifier. Components marked with an asterisk are only needed if an electret microphone is used. Increase the value of R2 to 18k with 12V supplies.

Potentiometers
VR1 VR2 10k enclosed carbon preset 100k SCREENED SIGNAL enclosed INPUT LEAD carbon preset
R1

R 2

R 3

R 6 C5

+
C7

+9V TO +12V
SCREENED LEAD TO POWER AMPLIFIER

+
C6 R 8

+ C2

+
C1 C3

c b e

TR1
R 5 R 4

c b e

OUTPUT

TR2
VR2

Capacitors
100m radial elect. 25V (2 off) C2 4m7 radial elect. 25V C3 100n polyester C5, C8 10m radial elect. 25V (2 off) C6 10n polyester C7 1000m radial elect. 25V *Only required if electret mic. used C1*, C4
VR1

C8 TO COMMON 0V POINT ON POWER SUPPLY P.C.B.

+
C4

R 7

3.4IN (87.5mm)

Semiconductors
TR1, TR2
1.8IN (46.5mm)

BC549C npn transistor (or similar see text) (2 off)

Miscellaneous
Printed circuit board available from the EPE PCB Service, code 350 (Dual Trans); audio screened cable; multistrand connecting wire; input and output sockets, type to choice; solder pins; solder etc.

Approx. Cost Guidance Only excluding microphone

350
Fig.10. Printed circuit board component layout, wiring and full-size copper foil master for the Low-Noise Two-Transistor Preamplifier.

Everyday Practical Electronics, June 2002

423

d.c. voltage on the input which could disturb the action of some signal sources. This circuit, and variations of it, form the basis of the front-ends of most high quality preamplifiers. With the component values shown, 33mV r.m.s. input will produce a 1V output before the onset of clipping. The noise introduced by the amplifier is about the same, or a little less, than that generated by the single transistor amplifier set for a gain of 150. The noise level could be further reduced by using low-noise, metal film resistors for R3 and R5.

Table 1: Reactance, in Ohms, of standard value capacitors at stated audio frequencies


Cap. 50 100 200 300 mF Hz Hz Hz Hz 1000 3W2 1W6 0W8 0W5 470 6W8 3W4 1W7 1W1 100 32 16 8 5 47 68 34 17 11 10 320 160 80 53 47 680 340 170 110 1 3k2 1k6 800 530 047 6k8 3k4 1k7 1k1 01 32k 16k 8k 5k3 0047 68k 34k 17k 11k 001 320k 160k 80k 53k 00047 680k 340k 170k 110k Reactance values rounded off 400 500 1 2 3 4 5 Hz Hz kHZ kHz kHz kHz kHz 4 3W2 1W6 8W5 6W8 3W4 1W7 1W1 40 32 16 8 5W3 4 3W2 85 68 34 17 11 8W5 6W8 400 320 160 80 53 40 32 850 680 340 170 110 85 68 4k 3k2 1k6 800 530 400 320 8k5 6k8 3k4 1k7 1k1 850 680 40k 32k 16k 8k 5k3 4k 3k2 85k 68k 34k 17k 11k 8k5 6k8 10 20 kHz kHz 1W6 3W4 1W7 16 8 34 17 160 80 340 170 1k6 800 3k4 1k7

FREQUENCY RESPONSE

Although inductors are sometimes used for tailoring the frequency response, the key components in networks which modify audio frequency response are normally capacitors. The resistance presented by a capacitor to the flow of alternating current (a.c.) decreases as frequency rises. This frequency dependant resistance is known as reactance. Capacitors combined with resistors form frequency dependant potential dividers which can be used to tailor the response. These RC networks can, of course, only attenuate signals. So called bass boost is obtained by reducing the response of the system to the higher audio frequencies. Table 1 lists the reactances of a range of standard capacitor values, at spot frequencies, across the audio spectrum. Referring to it, an 01mF (100nF) capacitor presents a resistance of 5300 ohms at a frequency of 300Hz. This rises to 32000 ohms at 50Hz and falls to 320 ohms at 5kHz. Fitting a blocking capacitor of this value to an amplifier with an input impedance of 5000 ohms will result in signal levels at 300Hz being halved. (Capacitor and input impedance act as a potential divider). This attenuation will increase as the frequency is lowered, and reduce as frequency is raised, at a rate of 6dB per octave. Fitting low value d.c. blocking capacitors to one or more stages will, therefore, roll-off the low frequency response. Capacitors connected from signal lines to ground; e.g. across the tracks of volume controls, will progressively attenuate high frequencies. Although simple, these measures can make a significant improvement in clarity and signal-to-noise ratio. Refer to Table 1 when selecting a capacitor to give the desired roll-off with a particular input impedance, then refine its value by trial and error.

Fig.11. This is the medium impedance transistor preamplifier illustrated in Fig.3 with negative feedback applied, via a frequency dependant network, from transistor TR1 collector to base. First published by P J Baxandall in 1952, the circuit has since been used, with minor variations, in most high quality preamplifiers. Potentiometers VR1 (Bass), and VR2 (Treble), control the impact of capacitors C1, C2 and C3 on the feedback network. Resistors R2 and R3 minimise interaction between the controls, and the circuit affords 15dB of boost or cut at 100Hz and 10kHz.

CIRCUIT BOARD

The printed circuit board component layout, wiring details and full-size copper foil master pattern are shown in Fig.12. This board is available from the EPE PCB Service, code 351 (Tone). Before undertaking any assembly work, see the general component, construction and interconnection notes at the end of the article.

IN-CIRCUIT

When circuits are cascaded, the Tone Control unit should always be the last in the chain; i.e. the one connected to the power amplifier. Most high quality preamplifiers consist of the two transistor circuit illustrated in Fig.9 followed by this Tone Control circuit.

The Bandpass Filter circuit diagram shown in Fig.13 cascades three high-pass (low frequency cut) sections between transistors TR1 and TR2, and three low-pass (high frequency cut) sections between TR2 and TR3. By this means, a roll-off of 18dB per octave is achieved above and below the desired frequency range. Filter networks of this kind need to be fed from a comparatively low impedance, and feed into a high impedance. The emitter follower stages, TR2 and TR3, are thus eminently suitable, and amplifiers of this kind have already been discussed. The input stage, transistor TR1, overcomes signal losses, or, with the slider of VR1 at TR1 emitter (e), ensures an overall circuit gain of around 25. Emitter to base feedback around TR2 and TR3, via the RC networks, improves the action of the filters. Component values have been selected to start the roll-off just within the pass band, and the response falls steeply below 300Hz and above 3kHz. Two capacitors have to be combined to produce a difficult-to-obtain value. To avoid confusion they are shown separately on the circuit diagram as C8 and C9.

CIRCUIT BOARD

BANDPASS FILTERS

FEEDBACK NETWORKS

Capacitors are used to make gain-reducing negative feedback networks frequency dependant; for example, capacitor C6 in the two-transistor Low-Noise Preamplifier shown in Fig.9. Reducing the emitter bypass capacitor C2, in the single transistor preamplifier shown in Fig.3, to 47mF, will progressively increase feedback, and reduce gain, as frequency lowers. This is another simple, but effective, way of securing low frequency roll-off.

TONE CONTROLS

Some means of continuously varying the frequency response is desirable when music is being reproduced, and a suitable Tone Control circuit diagram is given in

Reducing bandwidth to around 300Hz to 3kHz greatly improves the clarity of speech signals, and the practice is adopted by telephone companies around the world. Limiting the frequency response in this way significantly improves the signal-to-noise ratio. This is particularly desirable with sensitive radio equipment and surveillance systems, where the high level of amplification needed for the weakest signals brings with it a good deal of background and equipment generated noise. For best results, roll-off beyond the pass band should be fairly steep: the 6dB per octave afforded by a single RC combinaBandpass Filter (top) and Tone Control p.c.b.s. tion is not sufficient.

Details of the printed circuit board component layout, wiring and copper foil master are given in Fig.14. The Bandpass Filter board is also available from the EPE PCB Service, code 352 (Filter). See component, construction and interconnection notes before commencing building.

424

Everyday Practical Electronics, June 2002

TONE CONTROL

COMPONENTS
TONE CONTROL Resistors See R1, R3, R4, R6 4k7 (4 off) R2 27k R5 1M R7 470W page R8 100W All 025W 5% carbon film

SHOP TALK

Tone Control printed circuit board.


R8 100 C2 47n R6 4k7

Potentiometers
VR1, VR2 100k min. rotary carbon, linear (2 off) 2n2 polyester (2 off) 47n polyester 10m radial elect. 25V (2 off) 1m radial elect. 25V 47m radial elect. 25V 100m radial elect. 25V BC549C npn transistor (or similar see text)

Capacitors
+9V TO +12V

R1 4k7

VR1 100k

R4 4k7

C4 10

C1, C3 C2 C4, C8 C5 C6 C7

+
C8 10 R5 1M C5 1

R2 27k

TR1
BC549C c b e C7 100
SIGNAL OUTPUT

BASS

Semiconductors
TR1

R3 4k7
SIGNAL INPUT

Miscellaneous
SCREEN

C1 2n2

TREBLE

C3 2n2 C6 47

R7 470

SCREEN

VR2 100k 0V

B = BOOST END OF POTENTIOMETERS (VR1, VR2) MOVING CONTACT (SLIDER). VOLTAGE GAIN UNITY WHEN VR1 AND VR2 SET AT MID TRAVEL. BOOST AND CUT 15dB AT 100Hz AND 10kHz. CURRENT DRAIN AT 9V SUPPLY: 125mA.

Printed circuit board available from the EPE PCB Service, code 351 (Tone); metal case (optional), size and type to choice see text; audio screened cable; multistrand connecting wire; input and output sockets, type to choice; solder pins; solder etc.

Approx. Cost Guidance Only

Fig.11. Circuit diagram for the Tone Control (bass, treble boost and cut).
VR1 BASS

excluding case

+9V TO +12V
C2 R 1 R 2 C3 C1 R 3 C6 R 7

+ C4
R4

R 5

R 6

R8

SCREENED LEAD TO POWER AMPLIFIER

SCREENED SIGNAL INPUT LEAD

c C5

b e

+
C8

ANCHOR PIN FOR INPUT POTENTIAL DIVIDER RESISTORS

TR1

+
C7 TO COMMON 0V POINT ON POWER SUPPLY P.C.B.

VR2 TREBLE

2.8IN (71.0mm)

USE THIS RESISTOR NETWORK TO ATTENUATE INPUT SIGNAL. (SEE TEXT) R X SCREENED SIGNAL INPUT LEAD R Y

351
1.44IN (36.5mm)
C1

Fig.12. Tone Control printed circuit board component layout, interwiring and full-size copper foil master. The tape and CD player signal input attenuation resistors (see text) are shown in the inset diagram (left).

Everyday Practical Electronics, June 2002

425

SPEECH FREQUENCIES (300Hz TO 3kHz) BANDPASS FILTER


R12 100

R2 3k9 R1 1M C1 1 C3 10n C4 10n C5 10n

R5 1M

R10 1M

C11 100

+9V TO +12V

c b c R4 3k3 e R7 12k

TR2
BC549C

R8 12k

R9 12k

C10 100n b

TR3
BC549C C12 1

b e VR1 1k C2 47

BC549C

TR1

C7 22n

SIGNAL INPUT

SIGNAL OUTPUT

R3 6k8

R6 3k9

C6 15n

*C8
1n

*C9
470p

R11 3k9
SCREEN

SCREEN

*SEE TEXT

0V

VOLTAGE GAIN WITH PASSBAND, UNITY WITH VR1 SLIDER AT 0V RAIL; 25 WITH SLIDER AT TR1 EMITTER END. ROLL-OFF 18dB PER OCTAVE BELOW 30Hz AND ABOVE 3kHz. CURRENT DRAIN AT 9V SUPPLY: 4mA.

Fig.13. Circuit diagram for the Bandpass Filter for speech frequencies (300Hz - 3kHz).
+9V TO +12V SCREENED SIGNAL INPUT LEAD R 1 R 2 C5 R 5 R9 R 10 R12 SCREENED LEAD TO POWER AMPLIFIER

INPUT

+
C1

c b e C3 C4

c b e

R 8

c C10 C7 b e

+ TR3
R 11 C11

OUTPUT

TR1 +
C2 VR1 R 3 R 4

TR2
R 7 R 6 C6

+
C12

TO COMMON 0V POINT ON POWER SUPPLY P.C.B.

C8

C9

3.84IN (97.5mm)

Fig.14. Printed circuit board component layout, wiring and fullsize copper foil master for the Bandpass Filter.

352

COMPONENTS
BANDPASS FILTER Resistors R1, R5, R10 1M (3 off) R2, R6, R11 3k9 (3 off) R3 6k8 R4 3k3 R7 to R9 12k 1% metal film (3 off) R12 100W All 025W 5% carbon film, except R7 to R9

1.63IN (41.5mm)

See
Bandpass Filter printed circuit board.
C6 15n polyester C7 22n polyester C8* 1n polyester C9* 470p ceramic C10 100n polyester C11 100m radial elect. 25V *Combined (parallel) to give 1n5

SHOP TALK
page

Miscellaneous
Printed circuit board available from the EPE PCB Service, code 352 (Filter); audio screened cable; multistrand connecting wire; input and output sockets, type to choice; solder pins; solder etc.

Potentiometers
VR1 1k carbon preset 1m radial elect. 25V (2 off) 47m radial elect. 25V 10n polyester (5% or better) (3 off)

Capacitors
C1, C12 C2 C3 to C5

Semiconductors
TR1 to TR3 BC549C npn transistor (or similar see text) (3 off)

Approx. Cost Guidance Only

426

Everyday Practical Electronics, June 2002

PERSONAL EARPHONE JACK TIP RING

LEFT INPUT 470

47 SHANK TIP OV RAIL

RING SHANK

47

LINK HERE IF MONO INPUT REQUIRED

470 RIGHT INPUT

Fig.15. Method of connecting a Walkman tape or CD player.

SUMMARY

Operational amplifiers (op.amps) are more commonly used in filters of this kind but, when the need is simply for a unity gain buffer with a high input and low output impedance, the ubiquitous bipolar transistor can be made to serve our purpose just as well.

SIGNAL SOURCES

Radio Receivers The output from the detector or f.m. discriminator in a superhet radio receiver should fully load the power amplifiers described last month. After the usual filtering, the signal can be fed directly to the power amplifier, or via the Tone Control unit shown in Fig.11 and Fig.12. Microphones The single transistor preamplifiers shown in Fig.1 to Fig.8 will provide appropriate matching and sufficient gain for dynamic (moving coil), electret and crystal microphones when they are used for intercom purposes. (A circuit for line-powering electret microphones can be taken from Fig.9). The common emitter circuit given in Fig.3 should be used with moving coil units as these present an impedance of around 600 ohms. When electret or dynamic microphones are deployed for surveillance or sound capturing purposes, the two transistor circuit of Fig.9 will ensure a good degree of sensitivity. Electret microphones have an extended low frequency response. If this proves troublesome, reduce the value of the d.c. blocking capacitor C2. Try 47nF (0047mF) as a starting point. Gramophone Pick-ups The low output of moving-coil pick-ups necessitates the use of the two transistor preamplifier detailed in Fig.9. Omit preset VR1 and feed the signal to the base of transistor TR1 via capacitor C3. Low output ceramic pick-ups should be connected via a 1M (megohm) or 2M2 series resistor to preserve low frequency response. The F.E.T. Preamplifier circuit illustrated in Fig.7 is more suitable for high output ceramic and crystal pick-ups. Personal Tape and CD Players An arrangement for extracting the signal from personal cassette players and headphone radios is given in Fig.15. The 47 ohm resistors substitute for the 32 ohm earpieces, and the 470 ohm resistors attenuate the signal.

The chosen system Fig.16. Circuit arrangement for a stereo Balance control. must, of course, be It may help to start construction by first duplicated if stereo operation is required. placing and soldering in position the variTone and Volume controls are usually ous wire links on the chosen preamplifier ganged, and an additional potentiometer is p.c.b. This should be followed by the leadprovided to balance the gain of the two off solder pins, and then the smallest comchannels. ponents (resistors) working up to the With the simple circuit arrangement largest, electrolytic capacitors and presets. shown in Fig.16, the Balance potentiomeFinally, the lead-off wires (including the ter is connected across the ganged Volume screened cables) should be attached to the controls at the inputs to the two power p.c.b. amplifiers (VR1 on the power amplifier On completion, check the orientation of circuit diagrams). electrolytic capacitors and transistors, and examine the board for poor connections COMPONENTS and bridged tracks, before connecting the All of the components, for this part of power supply. The approximate current the series, are readily available from a varidrains are included with the circuit ety of sources. Transistor types are not critdiagrams. ical and almost any small-signal npn device will function in the circuits. INTERCONNECTIONS A low-noise, high gain transistor will, Overall voltage gain can be in excess of however, ensure the best performance, and 2000, and care must be taken to avoid hum the base connections for some alternative pick-up and instability. types are given in Fig.17. With European Hum pick-up is of two kinds, capacitatransistors, the suffix C indicates the tive and inductive. High impedance circuits highest gain grouping. are prone to the former, and low impedance to the latter. Housing the pre- and power b amplifiers in a metal case will do much to e c minimise these problems. e b c b c e If hum increases when a finger is BC549C BC169C BC109C BC239C 2N3711 brought near to the preamplifier, the pickup is capacitative. It can usually be cured by providing an earthed metal screen around the input wiring or even the entire g g g d s d s s d preamplifier board. BF245 MPF102 2N3819 BF244A All mains and a.c. power leads within UNDERSIDE VIEWS the metal case of the unit must be tightly twisted to minimise external fields, and the Fig.17. Base connections for suitable mains transformer should be sited at least transistors and f.e.t.s. 150mm (6in) from the input circuitry. Tightly twist power amplifier output leads, and keep them as far away as possible from If possible, use transistors with an hfe of preamplifier inputs. Keep all leads as short at least 450 for the input stage of the Lowas possible. Noise Preamplifier and for the various Run a separate negative power supply emitter follower stages (where high input connection from each of the p.c.b.s to a impedance depends on the use of a high common 0V point on the power supply gain device). board, or to the negative battery terminal. Do not connect one circuit board via CONSTRUCTION another to supply negative, or rely upon All the preamplifiers covered in this screened cable braiding or a metal case to part are assembled on printed circuit provide this connection. Make only one boards and construction is reasonably connection to any metal case, close to the straightforward. Solder pins, inserted at negative terminal on the power supply the lead-out points, will simplify any offp.c.b. board wiring. Remember to earth the If all of the above measures have been metal bodies of rotary potentiometers and adopted and hum problems still persist, try to use screened audio (mic.) cable for the disconnecting, one by one, the screens of leads to tone and volume controls to minthe audio cables, at one end only. imise hum pick-up. Reorientating the mains transformer can The single transistor preamplifiers all also effect a cure. use the same p.c.b. and wire links are required. If units are cascaded, and couNext Month: Mains power supplies, pling capacitors deleted, remember to loudspeakers and signal filtering will be install wire links to maintain the signal discussed. path.

STEREO

Everyday Practical Electronics, June 2002

Preamplification is not required, but readers may wish to use the Tone Control unit to process the signal. Provision is accordingly made, on the Tone Control p.c.b. illustrated in Fig.12, for a signal attenuating network; resistors Rx and Ry.

BALANCE CONTROL

LEFT CHANNEL INPUT

VR1a 10k
GANGED VOLUME CONTROLS TO POWER AMP LEFT

22k

RIGHT CHANNEL INPUT

TO POWER AMP RIGHT

VR1b 10k

427

SIMPLE AUDIO CIRCUITS


Part 4 Loudspeaker Enclosures, Tuning Oscillator and Resonance Detector

Constructional Project

RAYMOND HAIGH
A selection of pic-n-mix low-cost audio circuits from preamplifier to speaker!

and science collide in the design of loudspeaker enclosures and, transcending all the conflicting opinions, is the way a vibrating paper cone can reproduce sounds ranging from the human voice to a symphony orchestra with vivid realism. Last month we discussed speakers and crossover networks. In this final instalment, enclosures and the simple test equipment needed to optimise performance are covered.
RT

Sound waves formed by the front of the speaker cone are out of phase with those at the back. If the pressure variations can leak around the cone there will be cancellation, particularly at low frequencies, and sound output will be reduced. The primary duty of the enclosure is, therefore, to prevent this leakage.

WHY AN ENCLOSURE?

Speaker cones have a natural resonant frequency (just like a guitar string). The greater the mass of the cone, and the freer its suspension, then the lower the resonant frequency. At resonance, very little energy is required to make the cone vibrate vigorously. This has electrical drawbacks, which were discussed last month. It is also undesirable from an acoustical point of view for speaker sensitivity to peak sharply at one frequency. The second requirement of the enclosure is, therefore, to retain a volume of air which damps the cone and evens out the response of the system.

absorb the sound output from the rear of the speaker. Air trapped inside the box damps the cone, raising its resonant frequency by up to an octave (a doubling). Low frequency output falls off rapidly below resonance, and special speakers with high mass, high compliance (very low resonance) cones are sometimes used to offset the rise in resonant frequency. Absorption of the energy delivered by the rear of the cone, together with the high cone mass, result in an acoustic efficiency as low as 1 per cent. Our Twin TDA2003 125W Amplifier (82W into 8 ohms: see Part One) requires a more efficient speaker than this if windows are to rattle. Acoustic Labyrinth Acoustic labyrinth enclosures are, in effect, a duct one quarter of a wavelength long at the speakers resonant frequency (e.g., 7ft at 40Hz). Folding the fibreboard or plywood duct into a box shape produces a labyrinth, hence the name. Some designers fill the duct with acoustic wadding: others just line the interior surfaces.

ENCLOSURE TYPES

Ignoring simple open baffles, there are four basic types of enclosure. Infinite Baffles. Infinite baffles are no more than sealed boxes filled with acoustic wadding to

CURVE A WITH SEALED ENCLOSURE

CURVE B WITH VENTED AND TUNED ENCLOSURE

Fig.1. Speech coil impedance in region of resonance.

588

Everyday Practical Electronics, August 2002

The quarter wavelength air column imposes the desired heavy damping on the cone at its resonant frequency. As frequency rises through an octave (i.e., towards 80Hz in our example) the air column approaches half a wavelength. The phase of the radiation from the rear of the cone is then inverted, and it emerges from the duct to reinforce that from the front, thereby increasing output. Enclosures of this kind are not easy to construct or tune to suit different speakers. In our quest for good performance for a modest outlay of cash and effort, this highly regarded system has, therefore, to be rejected. Horns Loading the speaker cone with an expanding column of air in the shape of a horn results in very high efficiencies; of the order of 40 per cent to 50 per cent. The horn effects an impedance transfer: high at the throat and low at the mouth. The resulting heavy damping on the speaker cone, and the small cone excursions and low power input needed for a given sound output, greatly reduce distortion. Many ingenious designs have been produced for folding large, low frequency horns into cabinets. However, cost, size, and complexity of design and construction remove this system from our consideration. Bass Reflex Bass reflex enclosures, also known as acoustic phase inverters, are based on the work of a German physicist, Herman Ludwig Ferdinand von Helmholtz (18211894). Whilst exploring the nature of sound, he investigated the way air resonates inside vented chambers and close to the vent itself. The idea of mounting a loudspeaker in a Helmholtz resonator was patented, about half-a-century later, by A. L. Thuras. Enclosures of this kind are simple and cheap to construct and tune. Efficiency is comparatively high: some authorities suggest 15 per cent to 20 per cent depending on the size of the loudspeaker (the bigger the better). A reflex enclosure is, therefore, the natural choice when cost and effort are to be kept to a minimum and limited amplifier power demands good speaker efficiency.

A bass reflex enclosure is no more than a box with a small opening known as the vent or port. The mass of air within the box is tuned, by the vent, to resonate at the same frequency as the speaker cone. This imposes heavy damping and results in two smaller resonances, one of lower and one of higher frequency than the unvented cone resonance. Speaker output falls off rapidly below resonance, and the development of the lower frequency peak extends the speakers bass response by almost an octave. Phase inversion takes place over most of the low frequency range, and output from the vent augments that from the front of the cone (the operation of the system is complex, and phase inversion does not occur at all frequencies). Output falls off very rapidly below the lower peak but, in a well designed system, this will be in a region where there is little or no signal content. The damping effect of the vented enclosure is displayed graphically in Fig.1. A plot of speech coil voltage against frequency, it represents variations in impedance which are intimately related to resonances in the system. The single resonant peak (curve A) developed when the vent is sealed contrasts with the two lower peaks (curve B) which form when the vent is opened. Correct tuning is indicated when the peaks are of equal magnitude (as is the case here).

HOW IT WORKS

Crossover/Audio Filter selection switch and amplifier input terminals.


25 per cent and tune to resonance by reducing the vent area or providing a duct. When reflex enclosures are designed in this way, the frequency ratio between the two smaller resonances formed by tuning should be not less than 15:1 and not more than 24:1.

MODERN PRACTICE

DESIGN TECHNIQUES

Traditionally, designers matched enclosure resonance to the free-air resonance of the speaker cone on the basis of vent area being equal to effective cone area. This optimised low-frequency reinforcement by the vent but resulted in large enclosures. Readers who like to build on a grand scale might find the formulae in Table 1 helpful. Much simplified, they relate speaker size and cone resonance to enclosure volume. The relevant speaker parameters are listed in Table 2. Enclosures as large as this tune very broadly, and sizeable variations in vent area have only a modest effect on performance. As we shall see, enclosures can be too big, and it would be prudent to reduce the volume given by the formulae by, say,

During the 1960s, Australians, Neville Thiele and Richard Small, extended earlier loudspeaker research carried out by American, James Novak. They were able to show that, for optimum performance, enclosure size is dependant upon the relationship between the damping effect of the enclosed air and the compliance of the cone suspension. If, when the enclosure vent is sealed, the frequency of the single resonant peak is 15 to 16 times the free-air resonant frequency of the cone, the relationship is correct. Thiele and Small described an experimental method for determining suspension compliance, and produced formulae relating this, and other speaker properties, to enclosure size and vent area. Known as the Thiele-Small parameters, these speaker characteristics are now published by a number of manufacturers.

TABLE 1: TRADITIONAL ENCLOSURE DESIGN


Formulae relating enclosure volume to speaker cone size and resonant frequency f res Hz 40 Vol cu ft 3R 50 2R 60 14R 70 1R 80 08R 90 06R 100 05R 110 04R

TABLE 2: LOUDSPEAKER DATA Speaker Diameter (inches) Effective cone radius R in. Effective cone area sq. in. 8 3 28 10 375 44 12 475 71 15 6 113 18 75 177

Notes: (1) F res is the free air resonant frequency of the cone, in Hertz. Vol is the internal volume of the enclosure in cubic feet. R is the effective radius of the speaker cone in inches (see Table 2). (2) These formulae are derived from traditional design procedures. Calculations in accordance with current practice, which relates cone compliance to enclosed air compliance, usually result in a smaller enclosure (see text). (3) Although much simplified, the formulae will produce sufficiently accurate results (as size increases towards this maximum, tuning becomes less and less critical). (4) Formulae are based on enclosure port area being equal to the effective cone area. See Table 2 for details of effective cone areas.

Everyday Practical Electronics, August 2002

589

Cabinet dimensions should not be exact multiples of one another, and some experts maintain that deep enclosures perform better than shallow ones. Greater depth also permits a longer duct. Chamfers, formed around the enclosure front and reaching almost to the speaker aperture, are said to improve clarity at low frequencies, but this makes construction difficult. Keeping the front panel as narrow as possible is probably the best we can do to achieve this objective. The vent can be any shape provided its smallest dimension is not less than one inch. Circular vents can be ducted with a length of cardboard tube, but some builders may find rectangular openings and boxform ducts easier to fabricate.

liberal quantities of adhesive to fill the gaps. Use plastic foam draught excluder to seal the access panel.

MAKING DUCTS

Ducts need not be as rigid as the enclosures, and hardboard (Masonite in the USA) or very thick cardboard are suitable materials. Circular ducts can be formed by applying paste to a long strip of paper or thin card and winding it around a food or paint container until a thickness of 3mm (1/8in.) or so has been built up. Slide the duct from the former and place it somewhere warm for the paste to dry. It is not too difficult to combine two pipes to form an adjustable, telescopic duct.

D. B. Keel subsequently adapted the formulae for processing on a pocket calculator, but the procedure is still complicated. Readers with a mathematical turn of mind who want to optimise their enclosures in this way are urged to study the extensive literature on the subject.

BUILD AND TUNE

Theile-Small parameters are not usually available for the low cost, but often reasonable quality, speakers of Far Eastern origin (or for speakers in spares boxes). Even if they were, it is likely that many readers couldnt face the tedium of the calculations. An alternative approach is to make an enclosure of manageable dimensions, having regard to the size of speaker, and then tune it to optimise performance. Quite small enclosures can be tuned to frequencies in the 50Hz to 100Hz range. However, as volume is reduced vent area has to be reduced to secure resonance at a particular frequency. Eventually, a point is reached when vent output is negligible and the enclosure is performing almost like a sealed box. Moreover, as size is reduced, the smaller, stiffer volume of air increases damping on the cone and its resonant frequency rises unacceptably. The resonant frequency of a given vent and enclosure combination can be lowered by forming a duct or pipe behind the vent. The longer the duct the lower the resonant frequency. Although this involves more constructional effort, it does allow a reasonable vent area to be maintained when enclosure volume is small.

SIZE AND SHAPE

Speaker units were discussed last month, and it was clear that an extended and powerful low-frequency response becomes easier to achieve as speaker size is increased. It was suggested that speaker size ought not to be less than 8in, and this is especially true when an inexpensive unit is to be fitted. Readers may wish to use even larger speakers for the advantages they offer: some highly regarded studio monitors comprise a 15in bass unit in a 5 cubic foot reflex enclosure.

Tweeters can be mounted axially in front The above requirements, together with of the bass speaker to avoid the need for the desirability of a reasonable vent area another hole in the cabinet. Small hooks and the obvious influence of speaker diamand eyes and the kind of springy wire used eter, tend to determine the smallest acceptfor hanging net curtains are ideal for this able enclosure size. Suggested internal purpose. dimensions to suit standard speakers are If the wires are cut short to provide a litlisted in Table 3 and the general make-up tle tension the speaker will be held firmly of the enclosure is shown in Fig.2. in place. Strong rubber bands could be The enclosures for the 15in and 18in units used, but these may perish over time. are rather deep, and the speaker aperture and vent opening could be formed on the face with the larger dimension if desired (these cabinets are large enough for the cone to still be an adequate distance from what would then be the back). Whilst the width of the front is determined by the speaker chassis and cannot be reduced much, the other dimensions can be changed to suit materials that Using cutdown curtain wire, hooks and eyes to suspend the are to hand or a treble speaker over the bass speaker. particular space Bass reflex cabinets are resonators and in a room. When making changes, try not acoustic treatment should be applied sparto reduce the volume by more than 10 per ingly. The rear and top of the enclosure cent or so (especially with the 8in. and should, however, be lined with about 50mm 10in. units); and try to avoid dimension (2in.) of cellulose wadding to prevent the combinations that are exact multiples. reflection of mid-frequency sounds which CONSTRUCTION could otherwise escape through the speaker One of the best materials for cabinet cone and impair clarity. construction, acoustically speaking, is Cellulose wadding can be obtained from medium density fibreboard (MDF). This upholsterers and craft shops (it is used for material is reasonably heavy, easy to work, stuffing soft toys). has a desirable dead quality and is inexTESTBENCH SPEAKER pensive. Chipboard, blockboard and plyThe accompanying photographs show an wood are also perfectly acceptable. enclosure for an 8in. speaker, constructed Enclosures for the 8in., 10in. and 12in. in accordance with the earlier guidelines, speakers should be formed from 13mm and incorporating the crossover and audio (1/2in.) thick sheet with 19mm (3/4in.) filter unit described last month. It is intendsquare glued and screwed softwood corner ed for workshop use, and this is reflected in fillets. The two larger enclosures require the style and type of finish. Constructors 19mm (3/4in.) material and 25mm (1in.) wanting hi-fi speakers will have their square fillets. One or two lengths of 25mm own ideas for giving the units a more square softwood should be fixed across the domestic appearance. larger enclosures, from side-to-side, near The surface mounted grille is of the type mid panel, to inhibit vibrations. fitted to musicians speakers. The bezel The construction must be air-tight. If around the vent opening is formed from any of the joints are less than perfect, apply

CABINET SIZES

TWEETER MOUNTING

590

Everyday Practical Electronics, August 2002

TESTBENCH LOUDSPEAKER ENCLOSURE


TABLE 3: RECOMMENDED MINIMUM ENCLOSURE DIMENSIONS
Speaker Diameter Width A Height B Depth C Speaker Aperture diameter D Vent diameter E Vent area sq. in. Minimum distance F Enclosure Volume (cu. in.) Enclosure Volume (cu. ft.) 8 95 15 12 7 4 125 3 1710 1 10 115 18 145 9 5 195 4 3002 175 12 135 21 17 11 6 28 5 4820 275 15 17 27 21 1375 7 38 7 9639 55 18 20 33 24 165 8 50 8 15840 9 The parts list for the crossover unit was included with Part 3, last month.

LOUDSPEAKER ENCLOSURE . . . YOU WILL NEED


Bass Speaker: 8in. diameter, 8 ohms impedance, preferably with a free-air resonance below 70Hz (most speakers with a rolled surround will meet this requirement). Moving coil treble unit, 8ohms impedance (see text). Sheet of MDF, 1200mm x 600mm x 13mm (4ft x 2ft x 1/2in.) thick; softwood corner fillets 4m 19mm square (13ft of 3/4in. square); glue and screws. Speaker and vent grilles; material for any duct (see text); draught excluding strip; springy curtain wire and small hooks for mounting tweeter unit; finishing materials etc.

Notes: (1) All dimensions are in inches unless otherwise stated. (2) Enclosure volumes expressed in cubic feet are approximate. (3) Enclosures produced to these dimensions must be tuned for optimum performance (see text).

D B

Fig.2. Front and side elevations showing the speaker and vent apertures. Recommended enclosure dimensions are listed in Table 3 above.
hardboard and nylon mesh is used as a screen. Bezel and mesh are spray finished matt black. Photographs of the tweeter mounting were taken before the suspension wires were painted black to conceal them behind the grille. Car spray paints were used to decorate the cabinet, and the hard, smooth surface of the MDF makes it easy to obtain a good finish (spraying should be undertaken outdoors or where there is plenty of ventillation). Rub-down lettering, protected by varnish, is used for the panel annotations.

Main speaker and crossover filter (last month) mounted on the rear of the enclosure front panel.
Suitable tweeters are readily available at a fairly reasonable cost. The paper-coned unit mounted in the prototype is a cheap surplus component. It is sometimes desirable to adopt a cross-over frequency around 500Hz when large (15in. or 18in.) bass speakers are used. Suitable tweeters can be expensive, and experimentally minded readers may care to try one of the cheap Mylar cone speakers intended for alarm systems. The claimed frequency response extends up to 20kHz, and a 3in. or larger unit should cope with the lower cross-over frequency. Chassis perforations should be covered with several layers of sticky tape to prevent interaction with the bass speaker. Alternatively, isolate the tweeter by mounting it inside a small box formed within the main enclosure. Fill the box with cellulose wadding. A 3in. diameter Mylar cone speaker performed better than the purposemade tweeter mentioned above.

SPEAKERS

Lining the rear of the cabinet with sound-absorbent wadding.

Manufactured in the Far East, the bass speaker used in the model is an inexpensive 8in. diameter unit with a rolled surround. Speakers of this kind are widely retailed and cost between 8 and 15 ($12 and $22). A compliant suspension and robust cone give these units a free-air resonance in the region of 60Hz. Speakers with a free-air resonance much higher than 70Hz should be avoided if possible.

Everyday Practical Electronics, August 2002

591

TUNING OSCILLATOR
In order to tune our enclosure we need some means of exciting and detecting resonances. A simple Low Frequency Oscillator circuit diagram is shown in Fig.3, where IC1, a 741 op.amp, provides the necessary gain. A Wien bridge network, formed by C1, C2, R1, R2 and VR1a and VR1b, controls the phase of the positive feedback from IC1 output (pin 6) to the non-inverting input (pin 3). Potentiometer VR1 sets the frequency of oscillation. Negative feedback, from the output to the inverting input (pin 2), determines the gain, thereby controlling the level of positive feedback. Gain should be as low as possible consistent with reliable oscillation over the full swing of Frequency control VR1. Negative feedback increases, and gain reduces, as the slider (moving contact) of preset potentiometer VR2 is rotated towards resistor R3. The stabilising circuit usually incorporated into the negative feedback loop has been omitted in the interests of simplicity. Despite this, signal amplitude is constant over the frequency range and waveform is good when VR2 is correctly set.

OSCILLATOR CONSTRUCTION

Most of the oscillator components are assembled on a small single-sided printed circuit board (p.c.b.). This board is available from the EPE PCB Service, code 364. The topside component layout, interwiring and full-size underside copper foil master pattern for the Low Frequency Oscillator board are shown in Fig.4. Solder pins, inserted at the lead-out points,

simplify off-board wiring, and a holder for IC1 facilitates substitution checking.

R1 2k7 VR1a 10k ON/OFF C2 470n OUTPUT 7 3 2

S1a

+9V

+ IC1
741 4

SK1

VR1b 10k

R3 820 VR2 100 B1 9V

FREQUENCY

0V C1 470n R2 2k7 R4 390 B2 9V S1b 9V

Component layout on the completed circuit board.

Fig.3. Circuit diagram for a simple Low Frequency Oscillator for loudspeaker resonance checking.

COMPONENTS
Resistors
S1a S1b TO B1 +9V TO B2 9V

OSCILLATOR

R1 C 2 R3

R1, R2 2k7 (2 off) R3 820W R4 390W All 025W 5% carbon film

See
page

SHOP TALK

ON/OFF

VR1a C 1

IC1

Potentiometers
PHONO SOCKET SK1

VR1 VR2

VR1b

10k dual-ganged rotary carbon, lin. 100W enclosed carbon preset 470n polyester layer, 5% tolerance desirable (2 off) 741 gen. purpose op.amp

R2

R4

VR2

OUTPUT TO B1 VE

Capacitors
C1, C2

2.3IN (58.4mm)

TO B2 +VE

Semiconductors
FREQUENCY

IC1

364

Miscellaneous
1.3IN (33.0mm)

Printed circuit board available from the EPE PCB Service, code 364; small plastic case, size and type to choice; PP3 batteries and holders; pointed control knob; 8-pin i.c. holder; solder pins; multistrand connecting wire.

Fig.4. Low Frequency Oscillator printed circuit board component layout, interwiring to off-board components and full-size underside copper foil master pattern.

Approx. Cost Guidance Only

excluding batts.

592

Everyday Practical Electronics, August 2002

40 35

45 50 55 60 70 80 90 100 120

30

Packing the Low Frequency Oscillator components on the rear of the small plastic box lid.
Potentiometer VR1, On/Off switch S1, the p.c.b. and the batteries can be housed in a small plastic box. The compact internal layout inside the prototype unit is shown in the photographs. It is not necessary to know the precise frequency to tune the enclosure, but an approximate idea is useful. Component

26

126

Hz

tolerances will affect calibration, but the original dial should provide an approximate guide to the frequency control settings on other units. It is reproduced, full-size, in Fig.5.

Fig.5. Full-size front panel dial as used in the prototype Low Frequency Oscillator.

RESONANCE DETECTOR
Some test meters, set to the lowest a.c. range, could be used to monitor the voltage developed across the speech coil. However, unless the meter is sensitive, the sound level from the speaker under test would be distressingly loud. Further, a resistor has to be wired in series with the speech coil to facilitate the test. This could make it difficult for the amplifier to deliver sufficient output to produce a reading on an insensitive meter. Greater sensitivity can be achieved by rectifying the signal C1 and measuring the R1 D2 1 47 OA47 resultant d.c. on the a k lowest testmeter + range. A suitable TO DIGITAL + k loudspeaker C2 TO INPUT FROM R2 OR MOVING D1 SPEAKER AMPLIFIER 1 COIL METER 220k Resonance Detector OA47 a (2V RANGE) circuit is given in Fig.6, where diodes D1 and D2 are configured as a voltage Fig.6. Circuit diagram for the loudspeaker Resonance Detector.
+

doubler delivering almost the peak-to-peak value of the signal. When the Resonance Detector unit is connected to a high impedance digital meter, reservoir capacitor C2 slows the response to voltage changes, and resistor R2 is included to reduce the delay. Series resistor R1 increases the impedance of the signal source and magnifies the effect of changes in the impedance of the speech coil. The values of electrolytic capacitors C1 and C2 have been chosen to suit the frequencies involved.

DETECTOR CONSTRUCTION

All the components for the Resonance Detector are assembled on a small printed

COMPONENTS
Resistors

a TO POWER AMPLIFIER R 1

D2

RESONANCE DETECTOR

+ +
C1 D1 a

+
k C2

R 2

R1 47W R2 220k All 025W 5% carbon film

See

Capacitors
C1, C2
TESTMETER SET TO 2V D.C. RANGE

page 1m radial elect. 25V (2 off)

SHOP TALK

Semiconductors
D1, D2 OA47 or OA90 germanium diode (1N914 silicon if lower sensitivity can be tolerated see text) (2 off)

SPEAKER 2.15IN (54.6mm)

Miscellaneous

365

1.1IN (27.9mm)

Printed circuit board available from the EPE PCB Service, code 365; multistrand connecting wire; solder pins; solder, etc.

Fig.7. Printed circuit board component layout, interwiring details and full-size underside copper foil master for the loudspeaker Reasonance Detector.

Approx. Cost Guidance Only excluding speakers

Everyday Practical Electronics, August 2002

593

circuit board (p.c.b.). This board is available from the EPE PCB Service, code 365. The p.c.b. component layout, wiring and full-size underside copper foil master pattern details are illustrated in Fig.7. Construction is very straightforward and only the polarity of the capacitors and diodes needs special attention. Also, germanium signal diodes, D1 and D2, can be damaged by excessive heat and it is prudent to leave a good lead length and apply a heat shunt when soldering.

volume, so err on the long side when adjusting its length in this way. The speaker unit has an extended bass response and, when driven by the 8W amplifier described in Part One (May 02), sound levels are more than sufficient for a domestic hi-fi installation. Vent output makes a significant contribution at low frequencies (it will extinguish a candle held close to the aperture), and there are no audible resonances. The speaker is most certainly not a boom box with honking, one-note bass. The middle range is clear but there is some colouration at high power levels with music that has a heavy bass content. Performance at the higher audio frequencies depends very much on the tweeter used: the enclosure is certainly worth something better than the cheap unit fitted in the prototype. When the crossover network is switched to act as a speech frequency bandpass filter, signals overlaid by noise are greatly clarified. Communications enthusiasts, or readers involved in surveillance, may find this circuit of interest. It certainly makes the unit more versatile as a bench speaker.

PERFORMANCE

Completed circuit board for the Resonance Detector.


frequency and magnitude of the peak. It will now be at a higher frequency than the free-air resonance. Open the vent and sweep the oscillator, again noting the frequency and magnitude of the peaks. If the tuning is correct (most unlikely), two peaks of equal magnitude will be revealed on either side of the original, vent-sealed peak. If the higher frequency peak is of greater magnitude, the vent area is too small (or any duct attached to it too long). Enlarge the vent, or shorten the duct, and test again. If the lower frequency peak is of greater magnitude (more likely with the

GENERAL SUMMARY

No difficulty should be encountered obtaining any of the materials and components needed for the construction of the loudspeaker enclosure and the setting up equipment. Details of the cross-over unit were given last month. Silicon diodes (type 1N914) can be used in place of the germanium devices in the voltage doubling rectifier circuit of the Resonance Detector. The higher knee voltage (06V instead of around 02V) reduces sensitivity, but they will still reveal the resonance peaks when the sound from the speaker is not too loud, and this is the main requirement.

POWER CHECK
LOUDSPEAKER UNDER TEST

10k INPUT ATTENUATOR

OSCILLATOR

POWER AMPLIFIER

RESONANCE DETECTOR

TEST METER

The Low Frequency Oscillator and Resonance Detector units can, of course, be used to investigate any speaker system. The rating of resistor R1 in the Resonance Detector is only sufficient for testing at comfortable listening levels. If speakers are to be checked at high power, fit a 5W component and use silicon instead of germanium rectifier diodes. Although the test equipment will respond to very slight changes in venting, especially when the enclosure is small, only a refined ear could detect any audible difference, even when quite large adjustments are made. 6

Fig.8. Block schematic diagram showing the interconnecting set-up for checking speaker resonances.

FREE-AIR RESONANCE

The free-air resonance of the bass speaker should be checked before embarking on the construction of the enclosure. To do this, wire up the test circuit shown in Fig.8. Details of the connections to the Resonance Detector are given in Fig.7. The Oscillator output is in the region of 45V r.m.s., and the 10 kilohm input attenuator potentiometer will have to be turned well down. Hold the speaker, by the magnet, well away from other objects and sweep the Oscillator until the voltage across the speech coil peaks. The rise will be sudden and dramatic. Note the reading on the Oscillator dial. If an extended low frequency response is important, it ought not to be more than 70Hz.

design guidance given here), the vent area is too large or any duct is not long enough. Either reduce the vent area, add a duct, or increase the length of any duct already fitted, and test again. Repeat the procedure until the two peaks are of equal magnitude. Some experts tune to a slightly higher frequency. This depresses the higher frequency peak and, it is claimed, results in a more uniform bass response. The impedance plot of the test bench speaker, after tuning, is given in Fig.1.

DUCTING

ENCLOSURE TUNING

With the speaker now in the enclosure, connect it to the test circuit shown in Fig.7 (directly, not via the crossover). Seal the vent, sweep the oscillator and note the

It is preferable to install a duct, rather than reduce vent area, in order to lower resonant frequency. Hold ducts in place with sticky tape during the setting up process. If desired, a duct can be mounted externally and adjusted until its length is almost correct before fixing it behind the vent. Duct volume will then reduce cabinet

594

Everyday Practical Electronics, August 2002

GUITAR PRACTICE AMP


BART TREPAK
A low-cost amplifier that will allow the budding guitarist to improve his playing technique, without annoying the household or neighbours!
OMMERCIAL guitar amplifiers, even those intended for practising, tend to be fairly expensive and have many features such as gain and tone controls which are seldom used, while lacking more useful ones such as an extra input for a microphone or another guitar. The budding musicians money could be better spent on other accessories or even a better guitar, especially as a simple practice amplifier for use with headphones can easily be built around a cheap integrated circuit. Even a more ambitious version for driving a speaker providing an output of a few watts, which would be quite loud enough to annoy the neighbours or for playing in a small hall, only requires the addition of a cheap power amplifier i.c and a few more components.

Constructional Project

been designed for easy construction with virtually no off-board wiring apart from the mains transformer, speaker and an optional headphone socket. Since the printed circuit board is readily available, the circuit can be knocked up in a very short time and you should have some change from 25. The finished circuit can be mounted in the same cabinet as the speaker (these can be salvaged from a defunct hi-fi unit) and even if a speaker has to be purchased separately it should not set you back very much.

AMPLIFIER CIRCUIT

Although the cost and number of components required is small, audio power amplifier circuits do not lend themselves to a simple stripboard layout and the problems associated with designing and making a suitable printed circuit board are likely to put off all but the most cost conscious or determined constructors. The simple project to be described here solves this problem and has

EASY-BUILD

The full circuit diagram of the Guitar Practice Amp shown in Fig.1 is very conventional and consists of an inverting preamplifier stage, IC1, feeding a single chip power amplifier, IC2. The op.amp preamplifier IC1 has a variable gain set by preset VR1 to enable this to be set to any required level (up to 100) and should, therefore, be suitable for even the most inefficient guitar pick-ups. Many small commercial guitar amps often feature tone controls but these are really superfluous as most electric guitars have perfectly adequate tone controls fitted and so these have not been included in this design.

The output of the preamplifier stage (IC1 pin 6) is fed via Volume control VR2 to the power amplifier IC2, which is based around the popular TDA2030. This device can supply up to 24W of audio power depending on the supply voltage and speaker impedance used, provided we are not too bothered about the distortion which in this application can almost be considered to be an advantage. With the lower supply voltage specified, a more reasonable output power would be about 6W to 10W which should be more than sufficient for our purpose. The power output can be easily increased if required by reducing the speaker impedance or increasing the supply voltage, and no changes in the component values are required. It should, however, be remembered that the maximum supply voltage for both i.c.s is 36V. The TDA2030 is a very well protected device featuring both short circuit and over dissipation protection although from a reliability point of view it is certainly not advisable to run the device in either of these conditions. Music generally tends to have many peaks while the average power dissipated remains low so that in practice, despite the use of the relatively small heatsink specified, the temperature of the device will remain well within its safe limit even with prolonged loud playing. Also, as the circuit is permanently connected to the speaker (except when in headphone mode) the possibility of a short circuited output is much reduced. An (optional) output socket SK3 is also wired in circuit to enable headphones to be connected in place of the speaker LS1.

84

Everyday Practical Electronics, February 2002

9V

+VE

C13 2200

LS1 8

NC

NC

R9 10

R11 120

C11 100n

C12 2200

HEADPHONES

D3 1N4001

D4 1N4001

SK3

R7 4k7

R8 1k5

C10 470p

TDA2030

ELECTRET MICROPHONE

IC2

VR2 10k LOG

C5 47

GAIN

C4 100p

VR1 470k

TL081

IC1

C3 47

VOLUME

NC = NO CONNECTION

R3 10k

PL2

R4 10k

TIP 2 5 4 1 3 NC 1 2 5 6 SK2 SK1 GUITAR MIC TIP TIP MIC1 k D1 BZY88 4V7 ZENER a MICROPHONE C1 47

R2 4k7

Everyday Practical Electronics, February 2002

R1 22k

Fig.1. Complete circuit diagram for the Guitar Practice Amp.

COM(0V)

C2 47

Most of todays top hits are songs and playing chords on their own does not sound very good, it is far better if the artist can sing along while playing. With an electric guitar a microphone is required to avoid having to shout rather than sing. Nowadays headphones which include a microphone are available from any computer store for around 5 and these are eminently suitable for this application. Many practice amplifiers however, have only one input and cannot easily accommodate a microphone but this deficiency has been rectified in this design by adding a simple mixer. The microphones incorporated in these cheap headsets are usually electret types. The microphone element constitutes in effect a very high impedance source and a buffer amplifier (consisting of a field effect transistor or f.e.t.) is normally incorporated within the microphone capsule as shown inset in Fig.1. This requires a small supply voltage (between 15V and 5V) and a load resistor to operate and so the components associated with the microphone input have been added to supply this. A nominal 5V supply is derived from the main supply rail via resistor R1 and Zener diode D1 while R2 forms the load resistor for the f.e.t. inside the microphone capsule. Note that a stereo jack socket (SK2) is used for the microphone with the second terminal supplying the +5V while the signal is picked up from the centre pin (tip) and the outer earth (0V) connection in the normal way. (The centre pin and the second terminal are connected inside the microphone). This allows a different microphone such as a dynamic type for example, which does not need a supply voltage or resistor, to be connected and in this case the 5V supply will simply be

R6 1k5

C6 47

R5 47k

C7 100n

C8 47

C9 100n

D2 RED

R10 1k

VE

0V

The circuit is completed by a conventional power supply consisting of mains transformer T1, bridge rectifier REC1 and smoothing capacitors C12, C13. It provides a d.c. supply of +12V and 12V and although a single rail supply could have been used, the advantage here is that the usual large speaker coupling capacitor is not required. This may not seem to be such an advantage when it is realised that two capacitors are now required in the power supply, but it does mean that the annoying switch-on thump normally associated with these amplifiers (due to the speaker coupling capacitor charging up) is eliminated. The relatively low impedances in the circuit mean that hum and noise pick-up is low so that an l.e.d. D4 Power On indicator has been included to remind the user to switch off!

REC1 2A

0V

9V

POWER SUPPLY

T1

This is arranged so that inserting the headphone jack plug automatically disconnects the speaker. It also switches in a resistor, R11, in series with the headphones to prevent overloading, see Fig.1 and Fig.3. Both the resistor and the headphone socket are mounted off the board and it will be noticed that the headphones which normally have an impedance of 32 ohms (each) are connected in series.

230V AC MAINS SUPPLY

FS1 100mA (S-B)

230V

ON/OFF

S1

0V

85

shorted to earth by the microphones mono jack plug causing no damage to either the microphone or amplifier.

ALL MIXED UP

The signal from the microphone is fed to the input of the amplifier via another input resistor R4, the value of which together with the feedback control (resistor) VR1 defines the gain of this channel. A 10 kilohms resistor was found suitable in the prototype but this may be changed if required, a higher value resulting in a lower gain and vice-versa. This stage (IC1) of the circuit forms an ideal signal mixer since the inverting input (pin 2) of the amplifier is a virtual earth so called because the op.amp IC1 maintains the voltage at its inverting input at zero volts. It does this by changing its output voltage when a change in the input voltage tries to upset this and as the feedback preset VR1 has a higher value than

of course, have to be chosen carefully to avoid over driving the amplifier. The output of a CD player for example would be much larger than that of a guitar so that its resistor would need to have a higher value. Alternatively, each channel could have a separate volume control fitted as shown. It would also be a good idea to fit d.c. blocking capacitors to prevent any d.c. on the output of the CD player or other device upsetting the bias conditions of the op.amp. No separate provision for controlling the volume of the microphone channel has been made in this version as the relative volume of the guitar can be controlled at the instrument itself while VR2 controls the overall volume.

COMPONENTS
Resistors
R1 R2, R7 R3, R4 R5 R6, R8 R9 R10 R11 All 025W 5% VR1 VR2 22k 4k7 (2 off) 10k (2 off) 47k 1k5 (2 off) page 10W 1k 120W carbon film or better

See

SHOP TALK

Potentiometers
470k carbon preset, lin. 10k rotary carbon, log.

CONSTRUCTION

This is a mains operated circuit and its construction should not be attempted by those who are not suitably experienced or supervised. The use of a printed circuit board (p.c.b.) makes the circuit

Capacitors
C1 to C3, C5, C6, C8 4m7 radial elect. 50V (6 off) C4 100p ceramic C7, C9, C11 100n ceramic (3 off) C10 470p ceramic C12, C13 2200m axial elect. 25V (2 off)

Semiconductors
D1 D2, D3

the i n p u t resistor R4, the output voltage change is higher resulting in a voltage gain. Another way to visualise this is to realise that an op.amp always tries to maintain both of its inputs at the same potential which in this case is 0V. This means that the microphone channel will not be affected by any changes in the volume or tone settings of the guitar which is also connected to this point via its own resistor R1.

D4 REC1 IC1 IC2

BZY88 4V7 Zener diode 1N4001 50V 1A rectifier diode (2 off) 5mm red l.e.d. 2A 100V in-line bridge rectifier (see text) TL081 j.f.e.t. op.amp TDA2030 audio amplifier

Completed p.c.b. showing the supply smoothing capacitors, on/off indicator l.e.d. and in-line rectifier. The mains transformer, fuseholder and on/off switch are mounted off-board.
very easy to build and, with only five connections to the board, it should be possible to assemble the Guitar Practice Amp without any major errors. The topside p.c.b. component layout, interwiring and full-size copper foil master pattern are shown in Fig.3. This board is available from the EPE PCB Service, code 336. Miscellaneous
Rf

VIRTUAL EARTH

A general circuit of a virtual earth mixer is shown in Fig.2. and there is nothing to stop you connecting another guitar or other signal source such as a tape or CD player in the same way simply by adding another input socket, connected to IC1s inverting input by its own resistor as shown. The values of the resistors would,

DC BLOCKING CAPACITOR

VIRTUAL EARTH GAIN = Rf/R1 R1

SK1 SK2, SK3 MIC1

INPUT 1

VR1

OUTPUT

GAIN = Rf/R2 R2
INPUT 2

S1 FS1 T1

635mm (in.) moulded mono jack socket, with 2 switched break contacts 35mm stereo jack socket, with 2 switched break contacts (2 off) sub-min. omni-directional electret microphone insert s.p.s.t. mains rated on/off toggle switch 100mA 20mm slow-blow fuse 18VA 230V a.c. mains transformer, 9V-0V-9V secondaries (see text)

VR2

VR3 etc.
FURTHER INPUTS AS REQUIRED

0V

Fig.2. Adding extra inputs to the virtual earth mixer circuit.

86

Printed circuit board available from the EPE PCB Service, code 336; 8-pin d.i.l. socket; panel mounted fuseholder; aluminium heatsink, size 38mm x 58mm approx.; control knob; multistrand connecting wire; mains cable; 8W speaker, type to choice; solder pins; solder etc.

Approx. Cost Guidance Only

excluding speaker & case

24

Everyday Practical Electronics, February 2002

Assembly of the board should begin by inserting the terminal pins which will be used to connect the speaker and transformer to the p.c.b. These usually require a certain amount of force to insert into the board which could damage adjacent components if this were done at a later stage. Once the solder pins have been fitted, the board may be completed by mounting resistors, diodes, capacitors etc. in ascending order of height. Care should, of course, be taken to ensure that diodes and electrolytic capacitors are inserted the correct way around. Note also that a wire link (made from a discarded component lead) and a resistor (R10) are mounted under C12 and C13 so that these components must obviously be fitted before the electrolytic capacitors are mounted on the board. A second wire link is also required between C6 and VR2. Although IC1 is not a CMOS device, and thus not particularly sensitive to static, it is worth fitting an i.c. socket to prevent any possibility of overheating it during the soldering operation this will also facilitate its easy removal should this be required.

POWER AMP
The audio power amplifier IC2 is more difficult to fit and before this is done it is best to prepare the small heatsink according to Fig.4. In the prototype this was made from a piece of L-shaped aluminium extrusion normally sold in DIY shops but should this not be easily obtainable a suitable
38mm

piece of sheet aluminium bent to shape and drilled as shown will do just as well. IC2 should be mounted on the board but its leads should not be soldered for the moment. Once this has been done, the heatsink can also be mounted on the board and secured to it using two nuts and bolts. When it is secure, IC2 should be bolted to it, via its metal tab, and it is here that the
TAB CONNECTED TO V

12mm

12mm

10mm 5mm
22mm

38mm
INPUT(+) (1)

V(3)

OUTPUT(4)

+V(5) INPUT( ) (2)

Fig.4

20mm

Fig.5

Fig.4 (top right). Heatsink dimensions and bending details. Fig.5 (top, far right). Pinout details for IC2, the TDA2030 audio amp.
SK3 3 LS1 5 2 R 11

Fig.3. Printed circuit board component layout, wiring and full-size copper foil master. The wiring to the microphone insert jack plug PL2 is shown inset below.

1 4 TIP

C1

R 4 R 3

VR1

R 1

C10 C8

+
LS1

+ +

+
R2 a D1 k C2 + TIP SK1 2 5 SK2 3 4 1 C3 C6 C4

R 5

R 9

R R R 6 7 8

C 11 C12 1 2 C13 R 10

REC1

IC2

C9 C7 3

IC1

C5

4 5 k VR2 a D2 D3 a k k HEATSINK

a D4

MIC GUITAR

S1

FS1 9V 0V 9V

MIC1 1 2 0V 0V (3) PL2

SOLDER TAG E L 230V N

T1

0V

+V
(2) SK2

TIP (1)

336

6.4IN (162.6mm)

Everyday Practical Electronics, February 2002

1.5IN (38.1mm)

87

advantage of delaying the soldering of this device will be seen as this will allow a certain amount of tolerance in the final positioning of the device relative to the heatsink. Once IC2 is secured to the heatsink, its leads can be soldered and trimmed in the normal way. Note that it may also be necessary to bend the leads slightly to enable it to fit the holes in the board, see Fig.5. Most devices are supplied with the leads already pre-formed although it should be noted that the TDA2030 is available with the leads formed for both vertical and horizontal mounting. Both types are identical but the vertical device is to be preferred as quite a lot of lead bending would be required to fit the horizontal device. A smear of silicone grease between the heatsink and IC2s tab will help to conduct heat away from the i.c. but this was not found necessary on the prototype. What is important however is to ensure that there is a good electrical path between the tab and the negative supply p.c.b. copper track. For this reason no mica washers or any other insulation should be fitted between the tab of IC2 and the heatsink. The heatsink is used as a negative supply connection to the chip and it must not be earthed or connected to any other part of the circuit. The pinout details of the TDA2030 are shown in Fig.5 for reference. The only other component worthy of individual mention is the bridge rectifier where a 2A device is specified. A 1A device could also be used but this was not available in the authors spares box. These are available in many variants and shapes and although any of these devices will do, the board has been designed for an in-line package and so this type should be purchased if possible to avoid a lot of lead bending.

quoted as an r.m.s. value when delivering its rated current. After rectification and smoothing the final d.c. output will be nearer the peak value (approximately 14 times the r.m.s. value) and as amplifier circuits of this type draw a relatively low current when no signal is present, the final supply

Completed amplifier circuit board showing the audio power output i.c. bolted to its heatsink. The loudspeaker/headphones are wired to two output solder pins hidden behind the volume control.

HEADPHONES TIP (LEFT) PL3 COM R L

Fig.6 (right). Headphone jack plug PL3 wiring. The headphone jack socket (SK3) contacts break when the plug is inserted, disconnecting the loudspeaker LS1.

COMMON

LEFT

RIGHT RING (RIGHT)

voltage could be even higher depending on the transformer used. The supply voltage should, therefore, be measured to ensure that it does not exceed the ratings of the i.c.s (i.e. plus and minus 18V). The centre tap of the secondary must be connected to the 0V rail (corner terminal of the p.c.b.) while the other two leads may be connected to the other two terminals either way around. The mains wiring should be carried out carefully and all joints well insulated to ensure that they cannot be touched inadvertently when the unit is in operation. A mains On/Off switch and a fuse should also be fitted in the live mains lead and the mains cable securely clamped to the box or cabinet using a suitable strain relief mounting bush. The speaker will also need to be connected to the output terminals using suitable lengths of wire. If a socket for headphones is to be included, this should be arranged to disconnect the speaker when the jack plug is inserted so that a switched socket will be required (see Fig.1 and Fig.6). The finished p.c.b. is quite light and so no special mounting hardware is required. It should be adequately supported by the potentiometer spindle and the input jack sockets but the final details of this are left to the constructor and will depend to a large extent on the cabinet in which the p.c.b. and speaker are mounted. When fully assembled, check the wiring again, especially around the headphone socket and transformer primary and if all is well, connect the unit to the mains and switch on. The voltage across each of the two smoothing capacitors can be measured and this should be about 12V d.c but no higher than 17V. A slight hum or hiss may be audible

FINAL ASSEMBLY

PRELIMINARY CHECKS

from the speaker if the Volume control VR2 is turned up fully. Turn down the volume and connect a guitar which should now be heard. The only adjustment to be made is to set the gain of the preamplifier stage (IC1) and this should be done with the volume turned up to maximum on VR2 and the guitar. Starting with preset VR1 turned fully clockwise the gain should be increased until distortion is heard when a string is played. An oscilloscope is useful here but not necessary as it is the final sound that is important and not the apparent purity of the output waveform. If required, the headphones can be plugged in and, provided the wiring has been done correctly, this should switch off the speaker. With this adjustment complete, the stage act can be perfected without interference from the rest of the household. Take it away Eric . . . 6

After careful checking of the board to ensure that there are no solder splashes between the tracks and that all the joints are sound, the speaker and mains transformer connections should be made to the board. The transformer used in the prototype had wire leads but if another type is used, then wires may need to be fitted. Printed circuit board mounting types should be avoided as these usually lack mounting brackets and in this case the transformer will need to be mounted on a chassis or in the wooden cabinet containing the speaker. The final arrangement will depend to a large extent on circumstances and is therefore left to the individual to solve. Care should be taken to ensure that a transformer with a centre tapped secondary (or with two secondary windings which can be connected in series) is used and although a voltage of 9V-0V-9V is specified, a slightly higher output could also be used. It should be remembered that the output of a transformer is always

FINAL TESTING

88

Everyday Practical Electronics, February 2002

VERSATILE MIC/AUDIO PREAMPLIFIER


by RAYMOND HAIGH
Use one of the latest chips on the block to produce an audio pre-amp with AGC compression, limiting, and noise reduction.
Intended primarily as a means of processing microphone inputs to computers, the SSM2166P integrated circuit (IC) manufactured by Analog Devices has a wider range of possible applications. Public address and surveillance systems immediately spring to mind, and the device will be of particular interest to radio enthusiasts, especially now that the popular Plessey 6270 IC mic/pre-amp, with voice gain, is no longer available. This article describes how the new IC can be used for a variety of signal inputs, and additional circuitry is given for readers who
BUFFER AMPLIFIER OUTPUT 5 VOLTAGE CONTROLLED AMP INPUTS 3 4

require a signal-strength meter.

THE CHIP
The various amplifying and control stages built into the SSM2166 chip are shown in Fig.1. Signal inputs are buffered by opamp A, internally connected to a rectifier stage, B, which produces a DC voltage which varies in proportion to signal strength. After processing by the control circuit, C, the DC voltage is used to fix the large and small signal gain of a second opamp, D.

AMPLIFIERS
The input impedance of buffer amplifier, A, is 180 kilohms (180k) and its gain can be set, by external feedback resistors, between 0dB and 20dB. There is a standing DC voltage on the input, and a blocking capacitor must be used. The input and output impedances of the controlled amplifier, D, are 1k, and 75 ohms, respectively. A standing DC voltage necessitates the use of a blocking capacitor at the output.
LIMITING IS IMPOSED IN THIS REGION IN ORDER TO HOLD THE OUTPUT BELOW A PRE-DETERMINED LEVEL

SIGNAL OUTPUT

+5V
14

BUFFER AMPLIFIER BUFFER AMP INPUTS (AUDIO IN) 6 7

CONTROLLED AMPLIFIER

PROCESSED 13 OUTPUT 2 SET GAIN

THRESHOLD OF LIMITING SET BY VR4

IN THIS REGION GAIN REDUCES AS SIGNAL STRENGTH INCREASES IN ORDER TO COMPRESS THE DYNAMIC RANGE. COMPRESSION IS SET BY VR3

SET A.G.C. TIME CONSTANT

B
TRUE R.M.S. LEVEL DETECTOR

C
CONTROL CIRCUITRY

POWER 12 DOWN (STAND-BY)

1 GROUND (0V)

9 SET SQUELCH THRESHOLD

10 SET COMPRESSION RATIO

11 SET LIMITING THRESHOLD

DOWNWARD EXPANSION OR SQUELCH THRESHOLD SET BY VR2

IN THIS REGION GAIN REDUCES AS SIGNAL STRENGTH REDUCES IN ORDER TO PREVENT HIGH-LEVEL AMPLIFICATION OF NOISE UNDER NO-SIGNAL CONDITIONS

SIGNAL INPUT

Fig.1. Internal block schematic for the SSM2166P microphone preamplifier, with variable compression and noise gating.
Copyright 2000 Wimborne Publishing Ltd and Maxfield & Montrose Interactive Inc

Fig.2. Relationship between limiting, compression, and downward expansion or squelch.


EPE Online, May 2000 - www.epemag.com - 338

Constructional Project
Provision is made for setting the nominal gain of the controlled stage between 0dB and 20dB, but AGC action will increase amplification, at the lowest signal levels, to as much as 60dB. The output can be muted. Interestingly, the noise generated by the controlled stage is designed to be at a minimum when its gain is at a maximum, and this significantly improves the overall signal-tonoise ratio of the system.

RECTIFIER
The circuit of the rectifier, or level-detector stage (B), has been specially developed for this application. It produces a DC control voltage, which is proportional to the log of the true RMS value of the input signal. The speed at which the control voltage responds to changes in signal level, or the attack time, can be controlled by the user. Response to highlevel changes is automatically

speeded up by the IC in order to minimize the duration of any overload.

CONTROL CIRCUIT
The control circuit (C) enables the user to program the performance of the IC in a very comprehensive way, and the amount of signal compression can be set between zero and 60dB. Signal limiting can also be applied to prevent the occasional transient exceeding

the desired maximum output. It can be set at outputs ranging from 30mV to 1V. Above this threshold, the maximum compression ratio of 15:1 is applied. The response of the system to very low level inputs can be reduced in order to prevent the amplification of noise under nosignal conditions. The threshold of this downward expansion (the lower the signal the less it is amplified), can be set at inputs of between 250uV and 20mV.
+5V
OUT

IC2
COM

IN

+8V TO +18V

LM78L05 WIRE LINK SQUELCH LK1 VR2 1M LIMIT VR4 47k C10 470

R1 1k

C5 47 R4 1k R5 1k 9 11 14 12 C8 100n

AUDIO INPUT 1

R2 10k 6

R9 15k VR7 10k R7 2M2 VR8 10k

C1 47

C2 100n

IC1
7

SSM2166P 13

+
ME1 50A TO 1mA SIGNAL STRENGTH METER (SEE TEXT)

TR1
b

AUDIO INPUT 2 VR1 4k7 R3 10k

8 COMPRESSION

10

1 R6 1k

2 C9 1 OUTPUT SIGNAL LEVEL VR6 10k AUDIO OUTPUT

BC547

INPUT SIGNAL LEVEL C3 10n

VR3 100k C4 1 C6 47 C7 22

R8 SEE TEXT (TABLE 2)

R10 4k7

SCREEN 0V

GAIN VR5 22k

SCREEN 0V AUDIO IN (1): ELECTRET MICROPHONES AND INPUTS REQUIRING A D.C. BLOCKING CAPACITOR AUDIO IN (2) DYNAMIC (MOVING COIL) MICROPHONES

Fig.3. Complete circuit diagram for the Versatile Mic/Audio Preamplifier.


Copyright 2000 Wimborne Publishing Ltd and Maxfield & Montrose Interactive Inc

EPE Online, May 2000 - www.epemag.com - 339

Constructional Project
Provision is made for the device to be placed in a powerdown or stand-by mode, and this feature will be of particular interest when it is used in sophisticated surveillance systems. In this state, current consumption is reduced to around 10mA and the input and output ports assume a high impedance. User programmable control circuitry, coupled with the complex rectifier or leveldetector, contributes significantly to the chips performance. The relationship between the noise reduction, compression and limiting functions is displayed in Fig.2. Static discharges can damage the IC, and the usual precautions (discharging the body) should be taken when handling and connecting it into circuit. The SSM2166P is embedded in a 14-pin, dual-in-line package, and the suffix P refers to the standard-size version. This is the type most likely to be stocked by suppliers. However, surfacemount types are also manufactured: these carry the suffix S. circuit. Its value is appropriate for moving coil and electret microphones, and for audio signals derived from most transistor circuits. Keeping the

COMPONENTS
Resistors
R1, R4, R5, R6 1k (4 off) R2, R3 10k (2 off) R7 2M2 R8 (see Table 2) R9 15k R10 4k7 All 0.25W 5% carbon film

Potentiometers

CIRCUIT DETAILS
The full circuit diagram for the Versatile Mic/Audio Preamplifier, incorporating a signal strength meter, is given in Fig.3. Provision for controlling so many functions results in a plethora of preset potentiometer controls. However, they do enable the signal processing to be tailored to individual requirements, and their adjustment is not critical or difficult. A summary of their various functions is set out in Table 1. Preset VR1 permits adjustment of the input signal level to prevent overload and to optimize the performance of the

RATINGS
No doubt with computer circuit compatibility in mind, the SSM2166 is designed for a 5V supply. The absolute maximum supply voltage is 10V. Current consumption is approximately 10mA. The maximum input to the buffer is 1V, and the maximum output from the controlled amplifier is 14V RMS for 1 per cent total harmonic distortion. Frequency response extends well into the RF spectrum.

VR1 4k7 enclosed carbon preset, horizontal VR2 1M enclosed carbon preset, horizontal VR3 100k enclosed carbon preset, horizontal VR4 47k enclosed carbon preset, horizontal VR5 22k enclosed carbon preset, horizontal VR6, VR7, VR8 10k enclosed carbon preset, horizontal (3 off) C1, C5, C6 4u7 radial electrolytic, 10V (3 off) C2, C8 100n ceramic (2 off) C3 10n ceramic C4, C9 1u radial electrolytic 10V (2 off) C7 22u radial electrolytic C10 470u radial electrolytic TR1 BC547 (or similar, e.g. BC239, BC548) npn low-power transistor IC1 SSM2166 microphone preamp (Analog Devices) IC2 LM78L05ACZ +5V 100mA voltage regulator ME1 50uA to 1mA FSD moving coil meter (see text)

Capacitors

Semiconductors

Table 1: Preset Control Functions


Preset
VR1 VR2 VR3 VR4 VR5 VR6 VR7 VR8

Value
4k7 1M 100k 47k 22k 10k 10k 10k

Function
Set input signal level: clockwise to increase. Set threshold of downward expansion (squelch): clockwise to lower. Set compression: clockwise to increase. Set threshold of signal limiting: clockwise to lower. Set gain of controlled amplifier: clockwise to increase. Set output signal level: clockwise to increase. Set signal strength meter pointer at full scale (when strongest signal being processed): clockwise gives clockwise pointer movement. Set signal strength meter pointer at zero (under no-signal conditions): clockwise gives clockwise pointer movement.

Miscellaneous

Printed circuit board available from the EPE Online Store, code 7000260 (www.epemag.com); 14-pin DIL socket; screened cable, solder pins, solder, multistrand connecting wire, etc.

See also the SHOP TALK Page!

Approx. Cost Guidance Only (Excluding meter)

$27

Copyright 2000 Wimborne Publishing Ltd and Maxfield & Montrose Interactive Inc

EPE Online, May 2000 - www.epemag.com - 340

Constructional Project
value below 5 kilohms increases the stability margin of IC1. Power can be supplied to an electret microphones integral FET (field-effect-transistor) buffer via resistor R1, and C1 acts as a DC blocking capacitor. The input arrangements for alternative microphones and other signal sources are discussed at greater length later. The input signal to IC1 is applied to the non-inverting (+) input of the buffer amplifier stage (Pin 7 see Fig.1) via blocking capacitor C2. This IC has an extended frequency response and C3 introduces a measure of roll-off above 20kHz or so, again in the interests of stability. If desired, the gain of the buffer can be set at unity by deleting R3 and inserting a wire link in place of resistor R2 (to connect pin 5 and pin 6). Blocking capacitor C4 maintains the correct DC conditions. amplifier.

ATTACK TIME
The response or attack time of the AGC system can be controlled by adjusting the value of the rectifier reservoir capacitor C7. The IC manufacturer suggests a value within the range 22uF to 47uF, with smaller capacitors being suitable for music and the larger for speech. Too low a value will result in pumping effects, with background noise rushing up between bursts of speech. This will become increasingly apparent as the compression ratio is raised. Conversely, too high a value will excessively slow the response of the system to changes in signal level. The 22uF component specified for C7 has been found to work well with both speech and music inputs. The attack time is controlled mainly by the value of C7, but the much longer decay time is dependant upon this capacitor and the internal control circuit. Fast attack and slow decay help to reduce the pumping effect, which seems far less pronounced with this IC than with simpler audio AGC systems.

CONTROLLED AMPLIFIER
The output from the buffer stage (pin 5) is connected, via DC blocking capacitor C5 to the non-inverting (+) input (pin 3) of the controlled amplifier stage. A capacitor of identical value, C6, at pin 4 connects the inverting () input to ground (0V). (This connection makes any electrical noise on the ground rail appear as a common mode signal to the controlled amplifier and the differential input circuitry rejects it). The nominal gain of the controlled amplifier can be set, by preset VR5, between unity and 20dB. Resistor R6 ensures that the gain does not fall below unity. Switched muting can be achieved by grounding pin 2 via a 330 ohm resistor (the switch should be located at the ground or 0V rail end). Switch clicks can be suppressed by connecting a 10nF capacitor between pin 2 and ground. The IC can be put in standby mode by disconnecting pin 12 from ground and connecting it, via a 100 kilohms resistor, to the +5V rail. (Provision has not been made for muting or powering-down on the PCB.) The processed output is taken from pin 13 and connected, via DC blocking capacitor C9, to preset VR6. This enables the output signal level to be adjusted to suit the input sensitivity of the power

BUFFER GAIN
The gain of the buffer amplifier is set at 6dB by resistors R2 and R3, and this is likely to be sufficient for most purposes. Gain can be increased to a maximum of 20dB by decreasing R3 to about 12 kilohms. Adding this to the gain of the controlled amplifier results in an overall system gain, when signals are too small to initiate compression, of 80dB. This is a great deal of amplification in a small package, and particular care must be taken with the screening and routing of input and output leads, and the connections to a shared power supply, if instability is to be avoided. Separate ground, or 0V leads, from signal source circuitry, the preamplifier, and the power amplifier, should be run to a common point at the power supply. The screening braid of signal cables should be connected to ground at the preamplifier end only.
Copyright 2000 Wimborne Publishing Ltd and Maxfield & Montrose Interactive Inc

COMPRESSION
The amount of compression is determined by preset VR3, which connects pin 10 to ground. There is no compression with the potentiometer set to zero. When its resistance is at maximum, a 60dB change in input level (above the downward expansion or squelch threshold) changes

EPE Online, May 2000 - www.epemag.com - 341

Constructional Project
the output by less than 6dB. The onset of limiting is controlled by preset VR4. Setting this potentiometer to maximum resistance fixes it at 30mV. With VR4 at minimum resistance, it is around 1V RMS. Above the threshold of limiting, a 15:1 compression ratio is imposed, irrespective of the setting of compression control VR3. noted that, under a light load, a fresh 9V alkaline battery will usually deliver a higher voltage than this. However, in order to ensure the correct operation of the device, and provide a high degree of isolation from other equipment sharing the same power supply, a 5V 100mA voltage regulator, IC2, is included in the circuit. This enables supplies with outputs ranging from 8V to 18V (or more, depending on IC2 rating) to be used. Bypass capacitors C8 and C10 shunt the noise in the regulator output to ground. Note that C8 is essential to the stability of IC1 and it must be located as close as possible to pin 14, even when the unit is battery powered. Transistor TR1, configured as a DC amplifier, ensures that IC1s AGC line is only lightly loaded, even when a 1mA meter is used. It forms one arm of a bridge circuit, the other three being its collector load, R9, and the potential divider chain comprising preset VR8 and resistor R10. The bridge is balanced, and the meter set at zero under no-signal conditions, by preset potentiometer VR8. When a signal is being processed, the rising AGC voltage on the base (b) of TR1 increases its collector current and, hence, the voltage drop across resistor R9. This unbalances the bridge and drives the meter pointer over. Preset VR7 adjusts the sensitivity of the meter so that the pointer can be set just short of full-scale deflection (FSD) when registering a strong signal. The circuit can be made to accommodate meters with fullscale deflections ranging from 50mA to 1mA by adjusting the value of resistor R8. This resistor controls the flow of current through the base-emitter junction of transistor TR1, and values to suit a range of meter FSDs are given in Table 2. Bias resistor R7 provides a measure of negative feedback which helps to stabilize the operation of the circuit. Almost any small-signal npn transistor should prove suitable for TR1, and a 2N5827 or 2N5828 could be used in addition to the types listed in the Components list. These devices have different case styles and the base connections must be checked.

NOISE REDUCTION
Preset potentiometer VR2 sets the threshold below which downward expansion (gain reduces as the signals become weaker) is applied. With maximum resistance, downward expansion starts at signal levels in the region of 250mV. Turned to zero resistance, the threshold is raised to around 20mV. Gain rises to a maximum under no-signal conditions with all conventional AGC systems, and the amplification of external and internally generated noise produces a loud and tiresome hiss in the speaker or phones. The ICs noise reduction facility, which operates as a squelch control, is very effective in overcoming this. It can reduce output noise below the level of audibility when signal levels fall to zero. With any squelch system, a need to resolve very weak signals overlaid by noise compromises the usefulness of the feature. Radio enthusiasts with a particular interest in it could mount VR2 as a panel control so that the threshold could be adjusted to suit reception conditions.

SIGNAL STRENGTH METER


Some readers, especially those wishing to incorporate the unit into a radio receiver, may welcome the provision of a signal strength meter. This is included in the circuit diagram of Fig.1 and consists of transistor TR1, meter ME1 and associated components. The AGC control voltage appears on pin 8 of IC1. It ranges from 290mV under nosignal conditions to approximately 720mV with high level inputs.

Table 2: Signal Strength Meter ( Values of R8 for different meter sensitivities)


Meter FSD 50uA 100uA 500uA 1mA R8 1M 470k 100k 47k

POWER SUPPLY
The maximum safe supply voltage is 10V, and it should be
Copyright 2000 Wimborne Publishing Ltd and Maxfield & Montrose Interactive Inc

CONSTRUCTION
All the components, with the exception of the meter ME1, are

EPE Online, May 2000 - www.epemag.com - 342

Constructional Project
assembled on a small, singlesided, printed circuit board (PCB). The topside component layout, together with an (approximately) full-size underside copper foil master pattern, is shown in Fig.4. This board is available from the EPE Online Store (code 7000260) at www.epemag.com Commence construction in the usual way by mounting the smallest components first working up to the largest, but fit IC1, IC2, and TR1 last (see earlier comments about the static sensitive nature of IC1). A holder for IC1 will facilitate substitution checking. Solder pins, inserted at the lead-out points, will ease the task of offboard wiring.

MICROPHONES
The unit works well with
POWER SUPPLY NEGATIVE

dynamic (moving coil), electret, crystal, and ceramic microphones. Screened cable
+8V TO +18V

UNDERSIDE VIEW VR5 ebc BC547 BC239 BC548 UNDERSIDE VIEW IN COM OUT LM78L05ACZ VR6

IC2

GAIN R 6

IN COM OUT VR4 C9

OUTPUT SIGNAL LEVEL THRESHOLD OF LIMITING SIGNAL OUT R 5 VR3 COMPRESSION R8 R 4 VR8 VR2 SQUELCH DOWNWARD EXPANSION

C10 C4

R3

C5 R2 AUDIO INPUT 1 INPUT GROUND (0V) AUDIO INPUT 2 C1 VR1 R1 LK1 R 9 R7 e b c VR7

C2 C 3

C6 R10

SPOT-CHECKS
When all the components have been soldered in position on the PCB, double-check the orientation of electrolytic capacitors, the ICs, and the transistor. Also, check the PCB for bridged tracks and poor solder joints. Next, with IC1 out of circuit, connect a supply voltage of between 7V and 9V and check that the output from regulator IC2 is producing 5V. A fault in this device, or its wrong connection, could result in the destruction of IC1 when higher voltages are applied. Once all is well, place IC1 in its socket (checking orientation), connect, via screened cable, a signal source and a power amplifier. Adjust the various preset potentiometers until the processing meets your requirements. All preset functions are summarized in Table 1 for ease of reference.

TR1
SET METER AT FULL SCALE

Fig.4. Printed circuit board component layout, inter-wiring details and (approximately) full-size underside copper foil master.
EPE Online, May 2000 - www.epemag.com - 343

Copyright 2000 Wimborne Publishing Ltd and Maxfield & Montrose Interactive Inc

INPUT SIGNAL LEVEL

C7

+
SET METER AT ZERO

C8

+ +

SIGNAL STRENGTH METER

Constructional Project
R1, C1 AND VR1 ARE LOCATED ON THE PREAMPLIFIER P .C.B. R1 MICROPHONE CASE SCREENED CABLE HIGH IMPEDANCE MICROPHONE 2N3819 g s VR1 10 1k 0V d C1

+5V

2M2

Fig.5. The line-powered buffer stage built into electret microphones can also be used for ceramic and crystal types. (Most FETs will function in this circuit with the source grounded, eliminating the need for the source resistor and bypass capacitor.)
must, of course, be used to connect any type of microphone to the preamplifier. Very high quality studio microphones can be insensitive and require balanced feeders to minimize hum pick-up. The preamplifier described here is configured for unbalanced inputs, and is not likely to be suitable, as it stands, for microphones of this kind. A few words about the various types of signal input may prove helpful. Dynamic Microphones are manufactured with impedances ranging from 50 ohms to 600 ohms. Output tends to be greatest with the higher impedance units. This type of microphone should be connected to Input 2 (i.e., directly across preset VR1), and the wire link must be removed to isolate resistor R1 from the 5V rail. Electret Microphones are a modern development of the capacitor microphone (a permanently charged diaphragm, the electret,
Copyright 2000 Wimborne Publishing Ltd and Maxfield & Montrose Interactive Inc

Layout of components on the completed circuit board. The Signal Strength Meter components, except the meter, have been included on the board (bottom right).
eliminates the need for an external charging voltage). The output from the actual unit is low and at a high impedance, so these microphones have an integral FET buffer. The drain load for the internal FET is provided at the amplifier end of the cable (resistor R1 in Fig.3), to facilitate line powering. Electret microphones must be connected to Input 1, and the wire link must be in place to connect resistor R1 to the supply rail. The 1 kilohm drain load (R1), fed from the 5V supply, should ensure the optimum performance of most microphones of this kind. Crystal and Ceramic Microphones rely upon the piezo-electric effect to produce a signal voltage. The vibrating diaphragm induces stresses in a wafer of crystal, often Rochelle salt, or in a barium titanate element in the case of ceramic units. These microphones should be connected to Input 2. They have a high impedance, and feeding them into preset VR1 will reduce their response to low audio frequencies. Low frequency roll-off is, however, desirable for communications work, and more is said about this later. The use of long connecting cables will attenuate the signal but have little effect on frequency response (cable capacitance is modest compared to the selfcapacitance of these microphones, which can be as high as 30nF). If an extended frequency response is required from microphones of this type, the use of an external, linepowered, FET buffer, as built into electret microphones, is

EPE Online, May 2000 - www.epemag.com - 344

Constructional Project
Direct conversion and regenerative receivers will require a single transistor audio amplifier, after the product detector or regenerative detector, in order to ensure sufficient signal voltage for the SSM2166P. The output from the detector stage in most superhets will be more than adequate. Radio receivers should be connected to Input 1, and the wire link removed. The orientation of electrolytic capacitor C1 will usually be correct when the receiver has a negative ground or 0V rail. However, some diode detectors in superhets are configured to provide an output which is negative going with respect to ground (to suit the receivers AGC circuit). The polarity of C1 will need to be reversed when equipment of this kind is connected.

SIGNAL PROCESSING
The preamplifiers frequency response is reasonably flat from below 100Hz to more than 20kHz. Speech clarity, especially under noisy conditions, can be improved by rolling off frequencies below 300Hz and above 3000Hz, and active or passive band-pass filters are often used for this purpose. A big improvement can, however, be made by modifying some of the coupling and bypass components in the preamplifier. Constructors wishing to limit the frequency response in this way should reduce the value of capacitor C9, to 47nF (a Mylar or ceramic capacitor is then suitable). This will attenuate the lower frequencies. Wiring a 220nF ceramic capacitor across preset VR1 will attenuate the higher frequencies. Although extremely simple, these measures are quite effective.

recommended. This will also prevent signal losses when long cables are used. A suitable circuit diagram is given in Fig.5 and the circuit can be built inside the microphone case. When this arrangement is adopted, resistor R1 must, of course, be connected to the supply rail, and the signal must be fed to Input 1.

RADIO RECEIVERS
Audio derived AGC is often incorporated into direct conversion radio receivers. Even simple superhets can benefit from this form of control (sometimes a conventional RF derived system is not very effective when amateur singleside-band signals are being processed).

Copyright 2000 Wimborne Publishing Ltd and Maxfield & Montrose Interactive Inc

EPE Online, May 2000 - www.epemag.com - 345

Constructional Project

HANDY-AMP
TERRY DE VAUX-BALBIRNIE
A useful multi-purpose amplifier.

NTIL recently, the authors household hi-fi system had a piece of screened cable hanging down the back. This was left connected to the amplifiers high-level (auxiliary) input. When some piece of experimental audio equipment needed to be tested, the cable could be retrieved and connected to the circuit. It was then possible to listen to the result.

Readers should note that the amplifier has been designed to be small and relatively inexpensive to construct. Although the maximum output power (one watt approximately) and sound quality are perfectly adequate for the applications suggested earlier, it is not suitable for critical applications such as serious music listening. The stated power rating of 1W is applicable when the amplifier is connected to an 8-ohm load. If a 4-ohm loudspeaker were to be used, the rating would be some 15W. In fact, the subjective difference between the two is not great and battery life is reduced at the higher power. It is therefore recommended that an 8-ohm loudspeaker is used. A 4-ohm unit had to be used in the prototype for availability reasons. The low-level input has its own gain control while the overall gain is set using the master volume control. This allows just about any input source to be connected, including microphones, musical instruments and consumer audio equipment. Although not ideal, the headphone socket fitted to many pieces of electronic equipment provides a signal which will drive the high-level amplifier input. When doing this, the volume control on the equipment would need to be adjusted to obtain the correct input level. When this was tried with a small TV, the sound was better than from the TV itself. Most of the time, the amplifier will probably be used in conjunction with the

internal loudspeaker. However, better sound quality is obtained when using either headphones or a good-quality external loudspeaker. Although the amplifier is monophonic (that is, not stereo), when used with headphones, the output is applied equally to each one. This gives a more comfortable effect than with only one headphone operating.

ON THE PANEL

NOT GOOD

WATTS ENOUGH?

This method was far from satisfactory, so a small battery-operated bench amplifier was designed for such purposes. As well as having an in-built loudspeaker, it has the facility for connecting personal stereo type headphones or an external loudspeaker. Also, it will accept both highlevel and low-level input devices. Magnetic record player cartridges and dynamic microphones provide a low-level output while the line output socket fitted to many pieces of consumer equipment (such as CD players and video recorders) provide a high level. Many readers will, no doubt, wish to construct the amplifier for experimental purposes. However, it could have a variety of other applications. Examples include a small practice amplifier for electronic musical instruments and as the basis for an intercom, or toys and games. Being battery-operated it may be set up outdoors and, with just a microphone (possibly with an extension lead) and a pair of headphones connected, it could be used to listen to wildlife.

The completed Handy-Amp is shown in the photograph. For convenience, the rotary controls and all sockets and switches are mounted on the front panel. These are a jack and phono-type socket for the Low-level and High-level inputs respectively, together with the input selector switch, low-level Gain and master Volume controls, light emitting diode (l.e.d.) indicator and on-off switch, headphone jack socket, external loudspeaker sockets and output selector switch. On top, there is a matrix of holes to allow the sound to pass out from the internal loudspeaker. There are several possible battery arrangements and the one chosen will be determined largely by the space available inside the case. This, in turn, will depend to a great extent on the dimensions of the internal loudspeaker. Whatever battery is used, it must have a nominal voltage of 9V (say, six 15V cells connected in series). Cells should not have a capacity less than alkaline AA size. Note that a PP3 type battery would be totally unsuitable. The prototype unit was powered using two 45V alkaline 3LR12 batteries taped together and connected in series. These have around twice the capacity of alkaline AA cells. The standby current requirement of the circuit depends on the load. In the prototype, it is 100mA. However, there will be peaks of several hundred milliamps and, depending on how the amplifier is used (operating time, load and volume), a life of some 15 hours may be expected from a pack of AA alkaline cells. This would be sufficient for occasional use. With headphones connected, the standby current requirement of the prototype unit was only 50mA giving a longer battery life. The l.e.d. indicator reminds the user to switch off the unit after use. Although battery operation is convenient and safe, for long periods of operation the use of a plug-in power supply unit might be more appropriate. More will be said about this later.

572

Everyday Practical Electronics, August 2000

CIRCUIT DESCRIPTION

The full circuit diagram for the Handy-Amp is shown in Fig.1. The design uses two main integrated circuits (i.c.s), IC1 and IC2, together with voltage regulator IC3. Battery B1 provides a nominal 9V supply to the regulator which then gives a 5V supply for the main circuit. This will be maintained until the battery voltage falls to some 7V, whereupon the regulated output will fail. Thus, as the battery ages, the supply will remain constant throughout its useful life. Note that the l.e.d. on-off indicator, D1, is connected in series with current-limiting resistor R7 directly across the battery supply that is, it is not subject to the effect of the regulator. It will be obvious when the batteries need to be replaced because the amplifier output will become weak and distorted and the l.e.d. will become dimmer. Capacitors C9 and C10 promote stability of the regulator. Capacitor C11 charges up from the battery and can then maintain the supply on the output current peaks when the amplifier is delivering maximum power. This helps to provide a distortion-free output. If using a plug-in power supply unit, C11 will provide additional smoothing if a poorly-smoothed supply is used. This should not be necessary with a good-quality unit but will be useful with inexpensive ones.

FIRST BOOSTER

When a low-level device such a microphone is connected, via socket SK1, its output voltage is first boosted using a low-noise pre-amplifier, based on operational amplifier (op.amp), IC1. Highlevel (line) signals are input via socket SK2, thus bypassing IC1.

COMPONENTS
Resistors
R1 R2, R3 R4 to R6 R7 All 025W 5% VR1 VR2 VR3 680W 47k (2 off) 22k (3 off) 560W carbon film

See

SHOP TALK
page

Potentiometers
470k min. panel mounting, lin or log rotary carbon 10k min. panel mounting, log rotary carbon 47k min. preset, vertical carbon

Capacitors
C1, C3, C7 10m radial elect. 63V (3 off) C2 22m radial elect. 63V C4 4m7 radial elect. 63V C5 2m2 radial elect. 63V C6, C8 100n polyester, 5mm pin spacing (2 off) C9, C10 220n polyester, 5mm pin spacing (2 off) C11 1000m radial elect. 16V

Semiconductors
D1 IC1 IC2 IC3 SK1 SK2 SK3 SK4, SK5 S1, S2 S3 LS1 red l.e.d., 3mm NE5534AN op.amp SSM2211 power amplifier 7805 5V 1A voltage regulator 6.35mm plastic body mono jack socket phono jack socket, single-hole fixing (see text) 635mm stereo jack socket, plastic body 2mm socket or as required (see text) (2 off) s.p.d.t. toggle switch (2 off) s.p.s.t. toggle switch small 8-ohm loudspeaker, 2W rating minimum (see text)

Miscellaneous

Printed circuit board, available from the EPE Online store code 273; aluminium case, 203mm x 127mm x 51mm; 8-pin d.i.l. socket (2 off); 3mm l.e.d. panel clip; control knob (2 off); alkaline AA-size cells (6 off see text); holder and connector for cells (or as required).

Approx. Cost Guidance Only Fig.1. Complete circuit diagram for the Handy-Amp.

excluding batts. and case

$29.60

29.6

Everyday Practical Electronics, August 2000

573

The signal source is selected by switch S1 and, via volume control VR2, passed on to the power amplifier section centred on IC2. Op.amp IC1 is configured as a voltage amplifier used in inverting mode. Pins 7 and 4 are the positive and 0V supply inputs respectively. Blocking capacitor C1 allows the alternating current signal from a source connected to socket SK1 to pass via resistor R1 to the inverting input, at pin 2. The input impedance is set by the value of R1 and this will provide a good match for dynamic microphones. The op-amp non-inverting input, pin 3, receives a d.c. voltage equal to one-half that of the supply (nominally 25V) due to equal-value resistors, R2 and R3, which form a potential divider connected across the supply.

NO PAIN, NO GAIN

The pre-amp gain is set by the ratio of feedback resistance (R4 plus VR1) to input resistance, R1. With VR1 set to minimum, this provides a gain of about 32 and at maximum, rather more than 700 (the fact that this is an inverting amplifier and the gain has a negative sign is of no real consequence here and may be disregarded). This range of gain will suit microphones and other low-level input devices. VR1 is the low-level gain control (labelled simply Gain on the front panel). In use, this will be adjusted to take account of the sensitivity of individual input devices.

Fig.2. Block diagram of the SSM2211 power amplifier.


d.c. voltage equal to one-half that of the supply. This is due to the potential divider consisting of two internal 50kW resistors connected between supply positive (pin 6) and 0V (pin 7). Pins 2 and 3 are then connected to one end of the external bypass capacitor, C6, with the other end connected to the 0V line. This may be compared with the biassing arrangement used for IC1. Op.amp B is also configured as an inverting amplifier and because the internal input and feedback resistors have equal values (50kW), the gain is set at minus one. Thus, any signal appearing at Out 2 is an inverted copy of that at Out 1. In this way, the input signal at pin 4 has an amplified but inverted copy of itself at pin

5 and a straight copy of itself at pin 8 amplified by the same amount. This is known as a bridge output configuration. Correct working depends on the two op.amp sections being exactly balanced but, of course, this is not easy to achieve precisely. In theory, when no input signal is present, Out 1 and Out 2 will be at the same voltage. No current will then flow in a loudspeaker connected between them. When a signal is present, either Out 1 will drive current through the loudspeaker winding, which then sinks into Out 2, or Out 2 will drive current through the loudspeaker in the opposite direction and sink into Out 1. This will then reproduce the positive and negative excursions of the a.c. waveform presented to the input. In practice, there will be a small voltage difference between the outputs in the absence of an input signal. A small standing current will then flow through the loudspeaker coil and the lower its impedance, the greater this current will be. This is added to the small current needed by the i.c. itself (for the working of op.amps A and B, and for the current drain through the internal potential divider). The overall current requirement is therefore somewhat load dependent.

SHUTDOWN

MAKING THE SWITCH

The output from IC1 appears at pin 6. This passes via blocking capacitor C4 to the Low level (Low) contact of two-way Select Source switch, S1. With this set as shown, any low-level signal passes to the common contact and hence through the track of potentiometer VR2 to the 0V line. The sliding contact (wiper) of VR2 draws off the required fraction of the signal voltage and passes it, via capacitor C5 and resistor R5, to the input of the power amplifier (IC2 pin 4). If switch S1 is set to the alternative position (High), the output from IC1 is disconnected from VR2 but now any signal applied to the high-level input socket, SK2, is directed through VR2 instead. VR2 is the master volume control (labelled Vol on the front panel).

The SSM2211 amplifier used as IC2 has a shutdown feature. Thus, if pin 1 is made high, the i.c. is put into sleepmode and requires very little current. However, this feature is not used here and is disabled by connecting pin 1 to the 0V line along with pin 7. The gain of IC2 is calculated by the ratio of feedback resistance (VR3 plus

INTERNAL STRUCTURE

Power amplifier IC2 is an interesting device and a block diagram showing its simplified details is given in Fig.2. Basically, it consists of two operational amplifiers, A and B. The output of op.amp A (pin 5) provides one of the outputs (Out 1). However, it also feeds the inverting input of op.amp B (via the 50kW input resistor) whose output (pin 8) becomes Out 2. The loudspeaker is connected directly between Out 1 and Out 2. Op.amp A is configured as an inverting amplifier. Thus, the signal appearing at Out 1 is an amplified and inverted version of that at the input, pin 4. Referring back to Fig.1, its gain is set by the value of external fixed resistor R6 and preset VR3 connected between Out 1 and input pin 4. The non-inverting input (pin 3) is connected externally to pin 2 which sets it at a

Fig.3. Handy-Amp component layout and full size copper foil track master pattern.

574

Everyday Practical Electronics, August 2000

R6) to input resistance (R5) multiplied by two. This multiplied by two aspect comes about because of the bridged output configuration giving twice the voltage swing to the loudspeaker compared with a single op.amp. With VR3 set to minimum resistance the gain is therefore two, and at maximum resistance is just over six. VR3 is a preset potentiometer which allows adjustment for the desired gain.

OUTPUT ARRANGEMENTS

With switch S2 (Select Output) in the position shown in Fig.1, the internal loudspeaker is connected between Out 1 and Out 2. With the switch in the alternative position, the output is directed to both the Phones socket, SK3, and the external loudspeaker sockets (SK4 and SK5). It is thought unlikely that anyone would wish to connect an external speaker and a pair of headphones to the amplifier at the same time. However, even if they did, the load would not fall below the minimum impedance providing an 8-ohm loudspeaker was used. When using headphones, a greatly reduced power is available to them compared with a loudspeaker. This is because the (usually) higher impedance allows less current to flow. The impedance of typical personal stereo type headphones is about 30 ohms for each unit. In this design, the left and right units are connected in parallel giving a combined impedance of some 15 ohms. However, because headphones provide acoustic energy direct to the ears, only a very small amount of power is needed for them to sound with acceptable loudness.

to one another and to various points on the p.c.b. later. Begin by drilling the fixing hole in the p.c.b. then solder the sockets for IC1 and IC2 in position (but do not insert the i.c.s themselves at this stage). Follow with all fixed resistors and capacitors. Note that the resistors are mounted vertically. There are seven electrolytic capacitors and it is important to solder all of these with the correct polarity. The negative () end is clearly marked on the body and the corresponding lead is slightly shorter than the positive (+) one. Solder preset VR3 in place but not panel potentiometers VR1 and VR2 yet. Fit the control knobs to VR1 and VR2. Measure how much of each spindle needs to be cut off then remove the knobs again. Hold the end of the spindle (not the potentiometer body or it could be damaged) in a vice and cut off the required length using a small hacksaw. Smooth the cut edges using a file and check that the knobs fit correctly. Cut off the panel-location tags fitted to most potentiometers. If these are left in place, the bodies will not seat flat against the front panel when the p.c.b. is in position. The potentiometers should now be soldered to the p.c.b. Identify the l.e.d. end leads. The cathode (k) is usually shorter than the anode (a) lead. Also the body has a small flat to denote the cathode end. Solder the leads to the D1 pads on the p.c.b. observing the correct polarity. Bend them through rightangles, as shown in the photograph, so that the body ends up in line with the centre of the potentiometer spindles and standing out to about the centre of the bushes. Solder pieces of light-duty stranded connecting wire to the following points on the p.c.b.: Low-Level Input, S1 (L and C), Out 1 and Out 2. Using different colours of wire will help to prevent errors when connecting them up. Solder the red and black battery connector wires to the +9V and 0V points respectively on the p.c.b. (or use pieces of similarly-coloured stranded wire if soldered connections are needed to the batteries). Adjust the wiper of preset VR3 to approximately mid-track position. Solder regulator IC3 in position noting that the back is towards the centre of the p.c.b. (the part that protrudes is towards the edge).

Close-up detail of heatsink mounted on IC3.


Note that the specified regulator has a current rating of 1A. Although the average requirement of the circuit is much smaller than this, in use there are peaks of several hundred milliamps and this regulator will cope well. Due to the low average current, only a small heat sink is needed. In the prototype, this consisted of a piece of sheet aluminium size 50mm 15mm bent through right angles as shown in the photograph. It was drilled with a small hole and attached securely to the back of IC3. When choosing the internal loudspeaker also take into account the size of the battery pack to be used. The loudspeaker used in the prototype was a 90mm 50mm elliptical type as used in small radios and TV receivers. Make sure that the power rating is sufficient because many small loudspeakers are inadequate in this respect. Do not use one having a rating of less than 2W. Decide on a position for the p.c.b. and battery pack by arranging them on the bottom of the box. Consider also the loudspeaker which will be mounted on the lid section and the various off-board sockets and switches. When the p.c.b. is finally attached, there must be several millimetres of clearance between the copper tracks and the bottom of the box. This will avoid any possibility of short-circuits. Take care that the heat sink does not touch wiring or other electrical connections.

RIGHT LEADS

CONSTRUCTION

A metal case should be used as an enclosure for the Handy-Amp. A vinyleffect aluminium box was used for the prototype unit because it gave a good appearance. Do not use a plastic box since this will not provide any screening and hum pick-up might be a problem. Construction is based on a single-sided printed circuit board (p.c.b.). The topside component layout and full size underside copper track foil master are shown in Fig.3. This board is available from the EPE PCB Service, code 273. Most of the components are mounted on the p.c.b. although there are quite a few off-board parts which will be hard-wired

BORING BUSINESS

Mark the positions of the holes for the mounting bushes of potentiometers VR1 and VR2, also for the l.e.d. mounting clip. Drill these through and, gently bending the l.e.d. leads out of the way for the moment, secure the p.c.b. to the case using the potentiometer fixing nuts. Place washers (or spare fixing nuts) on the bushes on the inside of the case so that only a small amount of each bush protrudes through its hole. Mark through the p.c.b. fixing hole then remove the board again. Mark the positions of the switches, the low-level input jack socket (mono 635mm type), the phono socket, the headphones

Everyday Practical Electronics, August 2000

575

Fig.4. Off-board component connection details.


output jack socket (stereo 635mm see Important Note) and external loudspeaker sockets. In the prototype, 2mm sockets were used for the loudspeaker, but the type used will depend on personal requirements. Drill the holes and mount the sockets, switches and l.e.d. clip. The case itself is connected to 0V (earth). It is not acceptable for the headphone stereo output socket, SK3, to be of a type where any of its contacts touch the case. If they were to, a short-circuit would be formed and this could damage IC2. This precludes using the ordinary metal sleeved type of 35mm jack socket because, when mounted in position, its outer (sleeve) connection would make contact with the metalwork. There are various ways to avoid this. One method would be to use an insulating sleeve and insulating washers on a standard 35mm unit. However, the method used in the prototype was to use a 6.35mm plastic body stereo jack socket. This had all its connections isolated from the case. Headphones are then connected to it via a 635mm to 35mm converter. Mount the socket and check, using a meter, which tag is which and that none of its tags make contact with the case. The mono jack socket, SK1, used for the low-level input, unlike the headphone output socket must have its sleeve connected to 0V (earth). Since this socket will probably have a plastic body, it will not be done automatically and the sleeve connection will need to be hard-wired to a solder tag attached to the case. The phono socket (SK2 high-level input) must also have its sleeve connected to 0V. If using the specified single-hole fixing type, this will be done automatically. Note that this socket usually has a solder

IMPORTANT NOTE

tag on its bush and this may be used for the SK1 earth connection. If the phono socket is of a fully-insulated type, you will need to make a connection between the sleeve tag and the case using a solder tag (which will also be used for SK1). Referring to Fig.4, carry out all the internal wiring using light-duty stranded connecting wire. By using different colours, you will avoid errors (rainbow ribbon cable is ideal). Note that the two non-sleeve (tip) tags of the headphone socket are joined together so that both headphone units are connected in parallel. Remember to leave all wires interconnecting the various points on the p.c.b. with off-board components long enough to enable the p.c.b. to be removed without straining them, should this ever become necessary. Also, the loudspeaker wires

HARD WIRING

should be sufficiently long to allow the lid of the box to be removed without straining them. Place the loudspeaker in position and mark the fixing holes on the lid of the case. Take care to avoid the p.c.b. (especially the heat sink on IC3) and battery pack positions. Mark out the holes which are needed to allow the sound to pass through. Drill these using a small (say, 15mm) drill then increase the diameter to 5mm approximately. Work carefully because the appearance of the finished project will be spoilt if the holes are drilled carelessly. Carefully clean away any metal particles then attach the loudspeaker. Solder the wires to its tags and apply some strain relief so that they cannot pull free when removing the lid of the case. In the prototype, this was done using a solder tag

576

Everyday Practical Electronics, August 2000

having a long tail. This was attached to one of the loudspeaker fixings. The wires were protected using a short piece of sleeving and the tail of the solder tag was gripped gently around them. Take care that the wires are not so tightly held that a short-circuit is produced. Attach the p.c.b., making sure that it is parallel with the base of the box. Measure the clearance between the copper track side and the bottom of the case. Cut a plastic stand-off insulator to the same length. Slide it into position and secure the p.c.b. using a thin nylon nut and bolt. Gently bending the leads as necessary, push the l.e.d. into its clip. Attach the control knobs to the potentiometer spindles. If the knobs have a white line or spot, this should be arranged to be vertically upwards when the control is at its half-way position.

FINAL ASSEMBLY

Fit self-adhesive plastic feet to the bottom of the case to protect the work surface. Attach the battery pack using a small bracket or adhesive fixing pads (sticky Velcro pads were used in the prototype). Do not connect the battery yet. Immediately before unpacking and handling IC1 and IC2, touch a metal object which is earthed (such as a water tap). This will remove any static charge which might exist on the body. This is a wise precaution because the i.c.s are static-sensitive and could be damaged by such charge. Insert them into their sockets with the correct orientation. Place the lid of the case in position but do not actually attach it. Make a final check that nothing is obstructed and, especially, that the heat sink on IC3 is completely free of all wiring and internal components. Make sure switch S3 is off. Before connecting the battery, make certain the polarity is correct. The circuit will be damaged if the polarity is incorrect. Make certain the positive battery connection cannot make contact with the case or the battery will be short-circuited. This could result in damage to p.c.b. tracks. Switch on S3 and check that the l.e.d. indicator lights up.

Begin testing by using the amplifier with a high-level input source, such as the line output of a CD player, cassette deck or the audio output from a video recorder; if this is stereo, use only one channel. Connect it to the phono socket using a piece of mono screened lead with suitable connectors at each end. Set switch S1 to High and S2 to Internal Speaker. Turn both VR1 and VR2 fully anti-clockwise. Switch on the input device and slowly increase VR2. The sound should be heard clearly. Adjust preset potentiometer VR3 so that the sound is undistorted when VR2 is at maximum. You will find that the setting is not particularly critical. Almost maximum resistance was correct for the prototype (that is, the sliding contact almost fully clockwise when viewed from the right-hand edge of the p.c.b.). Leave the system operating for about ten minutes then switch off and check that the regulator heat sink is not excessively

TESTING

hot. If it is uncomfortable to touch, increase its area. When satisfied on this point, attach the lid section. Turn VR2 to minimum again. Switch S1 to Low and connect a dynamic microphone to the low-level input jack. Increase VR2 to approximately one-third of its total clockwise rotation then increase VR1 slowly while speaking into the microphone. The sound should be clearly heard. If the controls are turned up too far, or the microphone is placed too close to the unit, acoustic feedback will become evident. This usually manifests itself as a loud squealing noise from the loudspeaker. Acoustic feedback is a potential problem with any loudspeaker/microphone system. It comes about because sound from the loudspeaker re-enters the microphone and builds up in a positive feedback loop. To prevent it, turn down the controls, move the microphone away and/or point it in the opposite direction to the loudspeaker. Acoustic feedback may be largely eliminated by using headphones instead of a loudspeaker. It is unlikely that the low-level gain will need to be increased. If it is found to be necessary, decrease the value of resistor R1 to 560 ohms or even 470 ohms. Note that excessive gain leads to instability. When connecting an external loudspeaker, always remember to switch off the amplifier first. This will avoid any possibility of loudspeaker connections touching the case and possibly damaging IC2.

POWER ADAPTOR

Fig.5. Using a fuse and diode as protection devices when a power supply adapator is used.

If you wish to use a plug-in power supply instead of a battery, use a 9V d.c. type having a current rating of 800mA minimum. A fuse and polarity-protection diode need to be included (see later) if damage to the unit itself or to the circuit are to be avoided. Attach a power-in type socket to the rear of the box to suit the output plug on the power supply unit. If its sleeve connection does not make contact with the metalwork automatically, you will need to hard-wire this to a solder tag attached to the case. Preferably, the power supply unit should have a fixed polarity with the centre (pin) on the output plug being the positive and the sleeve the negative. If the polarity can be reversed, make sure that the pin is made positive. If the polarity is incorrect, the circuit will be damaged. This is why a diode should be connected in the positive feed wire. If the polarity is incorrect, the diode will not conduct and nothing will happen. The fuse protects against possible short-circuits. Referring to Fig.5, sleeve both end wires of a type 1N4001 diode. Solder the anode (non-striped end) to the centre (pin) connection of the socket. Attach a 20mm chassis fuse holder to the bottom of the box in such a position that the cathode of the diode can reach one of its tags. Solder this in position. A wire from the other fuse tag should then be taken to the +9V point on the p.c.b. Insert a 20mm 1A quick-blow fuse in the fuse holder. Make sure none of the connections to the diode or fuse can touch the case. Use insulation if necessary. 6

Everyday Practical Electronics, August 2000

577

HEADSET COMMUNICATOR
TERRY de VAUX-BALBIRNIE
Portable three-channel communication system
HIS communication system was originally designed to help in the production of short commercial videos. With it, the director is able to hold a two-way conversation with any one of up to three camera operators. It is also possible to speak to all the operators simultaneously. No doubt, such a system could find many other uses, such as in amateur stage work, concerts and sports events etc.

Constructional Project

In the prototype arrangement, the director sits at a small desk console and the remote operators wear units clipped on to their belts. Cables, which may be of any reasonable length, link the remote stations to the main unit. The director (Master) and remote (Slave) operators wear headsets which are plugged into their units. These headsets consist of a pair of headphones (or a single headphone) having a small boom microphone attached (see photograph). For the target applications, headsets are more convenient than loudspeakers. They provide hands free operation and allow the remote operators to move around freely (within the limits set by the interconnecting cables). Incoming speech cannot enter any microphone used to pick up the sound of the performance and cannot be heard by the audience. Headsets (while worn) are free from acoustic feedback (the howling noise which is produced when the sound from a loudspeaker re-enters a microphone and builds up in a loop). The close proximity of the microphone to the speakers mouth provides very clear communication even when there is a lot of extraneous sound or when he or she only whispers. Power is supplied using four AA size alkaline cells housed inside each unit. The current requirement is 25mA approximately (40mA for the master unit) and the specified batteries should provide at least 50 hours of operation. For safety reasons, the system MUST NOT be operated using a mains-derived supply such as a plug-in adaptor.

WIRED FOR SOUND

The Headset Communicator system units showing (left to right) the master unit, three slave units and a headphone with boom mic.

FREE SPEECH

MASTER UNIT

The Master unit is built in a sloping front instrument case (see photograph). The headset is plugged into a pair of sockets on the front and sockets on the rear panel connect the cables leading to the slave units. On the top, there is an on-off switch and associated l.e.d. (light-emitting diode) On indicator. There is also a three-position Slave Select rotary switch (S2) which selects which slave (A, B or C) is to be placed on line, a momentary-action pushbutton switch which provides the Talk to All function and a Volume control. Rotary switch S2 has three associated l.e.d.s (Red, Yellow and Green) which confirm the slave unit selected. These will be found useful when the unit is being used under dim conditions. Note that while the talk to all switch (S3) is being operated, only the remote station set by the S2 can be heard.

and the cable leading to the master unit, there is an on-off switch, l.e.d. on indicator and volume control. One particular feature of this circuit is that the operators voice is heard in his or her own headphones. This practice is used in telephony and helps the speaker to regulate his or her voice level. It also allows the user to hear someone speaking direct without the muffling effect of the headphones. The amount of voice feedback may be adjusted for each station at the setting-up stage. It may even be reduced to zero if required.

HOW IT WORKS

SLAVE UNIT

Each slave unit is built in a small plastic box having a belt clip attached (see photograph). As well as sockets for the headset

The basic circuit for the Headset Communicator is shown in Fig.1 and this is the same for both Master and Slave units. Each unit may be considered as having one input and one output the Listen (L) and Talk (T) lines respectively plus a common Earth. By linking the talk line of one unit to the listen line of another and the listen line of the first to the talk of the other and also making the common earth connection, two-way communication would be established. Of course, additional switching is

734

Everyday Practical Electronics, October 2002

Rear panel shows the three XLR type sockets for connecting up the Slave units.

Completed Slave unit with belt clip attached to the lid.

Fig.1. Circuit diagram for the Headset Communicator. This is the same for both the Master and each Slave unit.
needed in the Master unit to select the slave unit to be communicated with. This aspect of operation is looked at later. Six-volt battery B1, supplies current through On/Off switch S1 and diode D2. The diode provides reverse-polarity protection. Thus, if the supply were to be connected in the wrong sense, D1 would fail to conduct and no current would flow, thus preventing damage to semiconductor devices. Note that a Schottky diode is specified for D2. This introduces a smaller forward voltage drop than a conventional diode. Capacitor C8 provides a reserve of energy and allows peaks of power to be delivered especially when the battery is nearing the end of its useful life. Light-emitting diode, D1 is the on indicator and operates through current-limiting resistor R12. The microphone section of the headset, MIC1, is connected to the circuit via socket SK1. This microphone is of the electret type and so requires a power supply for its internal preamplifier. This is derived from the nominal 6V supply through resistor R1. The speech signal is applied, via capacitor C1 and input resistor R2, to the inverting input (pin 2) of operational amplifier (op.amp) IC1a. This is one half of a dual unit. The function of the other section, IC1b will be looked at presently. The non-inverting input of IC1 (pin 3) is connected to a nominal 3V reference derived from the potential divider comprising fixed resistors R3 and R4 working in conjunction with capacitor C2. Since the op.amp is powered from single supply rails (+6V and 0V), this procedure allows for a false zero to be set allowing both the positive and negative half-cycles of the input waveform to be amplified. Fixed resistor R5 and preset VR1 connected in series apply negative feedback between IC1 output (pin 1) and the inverting input (pin 2). The value of the feedback resistance divided by that of input resistor R2, determines the gain. With preset VR1 at minimum adjustment this will be unity and when at maximum 23. In fact, these values are negative but this has no practical consequence here. Preset VR1 will be adjusted at the end of construction to provide a suitable gain for the particular microphone used. If tests prove the gain to be too small, the value of resistor R2 could be decreased.

COMPLETELY BIASED

TALK TO ME

The output signal from IC1a flows, via capacitors C3 and C9, to the Talk (T) pin of input/output socket SK3. In addition, some of this signal flows through preset potentiometer VR2. The sliding contact selects a fraction of this and passes it, via capacitor C4 and resistor R6, to the inverting input (pin 6) of IC1b. The noninverting input (pin 5) biasing arrangements are the same as for IC1a, using fixed resistors R7 and R8 in conjunction with capacitor C5. A further signal arrives at IC1b inverting input from the Listen (L) pin of socket SK3 through capacitor C10 and resistor R9. This has been derived from the talk output of the remote unit. Op.amp section IC1b may be regarded as a mixer for the local and distant signals and since feedback resistor R10 is equal in value to input resistors R6 and R9, the gain

Everyday Practical Electronics, October 2002

735

is unity (actually 1). The level of the local (own voice) signal may be adjusted using preset VR2.

VOLUME CONTROL

COMPONENTS
Resistors

(Master + Approx. Cost one Slave) Guidance Only excl headset, leads, case & batts
Miscellaneous
s.p.s.t. rocker or toggle switch SK1, SK2 35mm stereo jack socket (or as required for headsets used) see text regarding headphone socket (2 off) B1 6V alkaline battery pack (4 x AA), with holder and connector clip Printed circuit board available from the EPE PCB Service, code 369; headset having electret microphone and an earphone or earphones (impedance 30 ohms approximately); 8-pin i.c. socket (2 off); commercial XLR leads (or homemade leads) total of 3 required; connecting wire; small fixings; solder, etc. S1

30

The output of IC1b (pin 7) is applied, via capacitor C6, to the top end of the potential divider comprising fixed resistor R11 connected in series with panel-mounted potentiometer VR3. A fraction of the signal is obtained from the sliding contact and applied to the input (pin 2) of power amplifier IC2. This device has been designed to allow an 8-ohm loudspeaker to be connected between its outputs (pin 5 and pin 8) to develop one watt approximately. Here headphones are used and, since these have a higher impedance than a loudspeaker (30 ohms approximately), the available power is reduced. However, only a small amount of power is needed to drive the headphones at full volume so this method works well. The headset volume may be adjusted using VR3. The specified power amplifier (type TDA7052 having no suffix) does not require a connection to pin 4. However, there are variants of this device having a suffix and which have a d.c. volume control. If one of these must be used, then pin 4 will be used to control its gain. To match the characteristics of the specified unit, it would be necessary to impose a voltage greater than 15V on pin 4 which sets it to maximum. This could be done using a potential divider and more will be said about this later.

ALL UNITS (Master and Slaves as required)


R1 10k See R2, R5 1k (2 off) R3, R4, R6, R7, R8, R9, R10, R11 47k (8 off) page R12 270W Rx 56k Ry 22k (Rx and Ry not needed if IC2 is as specified see text) All 025W 5% carbon film.

SHOP TALK

Potentiometers
VR1, VR2 VR3 22k sub-min. enclosed preset, vertical (2 off) 10k min. rotary carbon, log. 4m7 radial elect. 16V (2 off) 22m radial elect. 16V (2 off) 10m radial elect. 16V (4 off) 100n ceramic 220m radial elect. 16V 3mm red l.e.d. 1N5817 1A Schottky rectifier diode TL072 dual op.amp TDA7052 (no suffix) power amplifier (see text)

ADDITIONS FOR MASTER


R13 S2 S3 D3 to D5 270W 025W 5% carbon film 4-pole 3-way rotary switch d.p.s.t., momentary action, push-to-make switch 3mm l.e.d.s, one each red, yellow, green

Capacitors
C1, C4 C2, C5 C3, C6, C9, C10 C7 C8 D1 D2 IC1 IC2

Semiconductors

Sloping front instrument case with aluminium top and plastic sides, size 170mm x 143mm x 55/31mm; XLR panel mounting socket (3 off); plastic feet; solder tag.

MASTER SECTION

ADDITIONS FOR EACH SLAVE


Plastic box size 114mm x 76mm x 38mm; panel mounting XLR plug; belt clips if required; 6V alkaline battery pack (4 x AA) with holder and connector clip.

How the Master console is connected to the slaves is shown in Fig.2. The master Listen and Talk lines are directed to one of sockets A, B or C using switch S2. This switch is a 4-pole 3-position type. The talk and listen lines are connected via switch S2a and S2b respectively while the l.e.d. corresponding to the chosen socket receives current via S2c and current-limiting resistor, R13. Pole d is not used. The All Talk function (enabling the Master to speak to all slave units simultaneously) is provided by connecting the master talk line to all three sockets. This is

Fig.2. How the Master console unit is connected to the three Slave units.

736

Everyday Practical Electronics, October 2002

Fig.3. Printed circuit board component layout and full-size copper foil master.

carried out using a double-pole momentary action switch S3.

TAKING THE LEAD

In the prototype system, the interconnecting leads were of the commercial variety fitted with a 3-pin XLR line plug on one end and a matching line socket on the other. These connectors are widely used in the industry and are normally used for balanced audio applications. Before purchasing XLR leads, check that they are of the standard pattern. Some cheap cables intended for unbalanced microphones, have only one inner conductor with the screening connected to two of the pins. For this circuit, you need two available inner conductors plus the screening. You could, of course, use homemade leads constructed using two-core screened wire and stereo-type jack (or XLR) connectors.

CONSTRUCTION

Construction of the Headset Communicator is based on four identical single-sided printed circuit boards (p.c.b.s). This, of course, assumes that three slaves are required. These boards are available from the EPE PCB Service, code 369. The p.c.b. topside component layout and full-size underside copper foil master pattern are shown in Fig.3. Begin construction of each p.c.b. by drilling the two fixing holes as indicated. Next, solder the i.c. sockets in position, also the link wire connecting IC2 pin 2 with Volume control VR3 sliding contact, all resistors (including the presets) and the capacitors. Apart from C7, the capacitors are all electrolytics so take care with their orientation. Note that there are four holes which will have been left empty see later. Now solder pieces of stranded connecting wire to the talk (T), listen (L) and earth (E) points on the completed p.c.b. Connect similar pieces of wire to the MIC1 and VR3 positions. Use different colours to avoid errors later. Adjust presets VR1 and VR2 to approximately mid-track position.

TESTING

It is advisable to check the operation of each circuit board at this stage because it is then much easier to correct

minor problems. Solder the battery conITS WORKING nectors to the +6V and 0V p.c.b. pads, Do not put the headset on initially in taking care over the polarity (red wire for case of sudden loud clicks and other nois+6V). es. Satisfy yourself on this point before Solder jack sockets (or the required type putting it on. to match the headset) to the MIC1 and Connect the battery and note that the On Phones wires. Note that the sleeve of the l.e.d. operates. If acoustic feedback is evimicrophone plug must connect to rightdent (which should not occur when the hand MIC1 wire on the p.c.b. that is, the headphones are worn) adjust Volume conone connected to the 0V line. In the prototrol VR3. type unit, the microphone plug was a Listen to the headphones and speak into 35mm stereo jack type but either tip the microphone. If you can hear your voice connection could be used because they clearly, the circuit is working. If it is obviwere connected together internally. ous that the microphone gain is too small The prototype headphones were also (quiet sound even with VR1/VR2/VR3 set wired to a 35mm stereo jack plug. In this to maximum) reduce the value of resistor case, each tip connection was responsible R2 to 560 ohms (after switching off and for one unit while the sleeve was removing the i.c.s). common to both. This enables the headRepeat all this with the other circuit phones to be used individually for stereo boards then, observing the anti-static preapplications. cautions mentioned earlier, remove the i.c.s Here, both tips need to be connected from their sockets and replace them in their together so that the units appear in parallel anti-static packaging. De-solder the jack and provide mono operation. The common sockets, potentiometer and positive battery tips connect to one wire and the sleeve to connector lead. Connect a piece of strandthe other. This procedure may need to be ed wire to the +6V p.c.b. point instead. modified depending on the plugs fitted to the headsets. Referring to Fig.5, the Slave unit wiring, solder potentiometer VR3 tags to its wires in the sense shown. Adjust it to approximately mid-track position. Insert the i.c.s into their sockets. Since these are CMOS devices, they could be damaged by static charge which may have accumulated on the body. To avoid possible problems, touch something which is earthed (such as a metal water tap), before unpacking them and handling the pins. Do not throw away the packaging because it will be needed again later. Completed prototype circuit board.

Everyday Practical Electronics, October 2002

737

MASTER BOX

The sloping front aluminium instrument case used for the prototype Master unit gives a professional appearance, see photographs. There is an advantage in using a box that is of part plastic construction. This is because a case made entirely of metal will need additional insulation on the Phones output socket. Find the best positions for the switches, panel potentiometer, l.e.d. indicators and sockets. The headset socket should be located on a plastic part if possible. Decide whether commercial XLR leads are to be used or whether leads are going to be made up so that the appropriate connectors may be chosen. In the prototype, XLR sockets were used in the master with a matching plug on each slave unit. Drill holes for all these parts. Mark out and drill the holes for mounting the p.c.b., battery holder and any remaining parts, including one for the solder tag (in a metal part). Drill small holes to correspond with the anti-rotation tabs on the rotary switch and

potentiometer. This prevents their bodies possibly turning in service and breaking off soldered connections.

INTERWIRING

Attach all internal components and, referring to Fig.4, complete the interwiring to off-board components. Note how resistor R13 is connected. Apply some sleeving to the joints at the l.e.d. leads and any bare wires to prevent short circuits. Using a multitester, check that the solder tag makes good contact with the metal part of the case. The wires connected to it should be twisted together and hooked through the hole before soldering. Note that neither Phones socket connection may make contact with 0V (earth) that is, any metal part of the case. If, as in the prototype unit, the socket is mounted on a plastic part, there will be no problem. If the socket must be mounted on metal, the best approach would be to use a fullyinsulated jack socket. Unfortunately, most types make automatic connection of the sleeve to the case.

If necessary, you will need to make an insulating sleeve (or a shouldered plastic bush) and use plastic washers to isolate it from the metalwork. Use a multitester to check that the sleeve does not make electrical contact with earth before proceeding. Take care to wire up the Listen/Talk selector and the Talk to All switches correctly. The pole lettering and contact tag numbering (see inset dia.) is as shown on most switches of this type. If using XLR connectors, pin 1 should be connected to Earth (0V) along with the solder tag which connects to the metal body. In the prototype, pin 2 and pin 3 are used for the Talk and Listen connections respectively. All the wires connected to these sockets will need strain relief. In the prototype, this was done by means of a cable tie passed through slots in the bottom of the case. This will help in preventing the wires from breaking free in service.

MICROPHONE WIRING

The microphone input socket may be mounted on a metal part because its sleeve

Fig.4. Interwiring from the Master circuit board to off-board components.

738

Everyday Practical Electronics, October 2002

Fig.5.Interwiring details for one Slave unit.

Packing the components into the Slave unit.


the units switched off. Fit the batteries then plug in the interconnecting leads and headsets, with integral microphone booms. Turn all the Volume controls to minimum and switch the units on. The l.e.d. On indicators should operate. The headphones should be listened to with caution in case the Volume controls have been wired in the wrong sense and a sudden loud noise develops. Test the operation between the Master and each Slave unit. Preset VR1 should be adjusted in each unit so that the maximum volume set by VR3 is not too great and that there are no signs of instability. Adjust preset VR2 in each unit for the preferred degree of voice feedback. Check the talk to all function. When satisfied, attach the lids of the cases and label the controls. You will know when the batteries need to be replaced because the sound will become weak or distorted and the l.e.d.s will glow less brightly. In use, always start with the volume turned down to minimum and switch on all units before wearing the headsets. This will avoid any loud clicks.

must be connected to earth (0V). However, it will probably be mounted next to the phones socket for cosmetic reasons. If it is on plastic, you will need to hard wire its sleeve connection to the solder tag. Note the sense of the wiring to the Volume control (VR3) potentiometer tags. This gives conventional operation clockwise rotation increasing the volume. Note also that only one current-limiting resistor, R13, is needed for the slave indicator l.e.d.s. This is because only one l.e.d. can be illuminated at a time.

SLAVE UNITS

Choose plastic boxes of appropriate size for the Slave units and fit the belt clips if required. Check the layout of internal parts

and drill holes for them. Do not forget the small hole needed for the Volume control potentiometer anti-rotation tab. Attach all slave parts and, referring to Fig. 5, complete the internal wiring leaving plenty of slack in the wires. Note that certain connections will be close together so make sure they do not touch and cause a short-circuit. Use additional insulation as necessary. Check that the connections to the plug pins allow the interconnecting lead to make the appropriate connections (Talk to distant Listen and Listen to distant Talk). In the prototype, pin 2 was used for listen and pin 3 for the talk. Connect pin 1 to the solder tag on the plug that connects to the metal body. Take care over the sense of the connections to the potentiometer tags. Attach the control knobs to the spindles of the switches and potentiometers in all units. Leave the lids removed from the cases for the moment to allow presets VR1 to be adjusted. Observing the usual anti-static precautions, insert all the i.c.s into their sockets taking care over the orientation.

ALTERNATIVE POWER AMPLIFIER

Begin final General layout of components on the Master unit metal checking with all front panel.

FINAL CHECKS

If it is impossible to obtain the specified power amplifier (i.e. a TDA7052 without a suffix letter) and you must use one having a d.c. volume control, its gain will need to be configured to maximum to match the characteristics of the specified unit. This may be done by soldering resistors Rx and Ry in the unused positions on the p.c.b. Resistor Rx will be in the upper position which connects to IC2 pin 1 and Ry to the lower position which connects to IC2 pin 4. Resistor Ry may need a 1F capacitor connected in parallel with it. This could be placed on the underside of the p.c.b. Note that this set-up has not been tested and some experimentation may be needed to obtain correct operation. 6

Everyday Practical Electronics, October 2002

739

LOGIC GATE INVERTER OSCILLATORS


GEORGE HYLTON
Part Two
Good sine waves are obtainable at the LC circuit when VR1 is considerably less than the critical value, but to get a pure waveform at A2 output, VR1 must be set so that the circuit just oscillates. It may be simpler to pick off a sine wave output at A1 and extract it via buffer A3. This has a gain of R4/R3. The circuit may be used up to about 1MHz. If VR1 is calibrated it can be used to obtain a reasonably accurate indication of the dynamic resistance of the LC circuit. Simply adjust VR1 to the maximum value for oscillation. Then VR1 is the dynamic resistance. From this the Q can be calculated: Q = dynamic resistance / reactance of L or C at fo This circuit has overall d.c. positive feedback. It would latch up if the d.c. gain of A1 exceeded one. Fortunately, the low d.c. resistance of L keeps gain well below one, so it is d.c. stable. Resistors R1 and R2 set the gain of A2 to unity (1). Driving A2 directly would cause over-violent oscillation, The ratio R2/R1 could be increased to up the loop gain but this is not necessary with typical LC values. In A3, R3 and R4 set the gain and working point and R3 also provides some buffering. With VR1 set correctly there is no protection-diode conduction. This implies a VR1 of slightly less than the dynamic resistance 2FfLQ or Q/(2FfC). However, VR1 can be less than optimum without seriously impairing the sine wave at the LC.

Special Feature

A compendium of practical oscillator circuits for the creative experimenter, all based on inverting logic gates.

month we examined the basic principles which allow CMOS inverters to be used as oscillators, concluding with an example of a Colpitts oscillator. We conclude this two-part series by first examining a ccrystal oscillator circuit.
AST

CRYSTAL OSCILLATOR

The high frequency crystals used to set the clock frequency in computers can replace L in the Colpitts circuit of Fig.10. The circuit is then sometimes called a Pierce oscillator (Fig.11), although this nomenclature is dubious. Since a crystal blocks d.c., a resistance (R1) must be added to allow d.c. negative feedback to set the working point. This resistance should be high enough not to impair the oscillation. Crystal manufacturers specify the value of shunt capacitance needed to trim the frequency to its nominal value. In the pi-network, the two capacitances are effectively in series so each should be twice the quoted shunt capacitance. The frequency can be fine tuned by adjusting one or both of them. It is possible that oscillation may be too violent. A feedback control (VR1) may also be used as with the Colpitts oscillator. Crystal manufacturers may specify a safe operating voltage and VR1 can be set to

ensure that it is not exceeded. Generally speaking, it is sufficient to set VR1 so that reliable oscillation (in the face of falling supply voltage, etc.) is just feasible. For crystals designed to generate frequencies below about 1MHz, or above about 10MHz, special circuit arrangements may be needed. Consult the manufacturers data sheet.

TWO-TERMINAL LC

The need for transformers or twin capacitors can be avoided by using a socalled two-terminal oscillator circuit. This means that the frequency-determining LC circuit can be connected by just two leads, those marked X in Fig.12.

Fig.12. Two-terminal LC oscillator. A2 provides the required phase inversion. A3 can be added as a output buffer.
With R1 = R2, A2 has a gain close to one, so it is just a voltage inverter. Then A1 must provide the gain needed for oscillation. The critical condition is that VR1 should be just less than the effective resistance of the LC circuit at its resonant frequency fo. The effective resistance is called the dynamic resistance and is Q times the reactance of L or C at fo. For a usable coil the Q quality factor is unlikely to be less than five, and may be several hundred.

WIEN BRIDGE SINE WAVE OSCILLATOR

Fig.11. Pierce crystal oscillator. Here the crystal replaces L in the Colpitts circuit.

The reactive (RC) arms of a Wien bridge (Fig.13) can be used to set the frequency of a sine wave oscillator formed around an op.amp (Fig.14). In a Wien bridge, when R1 = R2, C1 = C2 (the usual case) balance (zero output) is obtained when V2 = V3, in which case C then has a reactance equal to R. This occurs when the input frequency fin is 1/(2FCR), usually called fo. Tuning is conveniently effected by using a two-gang potentiometer for the two controlling resistors (R1 and R2) so that they are always equal. In this way balance is maintained as these resistors are adjusted.

742

Everyday Practical Electronics, October 2002

resistance, VR1, has been added. Without it the circuit would cease to oscillate as R is reduced towards zero. The oscillation frequency is: fo = 1/(2pC(R + VR1)) In fact, there is a hidden component in the series arm: this is the output resistance of inverter A2 and it must be compensated for by an increased resistance in the parallel arm. If this is not done, feedback varies Fig.15. Inverter gate version of Wien oscillator. The A3 secas R is adjusted and it tion can be added to inject an external synchronising signal. is impossible to scales which are very cramped at the highobtain a good waveform over the tuning frequency end. Frequency sweeps range. (max./min.) of 10 are then a practical limit, though the circuit will oscillate over a AMPLITUDE LIMITING wider sweep. No device for automatic amplitude limThe circuit can be used as a selective iting is shown in Fig.15. The job could be amplifier with input injected via a highdone by substituting a thermistor for the impedance buffer A3. In this case VR2 is a feedback resistance across A2 as in Fig.16. sharpness control and for greatest selectivVR2 would then provide oscillation level ity is set for just not oscillating. The adjustment and should have a mid-value buffer amplifier may also be used, if equal to the working thermistor resistance. required, to inject a frequency-locking sigUnfortunately, there are really no suitnal into the oscillating circuit. able thermistors available to the average An injected signal of a few mV can synhobbyist. The sub-miniature bead thermischronise the oscillator. How long it stays tors needed are very expensive. Cheap synchronised depends on the frequency types are physically too bulky and do not stability of both the oscillator and the sync heat up enough at the small signal levels in input. Injecting a larger signal increases the circuit. the locking range but at the risk of false locks where one frequency bears some fractional relation to the other. (Often the waveform then shows some periodic distortion.) Multi-band operation is possible by switching-in different pairs of capacitors C. For consistent performance each pair must be very accurately matched.

Fig.13. Wien bridge.


In oscillators, use is made of the fact that RC arms of the bridge form a frequency-selective voltage divider whose output is greatest at fo. At frequencies away from fo, output falls. When this network is used as a positive-feedback path in an amplifier (Fig.14) and the gain is just sufficient for oscillation, a sine wave at fo is generated. Unfortunately, the Wien network is only very weakly frequency-selective. It does a poor job of discriminating against harmonics produced by the amplifier overloading. The waveform is distorted. A solution used in commercial Wien oscillators for audio work is to provide a distortionless means of automatically restricting gain to be just sufficient for oscillation. Very pure sine waves can then be obtained. A common method is to use a negative temperature coefficient (n.t.c.) thermistor for the R3 resistance. As oscillation builds up the signal warms the thermistor whose resistance falls. This increases the negative feedback to the inverting input terminal, damping down the oscillation. The standard circuit (Fig.14) does not translate into inverter-oscillator form because an inverter has only one input terminal. It can, however, be adapted to a 2inverter circuit, as illustrated in Fig.15. Inverters A1 and A2 are used in their linear mode and the parallel-RC arm now creates negative feedback to A1 while the series RC arm conveys positive feedback from A2 to A1. The circuit oscillates at fo when the gain of A2 (adjusted by VR2) slightly exceeds two. An extra preset

Fig.16. Using a thermistor in place of RF in Fig.15.


Vout must drive enough current through the thermistor to reduce its resistance sufficiently to obtain low distortion. Since CMOS inverters cannot deliver much current it is desirable to keep the thermistor resistance fairly high, say 10k. The a.c. voltage across it is unlikely to exceed about 3V r.m.s. The power available to warm the thermistor is then 09mW. For reliable operation over a range of ambient temperature this amount of power must cause a temperature rise of at least 20C. If very low distortion is not required, a fairly good sine wave can be obtained from the circuit as shown in Fig.15 if set-up carefully, as follows: Set R to maximum. Set VR2 for just oscillating. Set R to minimum (zero). Without altering VR2, set VR1 for just oscillating. Repeat this procedure then, if necessary, make minor adjustments so as to obtain the best compromise performance over the tuning range. The final result will depend on how well the two sections of the potentiometer are matched. Linear-law two-gang pots are usually better than log-law, but give tuning

DUAL INTEGRATOR OSCILLATOR

Fig.14. Wien bridge oscillator using an operational amplifier.

An inverter with feedback from output to input via a capacitor (as with A1 and A3 in Fig.17) has a gain which falls off as the frequency is raised. In a sine wave oscillator this reduces the harmonics which result from distortion. The ability to yield good sine waves without special amplitude control circuitry is especially useful at very low frequencies, where conventional control using thermistors is difficult. (The resistance of the control device varies over the oscillation cycle and causes distortion.) An inverter with capacitive feedback produces a phase shift. Two inverters, each giving a phase shift of 90 in the same direction, give a total of 180, which is phase inversion. When cascaded with a simple inverter and connected in a ring, the overall feedback is positive at the 90 frequency. Here this is the frequency for which the reactance of C equals R. An inverter with capacitive feedback is often referred to as a Miller integrator, or just an integrator. The frequency generated by the type of circuit in Fig.17 is the same as for a Wien network oscillator (fo = 016/(RC)). With the values shown the

Everyday Practical Electronics, October 2002

743

Fig. 17 Dual-integrator oscillator. Oscillation level is set by VR2. The two outputs V1 and V2 are equalized by VR1 and are 90 apart in phase.
range is roughly 300Hz to 3300Hz. The range can be switched by substituting other pairs of capacitors, accurately matched. When R is in megohms and C is in microfarads, the frequency is in Hertz (Hz). Because of the good discrimination against harmonics it is easier to achieve a respectable sine wave than with the Wien oscillator. The circuit also has the useful property of yielding two equal output voltages (V1 and V2) phased 90 apart (in quadrature). On the other hand setting up to achieve a good performance over the tuning band (by adjusting VR1 and VR2) involves using an oscilloscope and doing a fair amount of fiddling. Start with VR1 and VR2 set halfway. Trim VR1 to equalise V1 and V2. Trim VR2 for the best waveform. The tuning range is somewhat affected by these settings. To achieve the best amplitude stability one of the fixed resistances in series with the tuning resistances may need to be trimmed (at the h.f. end of the band). some frequency then, going round the loop, the three phase shifts add up to 180. This is inversion. The reactance is twice the resistance for series C, shunt R, and the reverse for series R and shunt C. The fed-back signal at A1 is now in step with the original signal. Feedback is therefore positive and the circuit oscillates. If the 180 phase shift occurs at only one frequency then that will be the frequency of oscillation.

the gain of an inverter so the circuit oscillates strongly. Unfortunately, the strong oscillation drives the internal protection diodes into conduction. The effect is to raise the frequency spectacularly but unpredictably. It would be possible to add swamping resistances but a better alternative is to use the circuit in Fig.18d. Here the phase shifting is done by incorporating the RC network into an integrator, the amplifier being one of the inverters. The inverter input terminal is now a virtual earth point and the signal level there is low enough to avoid the worst effects of protection-diode conduction. In a ring of three such integrators each produces a lagging phase shift of 60. The oscillation frequency is theoretically fo = 008/(CR) As before, fo is in Hertz when CR is in megohms times microfarads and so on.

RING VCO

PHASE SHIFTERS

Two standard ways of achieving phase shift are shown in Fig.18c to Fig.18d. The first is passive the required 60 shift occurs at the frequency at which the series arm has twice the impedance of the shunt arm. At that frequency the attenuation factor is two (i.e. half the voltage is lost). This is likely to be much less than

If, in circuits using Fig.18c, the resistances and capacitances are reduced to zero the circuit reverts to that in Fig.18a. It might be expected to display a stubborn stability. Far from it! It oscillates, but at a high frequency. The explanation is simple. We may have removed our Rs and Cs but the circuit has its own built-in equivalents. R is now the output resistance of each inverter and C the input capacitance of the following one. In a particular case R might be 10k9 and C might be 10pF. These act like those in Fig.18c. The 60 frequency is: fo = 1/(FRC) = 3MHz approximately.

RING OSCILLATORS

The three inverters of Fig.18a are connected in a loop or ring. If the input to A1 is positive then the output of A3 is negative. Since this is fed back to A1, it opposes the positive input. The ring is a negative feedback loop with total feedback and (accidents barred) it will be stable. Accidents do happen, though, as will be shown later. Referring to Fig.18b, if we now interpose between successive stages networks which produce 60 phase shift to signals at

Fig.19. Dual-quadrature oscillator. Each twin RC network produces 90 shift at fo.


Both the output resistance and the input capacitance of an inverter are affected by the operating voltage. The output resistance is especially strongly affected. In experimental tests using a CMOS 4069 inverter, biased to operate in the linear region of the input/output curve, the output resistance measured 16k9 when VCC was 5V, falling to 5k9 when VCC was 15V. This means that the zero component ring of Fig.18a is in reality a voltage-controlled oscillator, with VCC as its control voltage. Oscillation may be possible at VCC down to 2V, where the frequency is quite low. At high VCC it may be tens of megahertz. Note that there is a real risk, at high VCC, of the current drawn becoming excessive and overheating the chip. Note also that while standard CMOS i.c.s like the 4069

Fig.18. (a) Three-inverter ring. (b) With added phase-shift circuits. (c), (d) Alternative phase shift networks.

744

Everyday Practical Electronics, October 2002

are rated to work at up to 15V their modern equivalents like the 74HC04 have much lower maximum VCC ratings. It is possible to bring down the frequency while retaining voltage control. Add real capacitors for C while leaving R at zero.

4-PHASE SHIFT RING

A ring with three equal phase shifters (Fig.18b) is a neat means of generating a three-phase signal. But suppose you need some other number of phases. Any number over two can be provided, with one precaution. The total number of inverters in the ring must be odd. If it is even there is overall d.c. positive feedback and the circuit latches up. If you need an even number of phases you have to add one plain inverter (with no associated phase shift components) to keep the d.c. feedback negative. One potentially useful arrangement is to have four shifts of 45 each. This enables outputs to be selected at multiples of 45, notably 90. The necessary fifth inverter can be used as a gain-adjustable stage to set the oscillation level. The frequency is that at which R and C have equal impedances, i.e. fo = 1/2FCR. The loop shift must be 180. For a 3-section phase shift the average per section must be 60, for four sections 45, and so on. It is also possible to generate outputs phased 90 apart with a 3-inverter ring (Fig.19). Here two pairs of double RC networks each generate a 90 shift. The frequency is about 1/(2FRC). In theory, three or more RC (or CR) networks can be cascaded to give an overall phase shift of 180. A single inverting amplifier can then maintain oscillation, see Fig.20. These circuits are usually referred to as phase shift oscillators (though of course phase shifting is involved in all the oscillators we have just been discussing). Phase shift oscillators may look neat but they have two major disadvantages which stem from the fact that the second RC section loads the first, the third loads the second and so on. This greatly increases the attenuation at fo. For a network with three cascaded RC or CR

PHASE SHIFTING

sections, all with equal R and C, the gain needed to sustain oscillation is nearly 30. For a foursection network it is nearly 20. A single inverter may not provide enough gain. The second snag is that it is no longer possible to pick off outputs evenly spaced-out in phase. Also, the voltage diminishes at each successive section. A third problem is that the gain is not readily adjustable. If, however, one inverter provides more than enough gain a reduction can be made by shunting off some of the current into a sec- Fig.20. Phase-shift ond inverter (Fig.21), Four-section RC. which presents a load of R1 and can be used as an output buffer. (This trick can be used with other oscillators.) For a three-section RC network fo = 039/RC. For a four-section RC network fo = 019/RC. Attenuation can be reduced by tapering the networks. Successive resistances are multiplied by a factor N and successive capacitances divided by N. As N is made very large the 3-section attenuation factor falls towards eight and the 4-section towards four. Making N = 10 achieves most of the improvement and even N = 2 is worthwhile. The RC network discriminates against harmonics and even if the input to a multisection network is a square wave the output is a fairly pure sine wave. However, it occurs at a high-impedance point and can only be used if picked off by a very high impedance buffer. This adds its own quota of distortion.

oscillators. (a) Three-section RC. (b)


(they go well back into the valve era) have elicited from circuit analysts some formidable feats of mathematics. But if you need a low-distortion oscillator you will be well advised to leave them alone and stick to Wien or dual-integrator circuits! Whilst we have concentrated on the use of basic CMOS inverter gates, the principles can equally well be applied through the use of dual-input inverting gates, such as NAND and NOR. $

FORMIDABLE

Phase shift oscillators are fascinating circuits which over their long history

Fig.21. Gain-adjustment circuit. R1 acts as a load on A1.

TM

Radio Bygones
The leading magazine for vintage radio enthusiasts
The magazine is published six times a year, and is available by postal subscription. It is not available at newsagents. RADIO BYGONES, Wimborne Publishing Ltd, 408 Wimborne Road East, Ferndown, Dorset BH22 9ND. Tel: 01202 873872. Fax 01202 874562. Web sites: www.radiobygones.co.uk www.radiobygones.com

Tel: 01254 830761 Fax: 01254 830408


Email: sales@magtrix.co.uk Website: www.magtrix.co.uk Everyday Practical Electronics, October 2002

745

INTERFACE
Robert Penfold
Adding MSCOMM Active-X control to your PC
two Interface articles were Tdevotedpreviouspermit serialMSCOMM to the use of the ActiveX control to communiHE

cations with Visual BASIC programs. The advantage of this method is that it will work with any 32-bit Windows operating system, including Windows XP without the need for any third-party add-ons. The main drawbacks are that this control is not included with anything less than the Visual BASIC Professional Edition, and it is something less than straightforward in use.

With Microsoft Word for example, it is launched by selecting Macro from the Tools menu and choosing Visual BASIC Editor from the submenu. No form is produced when VBA has finished loading, but a form can be added by selecting User Form from the Insert menu. You then have something like Fig.1, which is similar to the normal arrangement in Visual BASIC. The next task is to go in search of the MSCOMM control, and the first step is to choose Additional Controls from the Tools

there are differences. The fact that VBA is not designed to produce standalone programs enforces a few changes, but there are differences in the code, such as the exact structure of conditional routines. Programs written for Visual BASIC will usually require at least a small amount of rewriting in order to make them work with VBA. This point is demonstrated in the first VBA listing (Listing 1), which is for a simple program that reads single bytes from a serial port and displays them on a label component.

MSCOMM and VBA


Software topics usually produce a certain amount of feedback from readers, and the pieces on MSCOMM are certainly no exceptions. A few readers pointed out that this control is included with Microsoft Word and Excel as part of VBA (Visual BASIC for Applications). On checking two PCs that had Microsoft Office installed but had never been loaded with Visual BASIC Professional, one had MSCOMM and the other did not. VBA is not only included with Microsoft applications, it is also provided with some software from Corel, Autodesk, etc. However, VBA is not always installed when the Typical option is chosen during installation. It is sometimes necessary to return to the installation disk in order to add VBA. The presence or absence of MSCOMM probably depends on the exact software installed on the PC. The more upmarket the software the greater the chances of success. It would certainly seem to be the case that it is not included with all versions of Microsoft Office. It is not difficult to ascertain whether MSCOMM is present on a PC. Launch Windows Explorer and then use the search facility to scan the hard disk for a file called MSCOMM.OCX. The MSCOMM ActiveX control is not installed if this file is not present on the hard drive. If this file is present, it would probably be possible to use it with one of the free versions of Visual BASIC as well as with VBA.

Fig.2 (below). Adding MSCOMM, if it is available.

Fig.1 (above). The Visual BASIC for Applications (VBA) set up and ready to use.

Same Difference
VBA is not really intended for producing normal software, and its usual role is in the production of extra commands for applications programs. However, at a pinch it can be pressed into service as a means of producing software for use with PC based projects. The first task is to launch VBA from within the host application, and it is normally accessed via the Tools menu.

menu. This brings up a window like the one of Fig.2, and it is then a matter of scrolling through the list looking for MSCOMM. It will not be called MSCOMM in this list though, it is more likely to be called Microsoft Communication Control version 6.0 or something similar to this. Having found the right entry in the list, tick its checkbox and then operate the OK button. A yellow telephone icon should then appear in the Toolbox, and this enables MSCOMM to be added to the form in the usual way.

In addition to MSComm and a form, it requires two buttons and a label. The captions for buttons one and two (CommandButton1 and CommandButton2) are respectively changed to START and EXIT.

Listing 1
Private Sub UserForm_Click() End Sub Private Sub CommandButton1_Click() MSComm1.PortOpen = False End End Sub

VB or not VB
Although VBA seems to be widely regarded as identical to Visual BASIC,

746

Everyday Practical Electronics, October 2002

Private Sub CommandButton2_Click() MSComm1.RThreshold = 1 MSComm1.InputLen = 1 MSComm1.Settings = 9600,n,8,1 MSComm1.CommPort = 1 MSComm1.InputMode = comInputModeText MSComm1.PortOpen = True End Sub Private Sub MSComm1_OnComm() If MSComm1.CommEvent = 2 Then Label1.Caption = Asc(MSComm1.Input) End Sub Operating the START button switches on communication with the serial port, selects the required port, and sets the required operating parameters. This works in the same way as the code for the Visual BASIC version described in a previous Interface article. The routine used for MSComm1 reads single characters from the port, converts each one to its ASCII value, and then writes that value to Label1. In the original program an If...Then...End If structure was used to check that the right OnComm event had occurred. If the right event had occurred (i.e. a new byte of

The latter is used to generate the values that are transmitted, and its MAX setting should be set at 255. It will then generate integers from 0 to 255, or single bytes of data in other words.

Listing 2
Private Sub CommandButton1_Click() MSComm1.PortOpen = False End End Sub Private Sub CommandButton2_Click() MSComm1.PortOpen = True End Sub Private Sub Label1_Click() End Sub Private Sub MSComm1_OnComm() End Sub Private Sub ScrollBar1_Change() MSComm1.Output = Chr$(ScrollBar1.Value) Label1.Caption = ScrollBar1.Value End Sub Private Sub UserForm_Click() End Sub

chose. Note that the main Word document can be empty, and there is no need to add any dummy text. To use the program on another occasion, load the relevant document and go to the Visual BASIC Editor again. This should contain the program. There can be a problem when tying to run the program, with an error message appearing. This points out that Macros have been disabled and that the program cannot be run. Macros are disabled by default as a means of reducing the risk from macro viruses. Selecting Macros from the Tools menu followed by Security from the submenu enables the security setting to be changed. A dialogue box appears and it has radio buttons that offer three levels of security. The lowest level enables macros to be run with no questions asked. You will be asked whether or not you wish to run the program if the middle setting is selected, and macros are blocked if the highest level is used. If you are used to VBA and its version of the BASIC dialect, VBA programs can be a valid approach to producing software for your PC projects. Even if you do not have MSCOMM on your computer system, VBA can still be used with third party add-ons such as Inpout32.dll to access the serial and parallel ports. One of the free versions of Visual BASIC probably represents a better

Fig.3. The serial reader program operating within VBA.


data had been received), the port was read, the conversion was made, and data was written to the label. With VBA the If...Then...End If structure is not quite the same, and the original routine just causes an error message when used with VBA. In this case the routine can be reduced to a single line of code, with no End If statement required at the end of the routine. In fact it must be omitted or an error message will be produced. The routine for the EXIT button simply closes communications with the serial port and closes the program. The VBA version of the program works as well as the original Visual BASIC version, and it can be seen working within VBA in Fig.3.

Fig.4. The serial transmission program. Values set on the slider control are transmitted from the serial port.
In this case the VBA program can be much the same as its Visual BASIC equivalent. It is the routine for the scrollbar that actually transmits the data, and the new value is sent each time that a change occurs. The Chr$ function is used to convert the value from the scrollbar into an equivalent ASCII character which is then sent to the serial port for transmission. The unprocessed value is displayed on the label component so that the user can see what values are being sent. Again, the VBA program works as well as the Visual BASIC version, and it is shown running in Fig.4. starting point for those starting from scratch. Either way, it is possible to get into visual programming at no cost.

Binary Mode
A couple of readers have pointed out methods of using MSCOMM in binary mode so that the string conversions can be avoided. This is a subject that will be considered in detail when the problem has been investigated fully. Strangely, the Microsoft documentation recommends that the text mode is used for all data transfers using MSCOMM. A possible reason for this is that some facilities of MSCOMM seem to disappear when the binary mode is used. The text and conversion method is a bit cumbersome, but it does have the saving grace that it actually works quite well.

Output
The second VBA listing is for a simple serial transmission program. The form is equipped with START and EXIT buttons, as in the serial port reading program. It also has a label, but this time it is used to show the value generated by a scrollbar.

Lockout Situation
Programs are saved using the Save Document option under the Edit menu. Once the document has saved, this option changes to Save XXXX where XXXX is the program name that you

Everyday Practical Electronics, October 2002

747

Constructional Project

VINYL TO CD PREAMPLIFIER
TERRY de VAUX-BALBIRNIE
Clean up those old records, dust down that turntable and lets get burning!

have a collection of old vinyl records? If so, you might wish to transfer them to CDs. By doing this, you will preserve their value because you will only need to play them once. It may even be possible to enhance the sound by removing some of the background noise and clicks which are found on worn recordings. If you have a CD player in your car or own a portable unit, you will also be able to play your work on the move.
O YOU

SYSTEM REQUIREMENTS

To transfer a recording to CD, you need a computer with a Compact Disc writer installed. Many new machines, of course, already have one of these. If yours is not so equipped, you will find that fitting a CD burner module is inexpensive and straightforward. You do not even need a particularly modern machine. A Pentium 133MHz PC may suffice but a new up-to-date machine will be much quicker (that is, produce a CD at the higher speeds allowed by the writer). Before purchasing any hardware, it is important to check compatibility with the supplier/manufacturer. To record sound files on to the hard drive before transferring them to a CD will require quite a lot of spare capacity. If your drive is almost full, you will need to back up files in order to clear sufficient space. To record stereo tracks in 16-bit resolution at 441kHz (CD quality) you will need some 600MB for one hour of work and you could run into trouble if you do not have at least 800MB available.

copy recording is deficient in bass (low frequencies) but have excessive treble (high frequency content). In other words, it will sound very tinny. More will be said about equalisation presently. A better method would be to use an existing hi-fi amplifier. The record deck would be plugged into its Phono input and a Line (high level) output obtained at the back (the one used for tape recording). This would be connected to the line input on the sound card using a piece of twinscreened wire fitted with the appropriate connectors. The phono connection would provide the necessary equalisation. Unfortunately, many modern amplifiers make no provision for playing old fashioned vinyl discs. You may therefore find that it has no phono input. Even if you do have a suitable amplifier, it may need a long connecting lead to reach the computer station and this could result in hum pickup and degraded performance.

of a magnetic cartridge to line-level. There are also Scratch and Rumble filter pushbutton switches. These may be used to reduce the effects of surface clicks and low-frequency motor or turntable noise respectively. As well as being useful for making CDs, the preamplifier will be found handy by enthusiasts who simply wish to play their vinyl records using a hi-fi amplifier that does not have a phono input. Some readers may even use it for tape or Mini Disc work or for making MP3 files to be sent over the Internet. In operation, the circuit requires some 40mA and the four AA size cells housed internally will provide up to fifty hours of service. A front panel mounted l.e.d. indicator requires some 15mA so, if the user can be trusted to switch the unit off after use, the l.e.d. may be omitted. This would give a significant increase in battery life. For extended periods of use, a larger battery could be placed externally. This unit must not be powered using a mains-derived low-voltage supply (such as a plug-in adaptor).

OVERVIEW

The circuit described here is a small battery-operated stereo preamplifier which provides equalisation and boosts the output

MORE EQUAL THAN OTHERS

Returning to the topic of equalisation, this must be applied if analogue recordings are to be reproduced with any degree of

It is not a good idea to link the record deck to the computer sound card direct by plugging it into the microphone input. Some people have done this thinking, quite correctly, that a magnetic cartridge provides a low-level output comparable with that of a dynamic microphone. Although this may work, the results will be very disappointing. This is because no equalisation has been applied to the signal. It will be found that the

METHODOLOGY

Everyday Practical Electronics, September 2002

665

fidelity. To understand why this is necessary, you need to know something about the recording process. Imagine the sound has three bands comprising the low, intermediate and high frequency content. When the groove was cut in the master disc, the low frequency part was reduced in level (volume) while the high frequencies were increased. Only the intermediate band was left unchanged. Leaving the low frequencies as they were in the original sound would have required more violent movements of the groove cutter (that is, heavier modulation). This would have produced a wider groove and a consequent reduction in available playing time. Also, the playing stylus might have difficulty following such a groove and it may tend to jump out. By reducing the level of the low-frequency sound, it is possible to obtain a uniform groove width and a longer playing time. Equalisation is the process by which the high and low frequency content from the cartridge are restored to their original state and, in theory, should be an exact mirror of that used during recording. Note that by restoring the high frequencies, the surface noise present during playback (which is made up chiefly of high frequencies) is reduced. It thus provides a simple means of noise reduction.

with a cheap unit and without such a cut would be accentuated due to the low-frequency boost made during equalisation. Before proceeding to construct this circuit, check that you have a good quality record deck available. This must be fitted with a magnetic cartridge (not a ceramic one). If you wish to transfer 78 r.p.m. records, make sure your turntable will operate at this speed (many are designed for 33/45 only) also that it is fitted with the correct type of stylus. The full circuit diagram for the Vinyl To CD Preamplifier is shown in Fig.2. This is built around three identical dual low-noise operational amplifiers (op.amps) IC1a/IC1b, IC2a/IC2b and IC3a/IC3b. Equalisation of left and right channels is centred around IC1 and IC2 respectively, while IC3 is a straight amplifier which boosts both channels to line level. It is only necessary to describe the action of one channel (the left-hand one) since the other is the same. Note that the component numbering for the right-hand channel is prefixed with a one hundred. Thus, R2 (left) corresponds with R102 (right). Components which are common to both channels, the i.c.s, switches and input/output sockets are numbered as if they belonged to the left channel.

It is, therefore, the higher-frequency signals which develop a greater voltage at IC1a pin 3. In other words, the low frequencies tend to be filtered out. With the Rumble switch contacts closed, the pair of capacitors C1 and C2 give the same effect as a single unit having a larger value. This decreases the overall impedance and the circuit rolls off at a lower frequency.

CIRCUIT DETAILS

OPERATING CHARACTERISTICS

MAINTAINING STANDARDS

NON-INVERTING Unfortunately, different equalisation AMPLIFIER standards have existed regarding the valThe first section of the circuit is a nonues of the cut-off frequencies defining the inverting amplifier IC1a. The signal low, intermediate and high bands and also obtained from the input cartridge (leftthe degree of cut or boost. The same hand channel) at SK1 is applied to the noncircuit will therefore not provide perfect inverting input, pin 3, via capacitor C2 (or results with all records. C1 and C2 in parallel if Rumble switch However, most vinyl discs produced S1a contacts are closed). since the 60s have followed the RIAA (Recording Industry Association of America) standard. In practice, an equaliser designed for this standard will also provide good results when applied to recordings using a different one (American Standard Record and British Microgroove format). It should also be suitable for 78s. Practical equalisation circuits can range from the sim- Fig.1. Equalisation graph (not to scale): a) roll-off; b) fall-off ple (which provide and c) high frequency filtering. only a coarse correction) to the very This, in conjunction with fixed resistor complex. This circuit lies somewhere near R3, determine the anti-rumble characteristhe middle of the range and provides good tics of the circuit (the roll-off below 10Hz results without special adjustment. labelled A in Fig. 1). Resistor R3 also The graph shown in Fig.1 illustrates the sets the input impedance making it suitideal (theoretical) RIAA equalisation comable for a standard magnetic cartridge. pared with that provided by this circuit. Anti-rumble processing comes about Note that this is for illustration only and is because the impedance of capacitor C2 not drawn to scale. rises as the frequency falls. High frequenThe section to the left-hand side labelled cy signals will then flow more easily A provides a roll-off of frequencies through resistor R3 and hence through below some 10Hz. This reduces the rumcapacitor C3 (which has a relatively high ble that is transmitted from the motor or value and therefore negligible impedance turntable bearing to the cartridge through the at these frequencies) to 0V. turntable. This is much more pronounced
666

The output of IC1a at pin 1 is connected to its inverting input (pin 2) through the parallel arrangement of resistor R5 and capacitor C4. This works in conjunction with resistor R4 to set the gain. The other end of R4 is connected to the mid-point of a potential divider consisting of equal-value resistors R1 and R2. This sets a d.c. voltage nominally equal to onehalf that of the supply that is, 3V. This provides a zero reference so that the a.c. input signal will rise and fall with respect to it. If the reference was a true 0V (the voltage of the 0V supply line), the negative half-cycles of the wave would not be amplified. This is because the output voltage cannot fall below 0V. As it is, the output signal will swing above and below the 3V level. Ignoring the effect of capacitor C4 for the moment, the gain of this section is approximately eight times. However, with C4 in place, the impedance of the feedback loop will fall as the frequency rises. This reduces the gain at higher frequencies and provides the fall-off characteristic shown by Fig.1 section B. Section IC1b of the circuit is configured as a unity-gain amplifier (buffer). The signal from IC1a output, at pin 1, passes through resistor R7 (or R6 connected in parallel with it when Scratch switch contacts S2a are closed) to IC1bs non-inverting input at pin 5. High frequency signals now flow more easily through capacitor C5 (due to its reduced impedance) and hence to a further false zero derived from the potential divider made up of resistors R8 and R9. The voltage appearing at IC1b pin 5 will therefore be less than with higher frequencies. The higher frequencies therefore tend to be filtered out (shown by section C in Fig.1).

SCRATCH MY BACK

With Scratch switch S2a contacts closed, resistors R6 and R7 are placed in parallel and provide near-RIAA high-frequency attenuation. With the switch contacts open, resistor R7 alone provides a more dramatic cut-off and provides the scratch reduction effect. These values may be experimented with or a tone control could be fitted to give a continuously variable effect. More will be said about this later. The output from IC1b, pin 7, is now equalised but still at a low level. The next section, centred around IC3a, is an amplifier used in inverting mode. This boosts the signal by a large factor making it suitable to drive the line input of a sound card or external power amplifier. Capacitor C7 allows the output signal from IC1b pin 7 to pass with little loss (due to its relatively low impedance at

Everyday Practical Electronics, September 2002

VINYL TO CD PREAMPLIFIER

COMPONENTS

Resistors
R1, R101, R2, R102, R8, R108, R9, R109 2k2 (8 off) R3, R103, R4, R104 R10, R110, R11, R111 47k (8 off)

See

SHOP TALK
page

R5, R105, R6, R106 330k (4 off) R7, R107 120k (2 off) R12, R112 15k (2 off) R13, R113 1M5 (2 off) R14 2709 All resistors 06W 1% metal film.

Potentiometers
VR1, VR101 1M min. enclosed carbon preset, vert. (2 off).

Capacitors
C1, C101 C2, C102 C3, C103, C6, C106 C8, C108 C4, C104 C5, C105 C7, C107 C9, C109 C10, C110 C11 470n polyester film (2 off) 330n polyester film (2 off) 22 min. radial elect. 16V (6 off) 10n polyester film (2 off) 2n2 polyester film (2 off) 1 polyester film (2 off) 10pF ceramic (2 off) 10 min. radial elect. 16V (2 off) 220 min. radial elect. 16V 3mm red l.e.d. NE5532AN dual lownoise op.amp (3 off)

Semiconductors
D1 IC1 to IC3

Miscellaneous
S1 to S3 d.p.d.t. interlocking pushbutton switch see text (3 off) B1 6V battery pack (4 x AA alkaline cells) SK1 to SK4 phono socket, single hole, panel mounting (see text) (4 off) Printed circuit board available from the EPE PCB Service, code 366; 8-pin d.i.l. i.c. socket (3 off); aluminium instrument case, size 150mm x 100mm x 75mm; battery holder and connector; 3mm l.e.d. clip; screened cable; multistrand connecting wire; solder, etc.

Approx. Cost Guidance Only

excl. batts. & case

24

Fig.1. Complete circuit diagram for the Vinyl To CD Preamplifier.

Everyday Practical Electronics, September 2002

667

audio frequencies) through resistor R12 and hence to IC3a inverting input at pin 2. Ignoring capacitor C9 for the moment, fixed resistor R13 connected in series with preset potentiometer VR1 provides negative feedback and, in conjunction with R12, sets the gain. This will be some 170 times with VR1 set to its maximum value and 100 times at minimum. Preset VR1 will be adjusted at the end to provide a suitable output for the particular cartridge being used. The value of resistor R13 could be increased to provide a greater gain if this is shown to be necessary at the testing stage. By adjusting preset VR1 in conjunction with its opposite number in the other channel (VR101), the circuit will also be balanced to provide equal outputs for both channels.

Secure the p.c.b. and make sure the switches operate freely. Attach the battery holder and the input and output sockets. If these are of the specified type, you will need to scrape away the paint on the inside surface of the box to allow the outer (sleeve) connections to make good metallic contact with the case. Attach one of the solder tags supplied with the sockets under the fixing nut of one of them. This will be used to earth the 0V wire leading from the circuit board.

If you are using sockets of the fully-insulated type rather than the specified pattern, the sleeve connection of each must be connected to the case (0V) using a separate solder tag. Referring to Fig.4 and photographs, complete the internal wiring. Take care that left and right inputs and outputs maintain their identity during the wiring process (that is, they do not become interchanged). Set all switches to the out position, insert the batteries and attach the lid of the case.

PROMOTING STABILITY

Returning to capacitor C9 which appears in IC3a feedback loop, its small value provides an extremely high impedance at audio frequencies. It therefore normally has negligible effect. However, if radio-frequency signals happen to be picked up by the circuit, the impedance of C9 will be low. This will lower the impedance of the feedback loop and reduce the gain at these frequencies. This prevents instability. The output signal finally passes from IC3a pin 1, via capacitor C10, to Line Output socket, SK3 (Left channel).

CONSTRUCTION

Construction of the Vinyl To CD Preamplifier is based on a single-sided printed circuit board. This board is available from the EPE PCB Service, code 366. The topside component layout and actual size underside copper foil master pattern are shown in Fig.3. Commence construction by drilling the three mounting holes as indicated. Solder the spring-loaded, pushbutton switches in position. If the specified type is not available, use toggle or slide units and hard-wire these to the appropriate points on the p.c.b. at the end of construction. Next, solder in position the three i.c. sockets. Follow with all resistors, preset potentiometers and capacitors taking particular care over the polarity of the electrolytics. Solder the battery connector to the +6V and 0V points on the p.c.b., again, taking care over their polarity. Adjust presets VR1 and VR101 to approximately mid-track position to provide a medium degree of gain for each channel.

BOXING UP

Note that this circuit must be housed in a METAL box to provide adequate screening against hum pick-up. Decide on a suitable layout for the internal components. Measure the positions of the switches and l.e.d. on the p.c.b. Mark these on the front panel of the box at the half-height level and drill them through. Mark and drill the p.c.b. mounting holes also those for the battery holder and the input and output sockets. Cut plastic stand-off insulators to the correct length so that, when the p.c.b. is in position, the switch buttons will pass through their holes with a little clearance.

Fig.3. Printed circuit board component layout and full-size underside copper track master pattern for the Vinyl To CD Preamplifier.

668

Everyday Practical Electronics, September 2002

TESTING

Unless the stylus on the record deck is known to have given very little service, renew it. Styli cost very little compared with that of your record collection. Also, a new stylus will give better results. If you are going to transfer 78s you must have the correct stylus fitted do not use one made for 33s/45s. It would be useful to have the turntable manual available to help make optimum stylus pressure and anti-skid adjustments. Sometimes a slightly greater pressure than normal will give better results. Although this wears the record more quickly it may be worthwhile since the record need only be played once. For initial testing, connect the output of the preamplifier to the line input of a hi-fi amplifier using twin-screened cable fitted with the appropriate connectors. Do not connect it to the computer sound card at this stage. Connect the turntable to the preamplifier input sockets. If possible, use a valueless record to make initial tests. Turn the Volume control on the amplifier to minimum and switch on both units. Check that the front panel l.e.d. operates. It may be found convenient to use headphones to monitor the sound. Start playing the record and gradually increase the amplifiers volume control. The music should be clearly heard. Compare the volume with that playing similar music from a commercial CD. If the levels are not similar, adjust VR1 and VR101 so that they are. If one channel is quieter than the other, adjust presets VR1 or VR101 as appropriate to bring the weaker channel to the level of the stronger one. This procedure ensures that the output is at line level and balanced between the channels. Check the effects of the Scratch and Rumble switches. The rumble effect is very subtle and may not be noticed. Note that, as described, pressing the switches in provides the anti-scratch and anti-rumble effects.

Layout of components inside the metal case.

SUBJECT FOR EXPERIMENT

The frequency balance and anti-scratch effects could be altered by changing the value of resistors R6/R7 and R106/R107. By increasing the appropriate resistor values slightly, the high-frequency response will be cut and vice versa. Beware small changes make a lot of difference! An alternative method would be to replace resistors R7/R107 with a dualganged, panel-mounted, potentiometer (stereo). This would allow for continuous variation and switch S2 could then be ignored.

Fig.4. Interwiring details from the printed circuit board to the rear panel mounted input and output phono sockets.
Before making a recording, clean the surface of the disc using a proprietary antistatic cleaner. If it is very dirty, it will need special treatment to remove the debris which will have become deeply embedded in the groove. You could try playing it once or twice in an attempt to allow the stylus itself to remove the contamination.

MAKING A RECORDING

MAKING TRACKS

When setting up the equipment to make CDs, the turntable should not be placed on the same surface as the computer (otherwise you could introduce hum due to vibration being transferred to the cartridge from the computer). Check that the turntable is true using a spirit level. Connect the preamplifier output to the line input of the PC sound card using twinscreened wire. Check that Left and Right channels are connected correctly.

STYLUS CHECKS

Check the stylus after every playing for any build-up of fluff and dirt. Leaving this will spoil the high-frequency response and also tend to cause the stylus to jump out of the groove. Use a proprietary stylus cleaning kit (a fine brush and cleaning fluid). Styluses are easily damaged so follow the instructions and work carefully.

Refer to your CD recording software instructions to make optimum sound level settings and make some tests using the old record. For your final recordings, you will probably be able to observe the file oscilloscope-style. It is then possible to remove the heaviest clicks by highlighting and deleting them. However, this must be done with great care. Some CD recording software allows for sophisticated restoration work to be undertaken. Automatic click suppression can be a problem because many sections of the intended waveform are click-like. One final point do not use the scratch filter unless the result sounds better. This is because it gives a markedly dull effect. 6

Everyday Practical Electronics, September 2002

669

READOUT
E-mail: editorial@epemag.wimborne.co.uk

WIN A DIGITAL MULTIMETER


A 31/2 digit pocket-sized l.c.d. multimeter which measures a.c. and d.c. voltage, d.c. current and resistance. It can also test diodes and bipolar transistors. Every month we will give a Digital Multimeter to the author of the best Readout letter.

John Becker addresses some of the general points readers have raised. Have you anything interesting to say? Drop us a line!

0 LETTER OF THE MONTH 0


SHOCK HORROR TALE!
Dear EPE, I was re-reading some old EPE issues while waiting for the latest to turn up here in New Zealand (I dont suppose you could print EPE every week, could you?), and something Alan Winstanley wrote in Circuit Surgery of Sept 00 made me laugh out loud. I hasten to say I have the greatest respect for Alans intellect which shines through everything he does, but I was reminded that there is sometimes a second, more amusing explanation for a set of symptoms. A reader had queried Alan about worrying electric shocks from his dishwasher, and yet his RCD (residual current device) had not tripped the power off, and the RCD checked out OK. Alan theorized a possible insulation fault but gave the excellent advice to get the dishwasher looked at by a professional. The following story from my time as an electronic repairman shows how a working RCD might not trip even though the machine it is attached to is giving you electric shocks. Some years ago I quickly attended a similar fault in an old, all-metal franking machine (stamps postage on envelopes) which had been relocated in an old office building and, while it was running well, had been giving electric shocks to everybody since the relocation even when it was switched off. I believe the NZ power distribution system is the same as UK, 230V a.c., 50Hz, multiple earthed neutral, so on the way to the fault I was mentally going over things like earth wire broken off in the old machine, wiring faults and errors in the building, etc. The ladies who used the machine were in some fear of their lives, and I had firmly advised them over the phone that this fear was well-grounded (is that a pun?). When I arrived onsite they were at first rather put out when they saw me dash in, wave my meter about the machine, glance around the room and burst out laughing. What I really did was to check the machine competently, and drew the conclusion that when they walked over the nice new carpet in their nice refurbished office in their pretty feminine artificial leather shoes to the machine, all those thousands of volts of static electricity they had built up found a ready path to ground through the well-earthed machine! And that was it. After my careful and sympathetic explanation to the ladies about how to minimise static buildup, and how it wasnt endangering their lives anyway, they ruefully saw the reason for my amusement, but still didnt want to touch the machine. In the end I suggested they leave the office scissors (metal) near the machine, they could pick up the scissors and, holding them firmly, touch the machine with the scissors first, thereby discharging themselves with a mighty crack! of spark and not feel a thing. And that would be one way that EPE readers could get even severely-felt electric shocks from a machine and yet the machines fully operational RCD wouldnt trip. Having said that, I very strongly advise people not to assume that electric shocks from machinery are just harmless static. Get it checked or plan your funeral, electricity is a good servant but a bad master! Stan Hood, Christchurch, New Zealand Reminds me of a situation in my late school years. While showering in the sports changing room, I frequently felt tingles in my hand when lightly coming into contact with the metal shower tap. For weeks the school authorities would not believe me when I said that the tingling was due to electricity being present on the water piping. Eventually the Electricity Board was called in yes indeed, there was an electrical problem, affecting the adequate earthing of that part of the building. A lot of digging in the road outside was required before the fault was found and cured! I would not be telling the tale had the current flow been more severe.

8051 FREEWARE
Dear EPE, I know that most of your projects that use microcontrollers are based around PIC devices, but I just want to let any of your readers who use the 8051 microcontroller, or its many derivatives, know about a very good freeware open source ANSI compliant optimising C compiler which I have been using for a few months, now called SDCC. Its available for download from sdcc.sourceforge.net. There are several discussion forums for its users also on the same site. It can also be targeted at Z80, Gameboy Z80, AVR and PIC14x microcontrollers, and comes with a freeware 8051 software simulator. Keep up the good work on your magazine, I have been a reader since I was a schoolboy hobbyist. Jez Smith, by email Thanks Jez, undoubtedly we have some readers who are 8051 users as well as PIC addicts. And thanks too for your continued interest in EPE!

BASIC STAMP
I have taught myself PICBasic and have a great interest in microcontrollers. What I would like to know is what industries use Basic Controllers and is it hard to start a career using and programming them? Any advice would be greatly appreciated. Alex, via email I suspect that in general industry does not use PICBasic types of program, preferring the more universally used assembler codings in various forms. Readers what are your opinions?

SMOKE DETECTION
Dear EPE, I am from Les Quennevais school in Jersey. For my business GCSE project I am going to make a photoelectric smoke detector, carbon monoxide detector and heat detector for the deaf. I am wondering if you could send me some circuit diagrams or tell me your suppliers as it would largely help me in my project. Any information that you could give would be very helpful Alan Morris, via email Our Teach In 2002 series looked at smoke detection in the June 02 issue, back issues can be ordered via our Online site, or according to the information published in each EPE issue. We have not done other smoke detectors in recent years.

PIC ALARM
Dear EPE, Ive been building your PIC Controlled Intruder Alarm (Apr 02) great application! It seems, though, that you can only arm the alarm when the entry zone is set-up to be normallyopen, is this so? In your article you suggest feedback would be welcome on the use of the RB4 interrupt for the panic switch. I have linked pins of the S3 connector but can still trigger the panic event by generating mains noise, even pulling the plug out and switching to battery power sometimes generates the event. Im planning to add mains suppression etc. Mark Jones, via email Feedback is always welcome Mark, thanks. The entry zone restriction was not intentional,

but in practice I have never encountered a situation where entering the main door zone could require a choice between normally-open and normally-closed contacts.

HOME SECURITY
Dear EPE, I am currently doing my final year project on a home security system which involves a 4 3 matrix keypad, PIR sensor, magnetic switch and glass break detector. Im using a PIC16F84 and PICBasic to write the software. Can you please give me some advice? Brendon, Malaysia, by email Sorry to disappoint you Brendon, but we cannot give specific advice for readers own designs, but you might find my PIC Controlled Intruder Alarm of April 02 of interest. That uses a matrixed keypad.

STYLOPIC OP.AMP
Dear EPE I am having problems finding the LM13600 transconductance op.amp for the StyloPIC of July 2002, the RS 304-453 is now listed as no longer stocked. Do you know what other device could be used as an alternative please? Mike Mackellow, via email You can use the LM13700 instead as a direct replacement no mods needed.

Everyday Practical Electronics, September 2002

671

STYLOPIC
Dear EPE, Following on from your StyloPIC in July 02, you might be interested in some info on the original. There were three variations of the pocket model standard, treble and bass. The treble and bass models being respectively an octave higher or lower (mine is the standard model). Its big brother, the 350S, had many extra features such as short or long envelope, staccato, two speed vibrato, wah-wah, and eight voices. An innovative feature is a light sensor (l.d.r.) for hand control of vibrato or wah-wah. It also has two styluses (for playing chopsticks?). An external amplifier was also available for either instrument, with tone and tremolo controls. On the technical side, the circuit diagram for the pocket version is in the back of the instruction book. Tone generation is by a programmable unijunction transistor so the waveform would be pulsed, however it is modified by what looks like a diode pump monostable so the mark-space ratio would vary depending on the note frequency (and presumably the harmonics generated). So the output waveform would be something like a square wave with slow rise and fall times. Vibrato is generated by a low frequency phase shift oscillator to vary the programming voltage of the unijunction transistor. I know John Becker likes to recycle his software so here is something to consider in a future incarnation. It gives greater flexibility of the output waveform. And, of course, you can have multiple waveform tables. This is only an example, other changes may be needed for it to work correctly. OUTIT: call WAVFORM movwf PORTA goto MAIN WAVFORM: andlw $7F ; Sinewave + 2nd harmonic movwf PCL; 128 entries, amplitude 0 to 63 DT 00,00,00,00,00,01,02,04,06,08,11,13,16,19,22,26 DT 29,32,35,39,42,44,47,49,52,54,55,57,58,59,59,59 DT 60,60,60,60,60,59,59,59,58,59,59,57,57,57,57,57 DT 58,58,59,59,60,60,61,61,62,62,63,63,63,63,63,63 DT 62,61,60,59,58,56,55,53,51,49,47,45,43,41,39,37 DT 35,33,32,31,30,29,29,28,28,29,29,30,31,32,33,34 DT 35,36,38,39,40,41,42,43,43,43,43,42,42,41,39,38 DT 36,34,32,29,27,24,21,19,16,13,11,08,06,04,03,01 ; (DT is Define Table of retlws in MPASM) Peter Hemsley, via email Thanks Peter. The technical stuff I did not find on the web. The table concept looks interesting. I dont know that Ill ever upgrade StyloPIC but who knows?!

where to order from abroad. The TENS replacement electrode pads you specify are easy to find at almost any Boots shop, but the staff there know nothing about leads, nor how to order. Ive tried at several Boots shops on my last trip to London. Cristian, via email Mine came from Boots in Wimborne. I dont know the order code, they were being supplied as normal stock items. If you cant get any, use flexible wire with crocodile clips to clip onto the chest pads. They dont need screening. You could try asking Boots HQ via email (www.google.com will provide a web address).

The only changes required are to the sweep2 and sweep3 routines: sweep2 bsf output bcf output2 decfsz freq,f goto sweep2 movfw nfreq movwf freq sweep3 bcf output bsf output2 decfsz freq,f goto sweep3 ; output high ; output2 low added

; output low ; output2 high added

SERIAL ADC PIC TRICK


Dear EPE, Readers might be interested in my PIC program for use with the TLC548/9 8-bit serial analogue-to-digital converter. I use file registers COUNT and TEMP as sort of standard registers, COUNT for timing etc and TEMP as a sort of second W. It helps me get a mental view of my progs. In the program this routine comes from, COUNT has previously been reset through DECFSZ, so I can get away with BSF COUNT,3. I have run this at 6MHz without problem, and it should go faster. The A-D value is stored in file UNIT. A2DIN: BCF PORTB,7 ; clear CS line to hold value to send BSF COUNT,3 ; set count to shift 8 bits (make sure that COUNT cleared before this section or use MOVLW etc) FETCH: RLF UNIT,F ; move bits one place left & store new value in UNIT BCF UNIT,0 ; set 0 value before Portb,0 bit test BSF PORTB,6 ; set A2D clock pin high, release bit for transfer BTFSS PORTB,5 ; is bit 0 (DOUT) set ? GOTO NEXT1 ; no, then leave UNIT bit 0 value as is BSF UNIT,0 ; yes, set bit 0 of UNIT NEXT1: BCF PORTB,6 ; clear clock pin DECFSZ COUNT,F ; is COUNT zero? GOTO FETCH ; no, get another bit! BSF PORTB,7 ; yes, 8 bits clocked out & held in UNIT, set CS line to get new value RETURN Graham Card, via email Useful, Graham, thank you Ive put it in the PIC Tricks folder on our ftp site.

Plus an extra define line: #define output2 gpio,5 ; inverted o/p to piezo sounder Nigel Goodwin, via email Thanks Nigel!

LOTTERY PREDICTOR
Dear EPE, I am studying GCSE Electronics. My father has been purchasing EPE since 1994 and is still enjoying each new edition. In the April 95 issue I came across the National Lottery Predictor project and am wondering if you could please send me as much information on that topic as possible to further my knowledge and passion. Gopyr, via email So sorry, but we cannot provide additional material for any published design. Regarding building a circuit from 1995, we normally advise against attempting to build a design that is over five years old since parts could well have become obsolete during that time. In this particular case, the p.c.b. is no longer available, nor will you be able to obtain the programmed PIC as we are no longer in touch with the authors, and they did not sell us the copyright to their software (that was before we began to insist that all project software must be made freely available to readers).

EARTH RESISTIVITY LOGGER


I am designing an Earth Resistivity Logger for archeological use, inspired by Robert Becks Earth Resistivity Meter of Jan/Feb 97. Mine is PIC controlled and will have its own non-volatile memory (data stays held even after switch off); possibly a graphics l.c.d. may show rough details of reading values as grey scale; serial interface for connection to PC for deeper analysis. I am not an archeologist and am approaching the design purely as an electronic problem to be solved send an output signal, retrieve it from a distance and store the value. I am in communication with a local archeological society, but I would be pleased to hear from any EPE readers involved in this field, with special regard to the following: * How many reading samples do you normally take on a site in one main session? * How many samples would you like the logger to store before download to PC? * Is powering it from a 12V car battery adequate, or do I need 18V as Robert had? * What probing techniques do you use? Im assuming the twin-probe technique is best, as described by Robert. * What maximum probe separation distance do you use? * How deep do you insert the probes? * Is a signal frequency of 137Hz as used by Robert the best to use? * In your experience, how likely is it that 50Hz mains frequency is likely to occur on a site being surveyed, and would thus need to be filtered out in some way? * Do you always plot the site squares in the same regular order, or would you prefer to sample in random order, telling the logger the square number being sampled? Any answers would be appreciated, my email is john.becker@wimborne.co.uk.

FLOW CHARTS
Dear EPE, PICs are not my strong point! However, Ive started to look at the code for your PIC Controlled Intruder Alarm (Apr 02) with a view to modifying it to suit my own purposes. Do you have a flow chart that you could send me? Trevor Brearley, via email No, sorry Trevor, I dont do flow charts for my software I keep concepts in my head and work to those! Readers who do like to work with flow charts will probably be interested in the Flow Code for PICmicro CD-ROM thats available via our CD-ROM pages in this issue, and in Terry de Vaux Balbirnies review of it, also in this issue.

FREEZER ALARM
Dear EPE, Ive been reading Humphrey Berridges Freezer Alarm in the May 02 issue, and Im extremely impressed with the low component count for the functionality achieved, but Id like to make a suggestion: The piezo sounder needs to be as loud as possible, but its only being fed with 5V pk-pk from pin GP4 to ground. If you connect the sounder between GP4 and GP5, and feed GP5 with an inverted signal, you will get 10V p-p drive in a bridge configuration twice the voltage, at no extra cost!

BIOPIC LEADS
Dear EPE, I am building the BioPIC Heartbeat Monitor (Jun 02) and need to know the order code for Boots lead pack, together with the information

672

Everyday Practical Electronics, September 2002

PRACTICALLY SPEAKING
Robert Penfold looks at the Techniques of Actually Doing It!
about finished projects Cfailing to for would-beprobably the work are main reason constructors
ONCERNS

failing to take the plunge. It is not a major concern for those with years of project building experience because they have the technical knowledge, equipment, and know-how to deal with practically any problem. The opposite is true for beginners who, on the face of it, have little chance of dealing with projects that refuse to work.

Keep it Simple
In reality the situation for beginners is better than it might seem. Provided you start with something reasonably simple and follow the instructions carefully there is a good chance of success. Prepublication checking for both books and articles containing electronic projects has increased over the years, and this has greatly reduced the chances of being led astray by printing errors. On the rare occasions that an error does creep in to an EPE article it is usually spotted quite early and corrected one or two issues later. In general, the complexity of modern projects is greater, but your chances of failure if the instructions are followed to the letter are much less than they were. Like any creative skill, electronic project construction would not be a worthwhile hobby if perfect results were guaranteed every time with no skills required. You have to be prepared to put in some effort and try to go about things the right way. It is worth repeating the importance of choosing a project that is within your capabilities. It is tempting to dive straight in with a project that will impress your friends, but the more complex the project the greater the risk that you will make a mistake. In the past it was not unusual to receive letters from readers having problems with projects that they clearly did not understand at all. You do not need to know how a project works in order to build it successfully, but you do need to have a proper understanding of what it is supposed to do and how it is used. Something like a household gadget is a more appropriate starting point than an advanced piece of test equipment where you need a degree in physics in order to switch it on! Fortunately, letters from readers who have bitten off more than they can chew are relatively rare these days, but it is still a problem to take seriously.

simple, if it connects to the mains supply it is certainly not suitable for a beginner. Start with projects that are battery powered. If you should make a serious blunder it is possible that one or two of the components will be damaged, but you should be perfectly safe. In most cases all the components will survive the experience as well. The two main construction methods used in modern projects are stripboard and custom printed circuit boards (p.c.b.s). While both types of board are pretty straightforward to use, custom printed circuit boards represent the more foolproof option. Stripboard is a multi-purpose circuit board that has a regular matrix of holes, and in most projects only a few percent of these are actually used. As its name suggests, a custom printed circuit board is specifically designed for a particular circuit and normally has just one hole per leadout wire or pin. With a custom board there is relatively little risk of making a mistake in the first place, and any errors that should creep in are likely to be spotted almost immediately. With stripboard there are hundreds of unused holes that are good at disguising mistakes, and some very careful checking is needed to detect them.

board, and in most cases they are easily removed using the bit of the soldering iron. If there is a lot of excess solder it is better to use a desoldering tool, and an inexpensive desoldering pump is ideal for this application. It is advisable to remove as much solder as possible and then redo any joints that have been desoldered.

Hidden from View


The more difficult problem is minute trails of solder that are often difficult or impossible to see with the naked eye. The situation can be made more difficult by the trails being hidden under excess flux from the solder. This tends to get liberally splattered across the underside of circuit boards during construction. There are various products that can be used to thoroughly clean the flux from boards, but vigorous brushing with a small brush such as an old toothbrush seems to do the job well enough.

Bridging the Gap


Having chosen a suitable project and put it together with due diligence, what do you do if the finished unit fails to work? When a newly constructed project is clearly failing to work properly it is not a good idea to leave it switched on. Leaving a faulty project switched on could result in damage to some of the components, and the semiconductors are particularly vulnerable. Always switch off faulty projects immediately and then recheck the component layout, wiring, etc. The prudent project builder checks all this sort of thing very carefully during construction, and spotting errors early can save a lot of hassle latter. In order to properly check the unit you may have to partially dismantle it in order to get proper access to the circuit board. Years of practical experience suggest that the vast majority of problems are due to short-circuits between copper tracks on the underside of the circuit board. This is not exactly a new problem, but the intricacies of modern boards make it even more problematic than in the past. Unless the board is coated with a solder resist that is designed to discourage solder bridges, it is likely that several will be produced per circuit board. Most of these bridges will be spotted while you are constructing the

A dry joint. Solder failed to flow.

A good joint, nice and shiny.

Photos courtesy Alan Winstanleys Basic Soldering Guide

Good eyesight is not sufficient to guarantee that any solder bridges will be spotted. Some form of magnifier now has to be considered part of the standard toolkit for electronic project construction, and even a small magnifying glass will greatly increase the chances of detection. An 8x or 10x loupe (also sold as lupes) is better though. The inexpensive types sold as photographic accessories for viewing slides and negatives are perfectly adequate for the present application. Provided the board is thoroughly cleaned first, a careful visual check using a magnifier should reveal any solder bridges. As solder bridges occur so often it is a good idea to clean and visually inspect all completed circuit boards prior to installing them in the case.

Hot Spots
Dubious soldering is a common cause of problems, particularly amongst beginners. Soldering is like any skill, and it is a case of practice makes perfect. The more projects you build the more proficient you will become at completing soldered connections. There is insufficient space here for a soldering tutorial, but a

Mains Point
The mains supply is potentially lethal, as are projects that connect to it. Mains power projects are only suitable for those with a reasonable amount of experience at project construction. Even if a project is very

674

Everyday Practical Electronics, September 2002

good one is available at the EPE web site. Some soldering irons and soldering kits are supplied with detailed instructions, and it is well worthwhile studying these. Probably the most common cause of so-called dry joints is the soldering iron being left unused for a few minutes before starting a new batch of connections. If there is a substantial amount of solder left on the bit, any flux in it will burn away and it will probably start to oxidise. If you produce the next joint without cleaning the end of the bit first, the joint will contain a significant proportion of old solder, which may not flow over the joint properly. The resultant joint might look plausible and could seem to have good mechanical strength as well. However, joints of this type usually provide only intermittent electrical contact or no contact at all, and are relatively weak mechanically.

Fig.1. The wide band indicates the cathode (k) leadout of multi-band diodes.

Heat of the Moment


Apart from semiconductors, modern components are reasonably tolerant of heat. However, it is still possible that damage will occur if you take too long to complete joints. Heat damaged components usually show some obvious signs of damage, such as a darkening in colour or being slightly misshapen. Always replace any off colour or deformed components, or any components that show significant signs of physical damage. Integrated circuits (i.c.s) are mostly fitted in holders, but transistors and diodes are often connected directly to the circuit board. Always take extra care when fitting these in place. As pointed out previously, it is a matter of practice makes perfect, and you can avoid a lot of problems by learning to solder quickly and neatly before dealing with transistors and diodes.

the markings on components such as diodes and electrolytic capacitors are usually quite explicit, so any errors should be easily spotted. One exception is the type of diode that has several bands rather than one at the cathode (k or +) end of the component. These have had something of a renaissance in recent times, so you may well encounter them. The bands indicate the type number using a variation on the resistor colour code. A wider band at that end of the body (Fig.1) indicates the cathode (k) lead. Light emitting diodes (l.e.d.s) can also be problematic. If a project works apart from a l.e.d. indicator, it is oddson that the l.e.d. is simply connected the wrong way round.

A Pressing Connection
Before too long practically everyone makes the classic mistake of forgetting to switch on the project or omitting that all-important component the battery. Battery connectors have always been notoriously unreliable. Try pressing the connector firmly onto the battery to see if it makes the project burst into action. Slightly compressing the female connectors with pliers usually gets a loose clip to work reliably. Battery holders for 15V cells are also something less than totally reliable. Ensure that the terminals of the batteries and the holder are clean by gently removing any contamination with fine sandpaper.

Shining Example
Always make sure that the bit is tinned with fresh solder prior to making joints. Practice soldering with some bits of wire, a few resistors, and a scrap of stripboard before you start building projects. This will cost very little and will greatly enhance your chances of success. Checks with a continuity tester or the continuity function of a multimeter should locate dry joints, but thoroughly checking even a small circuit board can be quite time consuming. Large amounts of excess flux are sometimes indicative of a bad joint, but this is of no help once the board has been cleaned. Good joints normally have a characteristic mountain shape and the surface of the solder is very shiny. Dry joints are often more spherical in shape and the solder tends to have a relatively dull surface, possibly with some crazing.

Multi-checks
A cheap multimeter is useful for checking that the battery voltage is actually getting through to the circuit board. It can also be used to check that the battery is in a usable state. Even if you do not have much technical knowledge, a multimeter can still be useful for numerous basic checks. For example, it can be used for making continuity checks on switches, which may not operate in quite the way you think they do? Have you confused the on and off settings? Often when a project seems to be working irrationally it is just that one of the switches does not function as expected. The high and low ranges are transposed, or something of this type. A multimeter is also useful for checking cables for short-circuits or broken leads, checking that that plugs and sockets connect together properly, etc. Even some of the cheaper digital types now have the ability to check resistors, transistors, diodes, and capacitors, which is clearly more than a little useful. A multimeter is a piece of equipment that no project builder should be without. Because modern components are very reliable you are unlikely to have a failure caused by a dud component. If you get everything connected together properly your projects will work, and it helps to keep this in mind. Of course, the projects will never work if you do not pluck up the courage to take the plunge and actually build them.

Try and Try Again


Having thoroughly checked both sides of the board and made any necessary repairs it is time to reassemble the project and test it again. Thoroughly check the hard wiring against the wiring diagram, as it is relatively easy to make mistakes here. If the project still does not work, the most likely explanation is that you have missed an error in the wiring or on the circuit board. With this type of thing there is a tendency to blame others and not accept that you could have made a mistake. In reality it is easy to make the odd mistake here and there, and even old hands make the occasional error. Start by checking that every component on the circuit board is in the right place and has the correct value. Work through the components methodically making sure that none of them are overlooked. If you have managed to miss out a component, this error should then come to light. With stripboard construction make sure that any link wires are present and correct. Ideally you should get someone else to check the unit against the construction diagrams. A fresh pair of eyes might spot something that you have consistently overlooked.

Clean Break
If any joints look suspicious it is probably worthwhile desoldering them and then re-soldering them. Before trying again it is a good idea to have a close look at the two surfaces. These days it is unusual for dirt or corrosion on one of the surfaces to cause problems. Modern components are less vulnerable to corrosion on the leadout wires and tags, and the flux in electrical solders is very efficient at dealing with contaminants. However, there can still be occasional problems though, and if there is any sign of contamination it is a good idea to clean both surfaces before redoing the joint. The best way to clean the surfaces is to gently scrape them with the small blade or a penknife, a miniature file, or something of this type. The driest joint of all is the one you forget to do! Missing joints are usually fairly obvious with custom printed circuit boards, but can be difficult to see with stripboard where there are numerous unused holes and no pads as such. Firmly pulling on resistors, capacitors, diodes, etc., will reveal any missing joints, or ineffective joints that look plausible.

Wrong Connection
The components that must be fitted the right way round are the most likely to give problems. Layout diagrams and

Everyday Practical Electronics, September 2002

675