Basics of Electricity and Electronics with Projects: Linear Electronic Circuits

Home | Articles | Forum | Glossary | Books

In a general sense, the entire massive field of electronics can be classified into two very broad categories: digital and linear. Digital pertains to those circuits and devices that operate on the basis of switching action, representing numbers or data by means of on-off pulses. The fundamentals of digital electronics will be covered in later sections.

In this section, you will examine the fundamentals of linear circuits. The term linear pertains to circuits that operate in a proportional manner, accepting inputs and providing outputs that are continuously variable (i.e., analog).

Although the field of linear electronics is very diversified, the basics of linear action are common to almost all of its facets. For example, the same techniques used to linearize (i.e., increase proportional accuracy in) an audio amplifier are used to linearize servo systems and operational amplifiers. An accurate understanding of the basic building blocks utilized in linear systems will aid you in understanding a great variety of electronic systems.

Because of a variety of factors, "discrete" (i.e., nonintegrated) circuitry is still utilized in a significant portion of the field of audio electronics. Therefore, this section focuses primarily on audio amplification circuits, since they provide a good beginning point to study the fundamentals of linear circuitry. In addition, the associative discussions make it a convenient point to detail printed circuit board construction (in conjunction with a few more "advanced" projects) and the newer computer-automated (also -aided or -assisted) design (CAD) techniques for designing and constructing PC boards.

Transistor Biasing and Load Considerations

FIG. 1 Basic audio amplifier circuit.

The circuit illustrated in FIG. 1 should already be familiar to you from the previous discussions of transistor amplifiers. It is a common-emitter amplifier because the output (to the speaker) is taken off from the collector, and the input signal to be amplified is coupled to the base. C1 is a coupling capacitor (blocking the DC bias voltage, but passing the AC audio signal). R1 and R2 form a voltage-divider network to apply the proper DC bias to the base. The emitter resistor (RE) increases the input impedance, and it improves temperature and voltage stability.

Transformer T1 is an audio transformer. It serves two important functions in this circuit. First, it isolates the DC quiescent (steady-state) cur rent flow from the speaker coil (speaker coils can be damaged by even relatively small DC currents). Secondly, it provides a more appropriate load impedance for a transistor collector than would a low-impedance 8-ohm speaker. A transistor amplifier of this configuration could not operate very well with an extremely low collector impedance. A typical audio transformer might have a primary impedance of 100 ohms, for connection into the transistor circuit, and a secondary impedance of 8 ohms for connection to the speaker. Generally speaking, an audio amplifier of this type performs satisfactorily for low-power applications. How ever, based on the basis of modern standards, it has severe problems and limitations.

To begin, examine the real-life problems relating to efficiency and biasing considerations. Choosing some simple numbers for discussion purposes, assume the source voltage in FIG. 1 to be 35 volts; T1's primary impedance, 100 ohms; and RE, 10 ohms. As you might recall, the volt age gain of this circuit is approximately equal to the collector resistor (or impedance) divided by the emitter resistor. Therefore, the 100-ohm collector impedance (T1) divided by the 10-ohm emitter resistor (RE) places the voltage gain (Ae ) at 10.

R1 and R2 are chosen so that the base voltage is about 2.1 volts. If Q1 drops about 0.6 volt across the base-emitter junction, this leaves 1.5 volts across RE. The 1.5-volt drop across the 10-ohm emitter resistor (RE) indicates the emitter current is at 150 milliamps. Because the collector current is approximately equal to the emitter current, the collector current is also about 150 milliamps. (The 150 milliamps is the "quiescent" collector-emitter current flow. The term quiescent refers to a steady-state voltage or current established by a bias.) Now, 150 milliamps of current flow through the 100 ohm T1 primary causes it to drop 15 volts. If 15 volts is being dropped across T1's primary, the collector voltage must be 20 volts (in reference to ground). Then 15 volts plus 20 volts adds up to the source voltage of 35 volts.

If a 500-millivolt rms signal voltage were applied to the input of this amplifier, a 5-volt rms voltage would be applied to the primary of T1 (Ae = 10). If T1 happened to be a "perfect" transformer, it would transfer the total AC power of the primary to the secondary load. In this example, the total power being supplied to the primary is 250 mW rms. Even with no T1 losses, 250 mW of power would not produce a very loud sound out of the speaker.

In contrast, examine the power being dissipated by Q1. As stated earlier, in its quiescent state, the collector voltage of this circuit is 20 volts.

The emitter resistor is dropping 1.5 volts; therefore, 18.5 volts is being dropped across the transistor (the emitter-to-collector voltage). With a 150-milliamp collector-emitter current flow, that comes out to 2.775 watts of power dissipation by Q1. In other words, about 2.775 watts of power is being wasted (in the form of heat) to supply 250 mW of power to the speaker. That translates to an efficiency of about 9%.

With a better choice of component values, and a more optimum bias setting, the efficiency of this amplifier design could be improved. How ever, about the best real-life efficiency that can be hoped for is about 25% at maximum output, in this class A amplifier.

There are actually two purposes to this efficiency discussion. The first, of course, is to demonstrate why a simple common-emitter amplifier makes a poor high-power amplifier. Second, this is a refresher course in transistor amplifier basics. If you had some trouble understanding the circuit description, you might want to review Section 6 before proceeding.

Amplifier Classes

Although some audio purists still insist on wasting enormous quantities of power to obtain the high linearity characteristics of class A amplifiers, such as the one shown in FIG. 1, most people who specialize in audio electronics recognize the impracticability of such circuits. For this reason, audio power amplifiers have been developed, using different modes of operation, that are much more efficient. These differing operational techniques are arranged into general groups, or "classes." The class categorization is based on the way the output "drivers" [transistors, field effect transistors (FETs), or vacuum tubes] are biased.

The amplifier circuit illustrated in FIG. 1 is a class A audio amplifier because the output driver (Q1) is biased to amplify the full, peak-to-peak audio signal. This is also referred to as biasing in the linear mode.

Again referring to FIG. 1, assume that the bias to Q1 were modified to provide only 0.6 volt to the base. Assuming that Q1 will drop about 0.6 volt across the base-emitter junction, this leaves zero voltage across RE.

In other words, Q1 is biased just below the point of conduction. In this quiescent state, Q1 would not dissipate any significant power (a little power would be dissipated because of leakage current) because there essentially is no current flow through it. If an audio signal voltage were applied to the input of the circuit in this bias condition, the positive half-cycles would be amplified (because the positive voltage cycles from the audio signal on the base would "push" Q1 into the conductive region, above 0.6 volt), but the negative half-cycles would only drive Q1 further into the cutoff region and would not be amplified. Naturally, this results in severe distortion of the original audio signal, but the efficiency of the circuit, in reference to transferring power to the speaker, would be greatly improved. This mode of amplification is referred to as class B, where conduction occurs for about 50% of the cycle.

Of course, the circuit shown in FIG. 1 (biased for class B operation) is not very practical for amplifying audio signals because of the high distortion that occurs at the output. But if a second transistor were incorporated in the output stage, also biased for class B operation, but configured to amplify only the "negative" half-cycles of the audio signal, it would be possible to re-create the complete original amplified audio signal at the output. This is the basic principle behind the operation of a class B audio amplifier.

There is still one drawback with class B amplification. At the point where one transistor goes into cutoff and the other transistor begins to conduct (the zero reference point of the AC audio signal), a little distortion will occur. This is referred to as crossover distortion. Crossover distortion can be reduced by biasing both output transistors just slightly into the conductive region in the quiescent state. Consequently, each output transistor will begin to conduct at a point slightly in advance of the other transistor going into cutoff. By this method, crossover distortion is essentially eliminated, without degrading the amplifier's efficiency by a significant factor.

This mode of operation is referred to as class AB. An efficiency factor of up to 78.5% applies to class AB amplifiers.

FIG. 2 A hypothetical class AB audio amplifier.

FIG. 2 is an example of a hypothetical class-AB audio amplifier. C1 is the input coupling capacitor, R1 and R2 form the familiar voltage divider bias network for biasing Q1, RE is Q1's emitter resistor, and RC is Q1's collector resistor. These components make up a typical common emitter transistor amplifier. Q2 and potentiometer P1 are configured in a circuit arrangement called an amplified diode. The purpose of this circuit is to provide the slight forward bias required on both output driver transistors to eliminate crossover distortion. Q3 and Q4 are the output drivers; with Q3 amplifying the positive half-cycles of the audio signal, and Q4 amplifying the negative half-cycles. C2 is an output-coupling capacitor; it serves to block the DC quiescent voltage from reaching the speaker, while allowing the amplified AC output voltage to pass.

The amplified diode circuit of Q2 and P1 could be replaced with two forward-biased diodes. In theory, each diode would drop about the same voltage as the forward biased base-emitter junction of each output transistor. The problem with this method is a lack of adjustment. If the for ward threshold voltage of each diode is not exactly equal to the base-emitter junction voltage of each transistor, some crossover distortion can occur. If three diodes are used, the quiescent conduction cur rent of each output transistor might be too high, resulting in excessive heating of the output transistors.

The amplified diode circuit could also be replaced with an adjustable biasing resistor for biasing purposes. Although this system will function well and eliminate crossover distortion, the adjustable resistor will not thermally "track" with the output transistors. As you might recall, bipolar transistors have a negative temperature coefficient, meaning that they exhibit a decrease in resistance with an increase in temperature. In reference to transistors, a decrease in resistance actually means an increase in leakage current. In other words, bipolar transistors become more "leaky" when they get hot. In bipolar transistor amplifiers, this is a major problem. As output transistors begin to heat up, the leakage current also increases, causing an increase in heat, causing an increase in leakage cur rent, causing an additional increase in heat, causing an additional increase in leakage current, and so forth. This condition will continue to degrade until the output transistors break down. A breakdown of this nature is called thermal runaway.

A means of automatic thermal compensation is needed to correct the problem. An adjustable resistor cannot do this (most resistors have a positive temperature coefficient), but that is the beauty of an amplifier diode circuit. Referring to FIG. 2, if Q2 is placed on the same heatsink as the output transistors, its temperature rise will closely approximate that of the output drivers. As the leakage current increases with a temperature rise in the output transistors, the leakage current through Q2 also increases. The increase of leakage current through Q2 causes the voltage drop across it to decrease, resulting in a decrease of forward bias to the output transistors. The decrease in forward bias compensates for the increase in leakage current, thus resulting in good temperature stabilization.

Additional Amplifier Classification

There are additional classes of amplifier operation, but they are not typically used for audio amplifiers. Class C amplifiers are biased to amplify only a small portion of a half-cycle. They are used primarily in RF (radio-frequency) applications, and their efficiency factors are usually about 80%.

Class D amplifiers are designed to amplify "pulses," or square waves. A class D amplifier is strictly a "switching device," amplifying no part of an input signal in a linear fashion. Strangely enough, class D amplifiers are available (although rarely) as audio amplifiers through a technique called pulse-width modulation (PWM). A PWM audio amplifier outputs a high-frequency (about 100 to 200 kHz) square wave to the audio speaker.

Because this is well above human hearing, the speaker cannot respond. But the duty cycle (on-time/off-time ratio) of the square-wave output is varied according to the audio input signal. In effect, this creates a proportional "power signal," which the speaker does respond to, and the audio input signal is amplified. Class D audio amplifiers boast extremely high efficiencies, but they are expensive, and they have drawbacks in other areas. Class D audio amplifiers are sometimes called digital audio amplifiers. Most class D amplifiers are more commonly used for high-power switching and power conversion applications.

Audio Amplifier Output Configurations

The circuit illustrated in FIG. 2 has a complementary symmetry output stage. This term means that the output drivers are of opposite types (one is NPN, and the other is PNP) but have symmetric characteristics (same beta value, base-emitter forward voltage drop, voltage ratings, etc.). Generally speaking, most audiophiles consider this to be the best type of output driver design. Transistor manufacturers offer a large variety of matched pair, or complementary pair, transistor sets designed for this purpose.

Another common type of output design is called the quasi-complementary symmetry configuration. It requires a complementary symmetry "predriver" set, but the actual output transistors are of the same type (either both NPN, or both PNP; NPN outputs are vastly more popular).

This type of output design used to be a lot more popular than it is now.

The current availability of a large variety of high-power, high-quality complementary transistor pairs has overshadowed this older design. However, it produces good-quality sound with only slightly higher-distortion characteristics than complementary symmetry.

Audio Amplifier Definitions

The field of audio electronics is an entertainment-oriented field. The close association between audio systems and the arts has led to a kind of semi-artistic aura surrounding the electronic and electromechanical systems themselves. As with any artform, personal preference and taste play a major role. This is the reason why there are so many disputes among audiophiles regarding amplifier and speaker design. My advice is to simply accept what sounds good to you, without falling prey to current trends and fads.

Unfortunately, there have been many scams and sly stigmas perpetrated by unethical, get-rich-quick manufacturers over the years. This has led to much misunderstanding and confusion regarding the various terms used to define audio amplifier performance.

The most heavily abused characteristic of amplifier performance is "power." Power, of course, is measured in watts. The only standardized method of designating AC wattage, for comparison purposes, is by using the rms value. Any other method of rating an amplifier's power output should be subject to suspicion.

Output power is also rated according to the speaker load. For example, an amplifier specification might rate the output power as being 120 watts rms into a 4-ohm load, and 80 watts rms into an 8-ohm load. You might expect the power output to double when going from an 8-ohm load to a 4-ohm load, but there are certain physical reasons why this will not happen. However, when comparing amplifiers, be sure to compare "apples with apples"; an amplifier rated at 100 watts rms into an 8-ohm load is more powerful than an amplifier rated at 120 watts rms into a 4-ohm load.

The human ear does not respond in a linear fashion to differing amplitudes of sound. It is very fortunate for you that you are made this way, because the nonlinear ear response allows you to hear a full range of sounds; from the soft rustling of leaves to a jackhammer pounding on the pavement. For example, a loud sound that is right on the thresh old of causing pain to a normal ear is about 1,000,000,000,000 times louder than the softest sound that can be heard. Our ears tend to "compress" louder sounds, and amplify smaller ones. In this way, we are able to hear the extremely broad spectrum of audible sound levels.

When one tries to express differing sound levels, power ratios, noise content, and various other audio parameters, the nonlinear characteristic of human hearing presents a problem. It was necessary to develop a term to relate linear mathematical ratios with nonlinear hearing response. That term is the decibel. The prefix deci means 1/10, so the term decibel actually means "one-tenth of a bel." The bel is based on a logarithmic scale. Although I can't thoroughly explain the concepts of logarithms within this context, I can give a basic feel for how they op e r at e . Logarithms are trigonometric functions, and are based on the number of decimal "columns" contained within a number, rather than the decimal values themselves. Another way of putting this is to say that a logarithmic scale is linearized according to powers of ten. For example, the log of 10 is 1; the log of 100 is 2; the log of 1000 is 3. Notice, in each case, that the log of a number is actually the number of weighted columns within the number minus the "units" column. The bel is a ratio of a "reference" value, to an "expressed" value, stated logarithmically. A decibel is simply the bel value multiplied by 10 (bels are a little too large to conveniently work with).

In this case, I believe a good example is worth a thousand words. Assume you have a small radio with a power output of 100 mW rms.

During a party, you connect the speaker output of this radio into a power amplifier which boosts the output to 100 watts rms. You would like to express, in decibels, the power increase. The power level that you started with, 100 mW (0.1 watt), is your reference value. Dividing this number into 100 watts gives you your ratio, which is 1000. The log of 1000 is 3 (bel value). Finally, multiply 3 by 10 (to convert bels to decibels), and the answer is 30 decibels.

Each 3-dB increase means a doubling of power: 6 dB gives 4 times the power [3 dB _ 3 dB equates to 2x power times 2x power; 2x(2x) _ 4x]. A 9-dB power increase converts to an 8x power increase [6 dB _ 3 dB, or 4x(2x) _ 8x].

Each 10-dB increase equals a 10-fold increase; for instance, 20 dB yields 100x [10 dB _ 10 dB, or 10x times 10x]. And finally, as per the example above, a 30-dB power gain means a 1000-fold increase [10 dB _ 10 dB _ 10 dB _ 10x3 _ 1000x].

Also, please be aware that negative values of decibels represent negative gain, or attenuation. A _3-dB gain means that the power has been halved. Similarly, a _10-dB gain represents a 10-fold attenuation, or a 0.1x change in power output.

These figures represent power logs. Voltage and current decibel logs are somewhat different: the square root of the power logs. This is because power is voltage times current, P _ IE; 30-dB volts is 31.620, and 30-dB amps is also 31.620. Thus, logP = logI x logE = 31.62 x 31.62=999.8x. Most electronic reference books have decibel log tables for easy reference. Just be aware that there is a difference between the power logs, and the volt age or current logs.

I recognize that if you have not been exposed to the concept of logarithms, or exponential numbering systems, this entire discussion of decibels is probably rather abstract. If you would like to research it further, most good electronics math books should be able to help you understand it in more detail.

Dynamic range is a term used to describe the difference (in decibels) between the softest and loudest passages in audio program material. In a practical sense, it means that if you are listening to an audio system at a 10-watt rms level, then for optimum performance, you would probably want about a 40-watt rms amplifier to handle the instantaneous high volume passages that might be contained within the program material (a cymbal crash, for example). Compact-disk and "hi-fi" (high-fidelity) videotape recorders offer the widest dynamic range commonly available in today's market.

Frequency response defines the frequency spectrum that an amplifier can reproduce. The normal range of human hearing is from 20 to 20,000 Hz (if you're a newborn baby, and had Superman as a father). In theory, there are situations occurring in music where ultrasonic frequencies are produced which are not audible, but without them, the audible frequencies are "colored" to some degree, causing a variance from the original sound. For this reason, many high-quality power amplifiers have frequency responses up to 100,000 hertz. The high-end and low-end frequency response limits are specified from the point where the amplifier power output drops to 50% (_3 dB) of its rated output.

Distortion is a specification defining how much an amplifier changes, or "colors," the original sound. A perfect amplifier would be perfectly "linear," meaning that the output would be "exactly" like the input, only amplified. However, all amplifiers distort the original signal by some percentage. In the mid-1970s, it was a commonly accepted fact that the human ear could not distinguish distortion levels below 1%. That has since been proved wrong. It is a commonly accepted rule of thumb today that even a trained ear has difficulty detecting distortion below 0.3%, although this figure is often disputed as being too high among many audiophiles. In any case, the lower the distortion specifications, the better.

Distortion is subdivided down into two more specific categories in modern audio amplifiers: harmonic distortion and intermodulation distortion. Harmonic distortion describes the nonlinear qualities of an amplifier. In contrast, intermodulation distortion defines how well an amplifier can amplify two specific frequencies simultaneously, while preventing the frequencies from interfering with each other in a nonlinear fashion.

Typical ratings for both of these distortion types is 0.1% or lower in modern high-quality audio amplifiers.

Load impedance defines the recommended speaker system impedance for use with the amplifier. For example, if the amplifier specification indicates the load impedance as 4 or 8 ohms, you might use either a 4- or 8-ohm speaker system (or any impedance in between) as the output load for the amplifier.

Input impedance describes the impedance "seen" by the audio input signal. This value should be moderately high: 10 Kohms or higher.

Sensitivity defines the rms voltage level of the input audio signal required to drive the amplifier to full output power. Typical values for this specification are 1 to 2 volts rms.

Signal-to-noise ratio is a specification given to compare the inherent noise level of the amplifier with the amplified output signal. Random noise is produced in semiconductor devices by the recombination process occurring in the junction areas, as well as other sources. High quality audio amplifiers incorporate various noise reduction techniques to reduce this undesirable effect, but a certain amount of noise will still exist and be amplified right along with the audio signal. Typical signal to-noise ratios are _70 to _92 dB, meaning that the noise level is 70 to 92 dB below the maximum amplifier output.

There are additional specifications that might or might not be given in conjunction with audio amplifiers, but the previous terms are the most common and the most important.

Before proceeding, here is a note of caution. Research has proved that continued exposure to high-volume noise (meaning music or any other audio program material) causes degradation of human hearing response. It saddens me to hear young people driving by in their cars with expensive audio systems blasting out internal sound pressure levels at 120 dB. Even relatively short exposures to this level of sound can cause them to develop serious nerve-deafness problems by the time they're middle-aged. Of course, exposure to high-volume levels at any age is destructive. It isn't worth it. Keep the volume down to reasonable levels for your ear's sake.

Power Amplifier Operational Basics

FIG. 3 Power amplifier operational basics.

Now that some of the basic audio terms have been established, this section will concentrate on the "front end" of modern audio amplifier design. In addition, this section establishes some of the basics relating to integrated circuit operational amplifiers, which will be discussed later in this guide.

FIG. 3 is a kind of semiblock diagram illustrating the input stage of most high-power audio amplifiers. Q1, D1, D2, and P1 form a circuit called a constant-current source. For discussion purposes, assume that D1 drops the same voltage as the base-emitter junction of Q1 (which should be a close assumption). That would mean that the voltage drop across D2 would also be the voltage drop across P1. If the voltage drop across D2/P1 is 0.7 volt, and P1 is adjusted to be 700 ohms, the emitter current flow will be 1 milliamp. Because the collector current of Q1 will approximately equal the emitter current, the collector current is also "held" at 1 milliamp. The important point to note here is that the collector current is regulated; it is not dependent on the collector load or the amplitude of the source voltage. The only variables controlling the collector current are D2's forward threshold voltage and the setting of P1. Therefore, it is appropriately named a constant-current source.

Transistors Q2 and Q3 form a differential amplifier. Think of a differential amplifier as being like a seesaw in a school playground. As long as everything is balanced on a seesaw, it stays in a horizontal position. If something unbalances it, it tilts, causing one end to go up proportionally, as the other end goes down. This is exactly how a differential amplifier operates with the current flow. As discussed previously, it is assumed that the constant current source will provide a regulated 1 milliamp of current flow to the emitters of Q2 and Q3. If Q2 and Q3 are in a balanced condition, the 1 milliamp of current will divide evenly between each transistor, providing 0.5 milliamp of current flow through each collector. If an input voltage is applied between the two base inputs (A and B) so that point A is at a different potential than point B, the balance will be upset. But as the collector current rises through one transistor, it must decrease by the same amount through the other, because the constant-current source will not allow a varying "total" current. For example, if the differential voltage between the inputs caused the collector current through Q2 to rise to 0.6 milliamp, the collector current through Q3 will fall to 0.4 milliamp. The total current through both transistors still adds up to 1 milliamp.

Notice that the output of the differential amplifier is not taken off of one transistor in reference to ground; the output of a differential amplifier is the difference between the two collectors. Now let's discuss the advantages of such a circuit.

Assume that the source voltage (supplied externally) increases. In a common-transistor amplifier, an increase in the source voltage will cause a corresponding change throughout the entire transistor circuit.

In FIG. 3, an increase in the source voltage (+V) does not cause an increase in current flow from the constant current source, because it is regulated by the forward voltage drop across D2, which doesn't change (by practical amounts) with an increase in current. Q2 and Q3 would still have a combined total current flow of 1 milliamp. The collector voltages of Q2 and Q3 would increase, but they would increase by the same amount, even if the circuit were in an unbalanced state.

Therefore, the voltage differential between the two collectors would not change. For example, assume that Q2's collector voltage is 6 volts and Q3's collector voltage is 4 volts. If you used a voltmeter to measure the difference in voltage between the two collectors, it would measure 2 volts (6 _ 4 _ 2). Now assume the source voltage increased by an amount sufficient to cause the collector voltages of Q2 and Q3 to increase by 1 volt. Q2's collector voltage would rise to 7 volts, and Q3's would rise to 5 volts. This didn't change the voltage differential between the two collectors at all; it still remained at 2 volts. In other words, the output of a differential amplifier is immune to power supply fluctuations. Not only does this apply to gradual changes in DC levels; the effect works just as well with power supply ripple and other sources of undesirable noise signals that might enter through the power supply.

One of the most common problems with high-gain amplifiers is noise and interference signals being applied to the input through the input wires. Input wires and cables can pick up a variety of unwanted signals, just as an antenna is receptive to radio waves. If you have ever touched an uninsulated input to an amplifier, you un-doubtably heard a loud 60-hertz roar (called "hum") through the speaker. This is because your body picks up electromagnetically radiated 60-hertz signals from power lines all around you. Fluorescent lights are especially bad electro magnetic radiators. FIG. 3 illustrates an example of some electromagnetic radiation causing noise pulses on the A and B inputs to the differential amplifier. Because electromagnetic radiation travels at the speed of light (186,000 miles per second), the noise pulses will occur at the same time, and in the same polarity. This is called common-mode interference. A very desirable attribute of differential amplifiers is that they exhibit common-mode rejection. The noise pulses illustrated in FIG. 3 would not be amplified.

To understand the principle behind common-mode rejection, assume that the positive-going noise pulse on the A input is of sufficient amplitude to try to cause a 1-volt decrease in Q2's collector voltage. Because the noise is common mode (as is all externally generated noise), an identical noise pulse on the B input is trying to cause Q3's collector voltage to decrease by 1 volt also. If both collectors decreased in voltage at the same time, it would require an increase in the combined total current flow through both transistors. This can't happen because the constant current source is maintaining that value at 1 milliamp.

Therefore, neither transistor can react to the noise pulse, and it is totally rejected. (Even if both transistors did react slightly, they would react by the same amount. Because the output of the pair is the difference across their collectors, a slight reaction by both at the same time would not affect their differential output.) This goes back to the analogy of the seesaw I made earlier. If a seesaw is balanced and you placed equal weights on both ends at the same time, it would simply remain stationary. In contrast, the desired signal voltage to be amplified is not common mode. For example, the B input might be at signal ground while the A input is at 1 volt rms. Differential amplifiers respond very well to differential signals. That is why they are called differential amplifiers.

One final consideration of FIG. 3 is in reference to RF and CF.

Notice that this resistor/capacitor combination is connected from the output back to one of the inputs. The process of applying a percentage of the output back into the input is called feedback. In high-gain amplifiers, this feedback is almost always in the form of negative feedback, meaning the feedback acts to reduce the overall gain. Negative feedback is necessary to temperature stabilize the amplifier, flatten out the gain, increase the frequency response, and eliminate oscillations. Various combinations of resistors and capacitors are chosen to tailor the frequency response, and to provide the best overall performance. Feed back will be discussed further in later sections.

Building High-Quality Audio Systems

You don't have to be an electrical engineer to build much of your own high-quality audio equipment. Even if your interests don't lie in the audio field, you are almost certain to need a lab-quality audio amplifier for many related fields. In this section, I have provided a selection of audio circuits that are time-proven, and which provide excellent performance.

FIG. 4 is a block diagram of a typical audio amplification system.

It is mostly self-explanatory, with the exception of the volume control and the two 10-uF capacitors. The volume control potentiometer should have an audio taper (logarithmic response). A typical value is 100 Kohms.

For stereo applications, this is usually a "two-ganged" pot, with one pot controlling the right channel and one pot controlling the left. The two 10-uF capacitors are used to block unwanted DC shifts that might occur if the volume control is rotated too fast.

FIG. 4 Block diagram of a typical audio amplification system.

FIG. 5 Preamplifier for use with high-Z signal sources.

FIG. 6 High-gain preamplifier.

FIG. 5 is a simple preamplifier circuit for use with high-impedance signal sources, such as crystal or ceramic microphones. It is merely a common-collector amplifier with a few refinements. R5 and C4 are used to "decouple" the circuit's power source. The simple RC filter formed by R5 and C4 serves to isolate this circuit from any effects of other circuits sharing the same power supply source. R3 and C2 provide some positive feed back (in phase with the input) called "bootstrapping." Bootstrapping has the effect of raising the input impedance of this circuit to several Mohms.

FIG. 6 is a high-gain preamplifier circuit for use with very low input signal sources. Dynamic microphones, and some types of musical instruments (such as electric guitars), work well with this type of circuit.

R1 provides negative feedback for stabilization and temperature compensation purposes. Notice that this circuit is also decoupled by R5 and C2.

FIG. 7 is an active tone control circuit for use with the outputs of the previous preamplifier circuits, or any line-level output. Active tone controls incorporate the use of an active device (transistor, FET, operational amplifier, etc.), and can provide better overall response with gain. Passive tone controls, in contrast, do not use any active devices within their circuits, and will always attenuate (reduce) the input signal. Line-level outputs are signal voltages that have already been pre-amplified. The audio out puts from CD players, VCRs, tape players, and other types of consumer electronic equipment are usually line-level outputs.

FIG. 7 Active tone control circuit.

FIG. 8 is a 12-watt rms audio power amplifier (the term power amplifier implies that the primary function of the amplifier is to pro vide low-impedance, high-current driving power to a typical loudspeaker system). It is relatively easy to construct, provides good linearity performance, and operates from a single DC supply (most audio power amplifiers require a dual-polarity power supply). A simple "raw" DC power supply, similar to the one illustrated in Fig. 4, constructed from a 2 amp, 24-volt transformer, a 2-amp bridge rectifier, and a 1000 uF filter capacitor at 50 working volts DC (WVDC) should power this circuit nicely (the rectified and filtered DC voltage produced from a 24-volt trans former will come out to about 36 volts DC). The complete circuit can be assembled on a small universal perfboard, or it can be constructed on its own PC board (PC circuit board fabrication is described later in this section; this project is mentioned here because it is ideal to test your first attempts at PC board making.) The parts list for this amplifier is given in Table 1.

Fig. 8--A 12-watt rms audio power amplifier.

TABLE 1 Parts List for 12-Watt Audio Amplifier Project

A functional description of FIG. 8 is as follows. Beginning at the left hand side of the illustration, the "audio signal input" is a line-level signal from a preceding audio device, such as an FM receiver or cassette tape deck. C1 is a coupling capacitor, blocking the DC bias on the base of Q1 from being applied to the preceding audio device.

Transistors Q1 and Q2 make up a differential amplifier, which functions as the first amplification stage of the amplifier. Differential amplifiers are often chosen as the first stage of an audio amplifier because they provide a convenient point of applying negative feedback (i.e., the inverting input; the base of Q2), and because they are capable of high current gain and high input impedance and are relatively insensitive to power supply fluctuations. A differential amplifier's unique quality of common-mode rejection is a paramount issue pertaining to their use in operational amplifiers (discussed in Section 12), but it is not used at all within the context of most audio power amplifiers. A single resistor, R3, is used in place of a constant-current source, which is typically adequate for many medium-quality audio power amplifiers. R4 is the load resistor for the differential amplifier. Note that Q2 doesn't have a load resistor; this is because an output signal is not needed from Q2.

Resistors R1 and R2 form a voltage divider that splits the power sup ply in half, biasing the base of Q1 at half of the power supply voltage, or approximately 18 volts. Resistors R5 and R6 provide the identical function for the base of Q2. The combined voltage-dividing effect of R1, R2, R5, and R6 is to cause the entire amplifier circuit to operate from a reference of half of the power supply voltage, thereby providing the maxi mum peak-to-peak voltage output signal.

The second amplification stage of FIG. 8 is made up of Q3 and its associated components. This is a simple common-emitter amplifier stage, in which the audio signal is applied to the base of Q3 and the amplified output is taken from its collector. R8 serves as the collector load for Q3.

Diodes D1, D2, and D3 provide a small forward bias to Q4, Q5, Q6, and Q7, to reduce the effects of crossover distortion. Capacitor C2 functions as a compensation capacitor. Compensation is a common term used with linear circuits (especially operational amplifiers) referring to a reduction of gain as the signal frequency increases. High-gain linear circuitry will always incorporate some amount of negative feedback (as discussed previously, to improve the overall performance). At higher frequencies, because of the internal capacitive characteristics of semiconductors and other devices, the negative-feedback signal will increasingly lag the input signal (remember, voltage lags the current in capacitive circuits). At very high frequencies, this voltage lag will increase by more than 180 degrees, causing the negative-feedback signal to be "in phase" with the input signal. If the voltage gain of the amplifier is greater than unity at this frequency, it will oscillate. Therefore, compensation is the general technique of forcing the voltage gain of a linear circuit to drop below unity before the phase shifted negative feedback signal can lag by more than 180 degrees. In gist, compensation ensures stability in a high-gain amplifier or linear circuit.

To understand how C2 provides compensation in the circuit of FIG. 8, note that it connects from the collector of Q3 to the base of Q3. As you recall, the output (collector signal) of a common-emitter amplifier is 180 degrees out of phase with the input (base signal). Therefore, as the signal frequency increases, causing the impedance of C2 to drop, it begins to apply the collector signal of Q3 to the base of Q3. The 180-degree out-of phase collector signal is negative feedback to the base signal, so as the frequency increases, the voltage gain of Q3 decreases. The result is a falling-off (generally called rolloff) of gain at higher frequencies, promoting good audio frequency stability of the FIG. 8 amplifier.

All of the voltage gain in the FIG. 8 amplifier occurs in the first two stages. Therefore, when an audio signal is applied to the input of the amplifier, the signal voltage at the collector of Q3 will be the "maxi mum" output signal voltage that the amplifier is capable of producing.

There are two other points regarding the signal voltage at the collector of Q3 that should be understood. First, it is in phase with the input signal. This is because the original audio signal was inverted once at the collector of Q1, and it is inverted again at the collector of Q3, placing it back in phase with the input. Second, the audio input signal was super imposed on the DC quiescent level at the base of Q1, which is set to half of the power supply voltage, or about 18 volts. Therefore, a positive half cycle of the audio signal on the collector of Q3 will vary from 18 to 36 volts. In contrast, a negative half-cycle of the audio signal will vary from 18 volts down to 0 volt (i.e., the amplified AC signal voltage is superimposed on a quiescent DC level of half of the power supply voltage).

The third stage of FIG. 8 consists of Q4, Q5, Q6, Q7, and their associated components. To begin, consider the operation of Q4 and Q6 only. Note that they are both NPN transistors, and the emitter of Q4 connects directly to the base of Q6. This configuration is simply a type of Darlington pair with a few "stabilizing" resistors added. As you recall, the purpose of a Darlington pair is to increase the current gain parameter, or beta, of a transistor circuit. Transistors Q4 and Q6 serve as a high-gain, current amplifier in this circuit. They will current-amplify the collector signal of Q3.

Under normal operation, the quiescent DC bias of this amplifier is such that the right-hand side of R7 will be at half of the power supply voltage, or about 18 volts (this condition is due to the DC bias placed on the input stage, as explained earlier). This also means that the emitters of Q4 and Q6 will be at about 18 volts also (minus a small drop across their associated emitter resistors). If a positive-going AC signal is applied to the input of this amplifier, transistors Q4 and Q6 will current-amplify this signal. However, as soon as the AC signal goes into the negative region, the signal applied to the base of Q4 will drop below 18 volts, causing Q4 and Q6 to go into cutoff. Therefore, Q4 and Q6 are only amplifying about half of the audio signal (i.e., the positive half-cycles of the audio AC signal). Since transistors Q5 and Q7 are PNP devices, with their associated collectors tied to circuit common, they are current amplifying the negative half-cycles of the audio AC signal, in a directly inverse fashion as Q4 and Q6. Simply stated, all audio signal voltages above 18 volts are amplified by Q4 and Q6, while all audio signal volt ages below 18 volts are amplified by Q5 and Q7. Consequently, the entire amplified audio signal is summed at the positive plate of C4.

Note that resistor R7 is connected from the amplifier's output back to the "inverting" side of the input differential amplifier (i.e., the base of Q2). R7 is a negative-feedback resistor. As you recall, the audio signal volt age at the collector of Q3 is in phase with the audio input signal. Q4, Q5, Q6, and Q7 are all connected as common-collector amplifiers (i.e., the input is applied to the bases with the output taken from the emitters). Since common-collector amplifiers are noninverting, the audio signal at the amplifier's output remains in phase with the input signal.

Therefore, the noninverted output signal that R7 applies back to the inverting input of the differential amplifier is negative feedback. Negative feedback in this circuit establishes the voltage gain, increases linearity (i.e., decreases distortion), and helps stabilize the quiescent voltage levels. The voltage gain (Ae ) of this amplifier is approximately equal to R7 divided by the parallel resistance of R5 and R6 (i.e., about 10).

Capacitor C4, like C1, is a coupling capacitor, serving to block the quiescent 18-volt DC level from being applied to the speaker. Note, however, that the value of C4 is very large compared to the value of C1. This is necessary because the impedance of most speakers is only about 8 ohms. Therefore, in order to provide a time constant long enough to pass low frequency audio signals, the capacity of C4 must be much greater.

Finally, resistor R13 and capacitor C3 form an output circuit that is commonly referred to as a Zobel network. The purpose of a Zobel net work is to "counteract" the effect of typical speaker coil inductances, which could have a destabilizing effect on the amplifier circuitry.

As you may have noted, the FIG. 8 amplifier consists of three basic stages, commonly referred to as the input stage (Q1 and Q2), the voltage amplifier stage (Q3), and the output stage (Q4, Q4, Q6, and Q7). Virtually all modern solid-state audio power amplifiers are designed with this same basic three-stage architecture, commonly referred to as the Lin three-stage topology.

Constructing the 12-Watt RMS Amplifier of FIG. 8

If you decide to construct this amplifier circuit on a universal bread board or solderless breadboard, the construction is rather simple and straightforward. You will need to provide some heatsinking for output transistors Q6 and Q7. If you mount the amplifier in a small metal enclosure, adequate heatsinking can be obtained by simply mounting Q6 and Q7 to the enclosure (remember to ensure that Q6 and Q7 are electrically isolated from the enclosure).

For optimum performance, diodes D1 and D3 should thermally track the temperature of transistors Q6 and Q7. This can be accomplished in several ways. The diodes could be glued to the case of the output devices with a small drop of epoxy, or you can bend a small solderless "ring" terminal into a makeshift clamp, with the ring held in place by the transistor's mounting bolt. If you construct the amplifier circuit similar to the FIG. 9a layout, you can simply bend the diodes into a touching position with the output transistors (remember to solder the diodes high above the board surface if you want to use this technique).

If you plan to use a current-limited power supply to test this amplifier circuit (such as the "lab-quality power supply" detailed in Sections 3 through 6), fuse F1 isn't necessary. If you provide operational power to this amplifier circuit with a simple raw DC power supply as mentioned earlier, F1 must be included for safety purposes. Also, keep in mind that if you accidentally "short" (short-circuit) the speaker output leads together, you will probably destroy one or both of the output transistors (i.e., Q6 and Q7).

FIG. 9 12-watt audio amplifier: (a) and (b) top views silkscreen lay out; (c) bottom view copper artwork.

FIG. 10 Illustration of a typical PC board fabrication kit.

Making Printed Circuit Boards by Hand

Making PC boards is not as difficult as you may have been led to believe, or as your past experiences may have indicated. Circuit board manufacture is a learned technique, and like any technique, there are several "correct" ways of going about it and many "wrong" ways of accomplishing disaster. In this section, I'm going to detail several methods that can be successfully used by the hobbyist. With a little practice, either of these methods should provide excellent results. The 12-watt audio amplifier described in the previous section is an excellent "first" project for getting acquainted with PC board fabrication, so I included a set of PC board layout illustrations in FIG. 9. See Section C for the full-size set to be used in your project.

To begin, you will need to acquire some basic tools and materials. If you are a complete novice regarding PC board construction, I recommend that you start by purchasing a "PC board kit," such as the one illustrated in FIG. 10. Typically, such kits will contain a bottle of etchant (an acid solution used to dissolve any exposed copper areas on the PC board), a resist ink pen (a pen used to draw a protective ink coating over any cop per areas that you don't want dissolved by the etchant), several "blank" (i.e., unetched) pieces of PC board material, and some miscellaneous supplies that you may or may not need. In addition, you will need a few very small drill bits (no. 61 is a good size), an electric hand drill, some fine grained emery paper, a small pin punch, a tack hammer, and a glass tray.

If you want to make a PC board of the FIG. 9 artwork by hand, the following procedure can be used. Begin by observing the illustrations provided in FIG. 9. FIG. 9a is commonly called the silkscreen layout.

It illustrates a top view of the placement and orientation of the components after they are correctly mounted to the PC board. FIG. 9b is another top view of the silkscreen, illustrating how the bottom-side copper artwork will connect to the top-mounted components-the PC board is imagined to be transparent. FIG. 9c is a reflected view of the bottom-side copper artwork. In other words, this is exactly how the cop per artwork should look if you turn the board upside down and look at it from the bottom. The reflected view of the copper artwork, FIG. 9c, is what you will be concerned with in the next phase of your PC board construction.

Make a good copy of FIG. 9c on any copy machine. Cut a piece of PC board material to the same size as the illustration. Cut out the illustration from the copy and tape it securely on the "foil" side of the PC board material. Using a small pin punch and tack hammer, make a dimple in the PC board copper at each spot where a hole is to be drilled.

When finished, you should be able to remove the copy and find a dimple in the copper corresponding to every hole shown in the artwork diagram. Next, drill the component lead holes through the PC board at each dimple position. When finished, hold the PC board up to a light, with the diagram placed over the foil side, to make certain you haven't missed any holes and that all of the holes are drilled in the right positions. If everything looks good, lightly sand the entire surface of the copper foil with 600-grit emery paper to remove any burrs and surface corrosion.

Using the resist ink pen, draw a "pad" area around every hole. Make these pads very small for now; you can always go back and make selected ones larger, if needed. Using single lines, connect the pads as shown in the artwork diagram in the following manner. Being sure you have the board turned correctly to match the diagram, start at one end and connect the simplest points first. Using these first points as a reference, eventually proceed on to the more difficult connections. When finished, you'll have a diagram that looks like a "connect the dots" picture in a coloring book. Finally, go back and "color" in the wide foil areas (if applicable) and fill in the wider tracks. The process is actually easier than it appears at first glance. If you happen to make a major mistake, just remove all of the ink with ink solvent or a steel wool pad and start over again-nothing is lost but a little time. You'll be surprised at how accomplished you will become at this after only a few experiences.

When you're satisfied that the ink pattern on the PC board corresponds "electrically" with the reflected artwork of FIG. 9c, place it in a glass or plastic tray (not metal!), and pour about an inch of etchant solution over it. Be very careful with this etchant solution; it permanently stains everything it comes in contact with, including skin. Wear goggles to protect your eyes, and don't breathe the fumes. After about 15 to 20 minutes, check the board using a pair of tongs to lift it out of the etchant solution. Continue checking it every few minutes until all of the unwanted copper has been removed. When this is accomplished, wash the board under cold water, and remove the ink with solvent or steel wool. When finished, the PC board will be ready for component installation and testing.

If you construct any PC boards using the aforementioned procedure, you will discover that the copper artwork on the finished PC board is slightly "pitted" in areas where the resist ink did not adequately protect the copper from the etchant. If you want to improve the finished quality of your PC boards, use inexpensive fingernail polish instead of resist ink. Obtain a dried-out felt-tipped marking pen (one with an extra-fine point), repeatedly dip the pen in the fingernail polish (like an old quill ink pen), and use it to draw your pads and traces in the same way that you would use the resist ink pen as described above. After etching, the fingernail polish can be removed with ordinary fingernail polish remover, and the copper foil surface of your PC board will be totally free of any pits or corrosion from the etchant.

Making Printed Circuit Boards by the Photographic Method

The previously described method of fabricating PC boards by hand is acceptable for some surprisingly complex PC patterns, but it is very time-consuming and non-repeatable (i.e., it is difficult to make exact replications). The serious hobbyist will most certainly develop the need to make PC boards in a more rapid and repeatable manner, and the best system that I have discovered for accomplishing this is by means of a photographic process. As in the case of the method discussed above, the photographic method of PC board fabrication is a learned technique, requiring a minimal amount of trial-and-error learning experience for optimum results. Don't get discouraged if your first attempt turns out less than ideal.

In addition to the previously listed materials, the photographic method of making PC boards will require some additional tools and materials. The PC board "blanks" will need to be pre-sensitized with a positive-acting photoresist. Such pre-sensitized PC boards are available from a wide variety of electronics suppliers, and the price is very reasonable. You will need a fluorescent light source, consisting of two ordinary 18-inch fluorescent lights positioned side by side, that can be placed approximately 2 inches above the surface of the PC board material. You will also need a supply of office transparencies that can be used with typical copy machines, available from office supply stores.

And finally, you will need some positive PC board developer (available from electronics suppliers; can be purchased as a dry powder or liquid) and a pane of glass somewhat larger than any PC board size that you intend to make.

The process of making a PC board with the positive photographic method is as follows. Begin by making a photopositive of the "reflected" PC board artwork that you intend to fabricate. This is accomplished by making a copy of the black-and-white reflected artwork illustration onto a sheet of clear transparency with a copy machine. The finished photopositive should look identical to the reflected artwork illustration (i.e., all of the copper areas to be saved are black; the copper areas to be etched are transparent), except that the white areas on the original illustration will be transparent on the photopositive. This dark area on this photopositive must be as opaque as possible. If you can see "leakage light" through the dark areas when holding the photopositive up to a light source, you should adjust the copier for a darker copy. If the copy is still too light, you may need to make two copies, and tape one copy over the other for increased opacity.

Once the photopositive is finished, you are ready to expose and develop the PC board artwork. Adjust the ambient lighting of your work area so that it is as dark as possible, but light enough to allow you to see what you're doing. Remove the protective plastic film from the pre-sensitized PC board material, and lay the PC board down (sensitized side up) onto a flat work surface. Place the photopositive on top of the pre-sensitized surface of the PC board, and then place the pane of glass on top of the photopositive (i.e., the photopositive is sandwiched flat between the pre sensitized surface and the pane of glass). At this point, make sure that the placement of the photopositive is correct on the PC board surface and that you haven't accidentally turned the photopositive upside down (getting the photopositive upside down is a very easy mistake to make- I've done it many times!). If all looks good, adjust your fluorescent lights to about 2 inches above the surface of the pane of glass. Turn on the fluorescent lights and expose the PC board for about 8 minutes.

After the exposure time is completed, remove the glass pane and photopositive, and store them for future use. Place the exposed PC board in a plastic or glass tray containing the "positive PC board developer solution." The recommended quantities of developer solution and mixing instructions will vary with different manufacturers, so follow the directions for preparing the developer solution as indicated on the specific type that you purchase. When the exposed PC board is placed in the developer solution, the photoresist coating that has been exposed to the fluorescent light will begin to dissolve, leaving a clear pattern of the desired PC board artwork. When all of the exposed photoresist has dissolved away, leaving a clean and bright copper surface underneath, remove the PC board from the developer solution and rinse under cold running water.

After the developed PC board has dried, examine the photoresist pat tern carefully. If there are numerous "pits" or "light-colored areas" in the photoresist surface, this means you suffered some bleed-through of fluorescent light through your photopositive. In this case, you need to increase the opacity (i.e., darkness) of the foil pattern on your photopositive. If the pitting and/or lightness of the photoresist is not excessive, you can touch up the photoresist pattern with a little fingernail polish.

Otherwise, you should scrap the PC board, correct your photopositive, and try again. In contrast, if some of your exposed copper areas (i.e., the areas intended to be etched) have a little remaining photoresist on them, this means that you didn't leave the PC board in the developer solution long enough. The only remedy is to scrape the excess photoresist off of the copper areas intended to be etched, but if the PC board artwork is more than moderately complex, it will probably be more practical to scrap the PC board and try again.

Assuming that you examine the developed PC board and everything looks good (i.e., the copper areas to be etched are bright and clean, the photoresist areas are smooth and pit-free, and all of the artwork edges are neat and sharp), the PC board is now ready for etching. This is accomplished in the same manner as described previously for making PC boards by hand. After the etching is completed, the photoresist can be removed with a fine steel wool pad, fingernail polish remover, or a variety of solvents. Personally, I like to use a spray can of "flux remover" for removing the photoresist.

If this whole photographic fabrication process sounds a little formidable, it is probably due to the lengthy and detailed description of the process that must be contained in a textbook of this sort. You might be a little shaky during your first attempt, but after a few experiences you'll easily get the hang of it. My wife has observed me making PC boards on many occasions, and she considers the whole process much less complicated than her recipe for good chicken gravy!

Other Methods of Fabricating PC Boards

As your experience in the electrical and electronics fields continues to grow, you will doubtless encounter various other methods of fabricating PC boards. Depending on your talents and resources, you may want to try a few. Generally speaking, the following is what I have discovered regarding some other popular methods.

The various methods used by large manufacturers and professional PC board fabrication companies can certainly produce professional-quality PC boards. However, the equipment and techniques required are typically too expensive for the hobbyist.

Although the photographic method, as previously described, will take more time per unit, the end results can be just as good as the professional manufacturing techniques.

The various types of "stick-on transfers" are adequate for small, uncomplicated, one-of-a-kind PC boards. However, for PC boards of medium complexity, the "fabrication by hand technique," as first described in this section, is typically faster and provides better results (it's also less frustrating than trying to work with those little sticky lines that seem to want to stick to everything except your PC board!).

The negative-acting photographic method is equal to the positive-acting method as described above. However, the materials are not as readily available, and, in some cases, it is more difficult to obtain a good "photonegative" of the PC board artwork.

The iron-on image-transfer system has become popular. With this method, you must purchase a special film intended for this technique from one of several manufacturers. The PC board artwork is copied onto the film using a copy machine or laser printer (the toner causes a chemical change on the film, making the transfer possible). The PC board artwork image is then directly ironed onto the PC board using a clothing iron or some other suitable heat source and pressure. Some hobbyists appear to favor this method, but I have not found the quality of the results to be equal to those using the photographic method. Also, the results are not as repeatable.

CAD Methods of PC Board Development

Before leaving this discussion on PC board fabrication, I would like to briefly describe the newer generation of CAD (i.e., computer-automated design) programs available for PC board development. In bygone days, an electrical engineer was forced to sit down with a schematic diagram, a tablet of graph paper, and a sordid collection of templates, rulers, and other drawing instruments, in the attempt to design a PC board layout for a new circuit design. This was usually an aggravating and time-consuming project, requiring the engineer to draw out trial-and-error track runs, reposition component placement, and redefine board dimensions, and all the while the designer had to think in an upside-down perspective as to how the track runs on the bottom of the board would connect to the component leads on top. As PC board complexities increased, this process became proportionally more difficult-in the case of modern multilayer PC boards, it became virtually impossible.

During the 1970s and 1980s, industry began developing various CAD programs to facilitate the difficult task of PC board layout design. Later, these CAD programs were "humanized" to make them more "user friendly," and marketed to electronic hobbyists for the purpose of designing "home-brew" PC boards. Currently, these PC board layout pro grams are available at modest pricing, and they are so easy to use that almost anyone can design PC boards at home. If you anticipate that you will be involved with electronics as an ongoing hobby or profession, I highly recommend that you purchase a good PC board layout program. It will be well worth the investment.

If you are new to CAD PC board layout programs, you may be wondering how you would use such a tool in a practical way. Allow me to provide a hypothetical example. Suppose you wanted to construct a small, portable guitar amplifier with tone control. You could combine the preamplifier circuit of FIG. 6 with the tone control circuit of Fig. 7 and the 12-watt power amplifier of FIG. 8. Obviously, you could con struct all of these circuits using perfboard or a universal breadboard, but the process is time-consuming and the finished product doesn't look very professional. In contrast, it would probably take you about an hour and a half to combine all three of these circuits into a single circuit design, download to a layout program, and come up with a single, finished, professional-quality PC board design. You would then use your inkjet printer to print out a full-size reflected artwork diagram directly onto a transparency. It may take you another hour to expose, develop, and etch the PC board for your project using this transparency. If you continued to work consistently, it might take you another 2 hours to drill the holes into the PC board and construct the finished project (provided you had all of the components on hand). In other words, you took a moderately complex project from the "idea" stage to the "professional-quality, completed project" stage in less than 5 hours! Your materials cost for this project was the cost of the pre-sensitized PC board blank (possibly two or three dollars, depending on size), less than one dollar's worth of chemicals, the transparency used in your inkjet printer (about 60 cents), and the cost of the electronic components.

If you are intimidated by the thoughts of "designing with computers" or "running complex programs," you shouldn't be. Most of these types of user-friendly CAD programs are self-explanatory and very easy to become comfortable with. Personally, I use the Electronics Workbench brand Multisim and Ultiboard design programs, and I can testify that on the very first time I tried using these programs, I was designing circuits in less time than it took me to figure out what the opening icons meant in Windows 95. And don't get the idea that I'm some sort of genius computer hacker-I still haven't made it through Windows ' 95 for Dummies.

A Professional-Quality Audio Power Amplifier

FIG. 11 Schematic diagram of a professional-quality audio power amplifier.

FIG. 11 is a schematic illustration of a very high-quality professional audio power amplifier. Modern high-quality, solid-state audio power amplifiers are a culmination of almost all of the commonly used linearization techniques applied to the entire field of linear electronic circuits. Therefore, a detailed dissection of the building blocks incorporated into a modern power amplifier is a very good method of understanding much of the theory and physics involved with the broad spectrum of linear electronics, including such subjects as operational amplifiers, servo systems, and analog signal processing. Also, it should be remembered that the humble audio amplifier is probably the most common electronic subsystem to be found in the vast field of consumer electronics, including such common devices as televisions, stereo systems, radios, musical instrument amplifiers, multimedia computer systems, and public address systems.

Preconstruction and Preliminary Design

The amplifier design of FIG. 11 is not a typical commercial-quality design; it is significantly superior. I chose this particular design to pro vide a baseline for discussion and illustration of various discrete building blocks that are commonly used in linear circuits. Also, it is a good project if you wish to construct something a little more advanced, and I have provided the PC board artwork if you would like to practice your PC board construction skills. (Unfortunately, space restrictions will not allow me the opportunity to provide PC board artwork for the majority of projects in this textbook-this is where a good CAD PC board design program will come in very handy!) Beginning with C1 and C2 in FIG. 11, note how they are connected in series, oriented with opposing polarities. This is a common method of creating a nonpolarized electrolytic capacitor from two typical electrolytic capacitors. The combination of C1 and C2 will look like a single nonpolarized capacitor with a capacitance value equal to half of either individual capacitor and a voltage rating equal to either one of the capacitors individually. In other words, if C1 and C2 are both rated at 22 uF at 35 WVDC, the equivalent nonpolarized capacitor will perform like an 11-uF capacitor at 35 WVDC. A nonpolarized coupling capacitor is needed for this amplifier because the quiescent volt age on the base of Q1 will be very close to circuit common potential, so the AC input signal current flow through C1 and C2 will be in both directions.

Degeneration (Negative Feedback)

R1 establishes the input impedance of the amplifier at approximately 10 kohms (the input impedance at the base of Q1 is much higher than R1). C3 provides some high-frequency filtering to protect the amplifier from ultrasonic signals that could be superimposed on the audio signal. R2 and R3 are degeneration resistors for the differential amplifier, consisting of Q1 and Q2. The term degeneration is used to describe a type of negative feedback and usually applies to resistors placed in the emitter circuits of transistors, but it can be applied to other techniques in which a single resistor is placed in the signal path of a gain stage, providing a voltage drop that opposes the gain action of an active device, thereby reducing gain and improving linearity. In simple terms, degeneration resistors flatten out the gain response of transistor circuits and improve their linearity.

Constant Current Sources

The circuit consisting of Q3, Q4, and R4 is an upside-down version of the circuit illustrated in Fig. 8d of the Transistor Workshop section of Section 6. If you refer back to Fig. 8d and the corresponding description, you will note that the Fig. 8d circuit serves the function of providing current-limit protection. In other words, it allows current to flow up to a maximum level, and then restricts it at that maxi mum level regardless of increases in the load or input signal. Referring back to FIG. 11, note that R5 has been added to the basic circuit of Fig. 8d, which provides a constant maximum base signal to Q4, which forces Q3 into a constant state of current limiting. Since the collector current of Q4 is constantly at the current-limit level (determined by the resistance value of R4), the circuit of Q3, Q4, R4, and R5 makes up a constant current source.

Actually, this type of constant-current source provides better performance than do the more conventional types of constant-current sources, such as illustrated in Fig. 6-8a. This is due to the fact that you have the combined effect of both transistors in regulating the current, instead of a single transistor referencing a voltage source. Note that the circuit consisting of Q10, Q11, R9, and R12 makes up an identical constant-current source for the voltage amplifier stage; however, the level of regulated current is set differently because of the resistance value of R12.

Current Mirrors Now, turn your attention to the peculiar-looking circuit consisting of Q5, Q6, R6, and R7 in FIG. 11. If you ever study the equivalent schematic diagrams of linear integrated circuits, you will see tons of these circuits represented. They are called current mirrors.

Current mirrors "divide" current flows into equal portions. For example, if 4 milliamps of current is flowing through the collector of Q4, the current mirror of Q5 and Q6 will force 2 milliamps to flow through Q1 and 2 milliamps to flow through Q2-it will divide the total current flowing through the differential amplifier equally through both legs.

This is important because a differential amplifier will be most linear when the current flows through both legs are "balanced" (i.e., equal to each other). Therefore, by incorporating the current mirror of Q5 and Q6, the linearity of the differential amplifier (i.e., Q1 and Q2) is optimized. The operational physics of Q5 and Q6 is rather simple. Since the collector of Q6 is connected directly to its base, the current flow through the collector of Q2 will be directly proportional to the voltage drop from the base to emitter of Q6. Since Q5's base-emitter junction is connected in parallel with Q6's base-emitter junction, Q5 is forced to "imitate" the collector current flow of Q6. Therefore, whatever the total current flow through the differential amplifier happens to be, it will be forced into a balanced state. R6 and R7 are degeneration resistors, serving to flatten out the small parameter differences between Q5 and Q6.

Voltage Amplification

Q8 provides the same function in this amplifier design as Q3 provides in the FIG. 8 amplifier. Namely, it provides the majority of "voltage gain" inherent to the amplifier. Likewise, C7 is the compensation capacitor for this amplifier, serving the same function as C2 does in the FIG. 8 amplifier.

Instead of utilizing three forward-biased diodes (to provide a small forward voltage drop, reducing the effects of crossover distortion) as in the design of FIG. 8, the amplified diode circuit of Q9, P1, and C8 is incorporated into the FIG. 11 design. Amplified diode circuits are often called Vbe multiplier circuits. An amplified diode circuit has the advantages of a more precise control of the continuous forward bias applied to the output transistors, as well as improved thermal tracking characteristics when the Vbias transistor (Q9 in this case) is physically mounted to the same heatsink as the output transistors. C8 provides smoothing of voltage variations that would normally appear across the amplified diode circuit during signal variations.

Active Loading

If you compare the amplifier design of FIG. 11 with the design of FIG. 8, you'll note that the collector load for the voltage amplifier transistor (R8 in FIG. 8) has been replaced with the Q10/Q11 constant-current source in FIG. 11. The technique of using a resistor (a passive device) for a collector load in a transistor amplifier is appropriately called passive loading. In contrast, the technique of using an active circuit for a collector load, such as the Q10/Q11 constant-current source, is called active loading.

The principles and reasons for active loading relate back to the basic operational fundamentals of a simple common emitter amplifier. As you may recall, the voltage gain of a common emitter amplifier is deter mined by the ratio of the emitter resistance to the collector resistance. In theory, if you wanted a common-emitter amplifier with a very high gain, you would be forced to make the collector resistor a very high resistance value. Unfortunately, as you may also recall, the output impedance of a common-emitter amplifier is approximated as the value of the collector resistor. Therefore, if you increase the resistance of the collector resistor to extremely high values, the amplifier's output impedance also increases to such high values that the amplifier is no longer capable of "matching" to any kind of practical load. In other words, it can't "drive" anything. The optimum collector load for a common-emitter amplifier would be a load that appeared to represent extremely high resistance (i.e., a high internal impedance), while simultaneously providing its own "drive current" to an external load. This is the effect that a constant-cur rent source provides to a common emitter amplifier.

The easiest way to understand the desirable effects of active loading is to imagine Q8 in FIG. 11 to be a variable resistor (signal voltages applied to the base of Q8 will cause the collector-to-emitter impedance of Q8 to vary proportionally, so this imaginary perspective is not in error). Also, consider Q7 and the amplified diode circuitry to be out of the circuit for purposes of discussion. Imagine that the variable collector-to-emitter impedance provided by Q8 can range from the extremes of 100 ohms to 6 Kohms, which is a reasonable estimation. Assume that throughout this entire drastic range of collector-to-emitter impedance variations, the constant current source of Q10/Q11 maintains a constant, regulated current flow of 6.7 milliamps. The only way that the same "constant current" effect could be accomplished using "passive" loading would be to make the collector load resistance for Q8 extraordinarily high, so that Q8's collector-to-emitter impedance variations would be negligible. In other words, if the Q10/Q11 constant-current source were replaced with a 1-Mohm resistor (i.e., 1,000,000 ohms), the 100-ohm to 6-Kohm collector-to-emitter impedance variations of Q8 would have a negligible effect on the circuit current (i.e., the dominant current-controlling factor would be the very high resistance of the collector resistor). However, to accomplish a 6.7-milliamp current flow through a 1-Mohm resistor, you would need a 6700-volt power supply! Consequently, the "active load" provided by the constant current source of Q10/Q11 looks like a very high-resistance collector load to Q8, which improves the gain and linearity factors of this voltage amplifier stage.

Output Stage Considerations

Note that the collector circuit of Q8 must provide the signal drive current to the bases of Q14 and Q15 (through resistors R13 and R14). The technique of active loading, as previously described, also aids in providing some of the signal drive current to Q14 and Q15, which makes the output impedance of the Q8 voltage amplifier stage appear reasonably low. This desirable condition also improves the linearity of the voltage amplifier stage.

Transistors Q14, Q15, Q16, and Q17 in FIG. 11 perform the same function as described previously for transistors Q4, Q5, Q6, and Q7 in the 12-watt amplifier of FIG. 8. Collectively, these transistors constitute a "near-unity voltage gain, current amplifier." They are biased in a class B mode, meaning that Q14 and Q16 provide current gain for the positive half-cycles of the AC signal, while Q15 and Q17 provide current gain for the negative half-cycles of AC signal.

Capacitor C9 in FIG. 11 aids in eliminating a type of distortion known as switching distortion. High-power bipolar transistors have to be manufactured with a rather large internal crystal geometry to be able to handle the higher current flows. The larger crystal area increases the inherent "junction capacitance" of power transistors. The higher junction capacitance can cause power transistors to be reluctant to turn off rapidly (i.e., when they are driven into cutoff). If one output transistor is so slow in turning off that it is still in conduction when its complement transistor turns on, a very undesirable condition occurs which is referred to as cross conduction (i.e., both complementary output transistors are on at the same time). Cross-conduction creates switching distortion. Note how C9 is connected across both base terminals of the output transistors in FIG. 11.

Under dynamic operating conditions, C9 acts to neutralize charge carriers in the base circuits of the output transistors, thereby increasing their turn off speed and eliminating switching distortion. Although cross-conduction creates undesirable switching distortion in audio power amplifiers, it can be destructive in various types of motor control circuitry utilizing class B outputs. Consequently, it is common to see similar capacitors across the base circuits of high-current class B transistor outputs utilized in any type of high-power control circuitry.

Diodes D4 and D5 in FIG. 11 can be called by many common names. They are most commonly known as freewheeling, catching, kickback, or suppression diodes, although I have heard them called by other names.

Their function is to suppress reverse-polarity inductive "kickback" transients (i.e., reverse electromotive force pulses) that can occur when the output is driving an inductive load. As you may recall, this is the same function performed by D1 in Fig. 7-9 of Section 7.

Most high-quality audio power amplifiers will incorporate a low inductance-value air-core coil on the output, such as L1 in FIG. 11. L1 has the tendency to negate the effects load capacitance, which can exist in certain speakers or speaker crossover networks. Even small load capacitances can affect an audio power amplifier in detrimental ways, because it has a tendency to change the high-frequency phase relationships between the input and negative-feedback signals. Compensation is incorporated into an amplifier to ensure that its gain drops below unity before the feedback phase shift can exceed 180 degrees, resulting in oscillation. However, if the amplifier's speaker load is somewhat capacitive, it can cause a more extreme condition of phase lag at lower frequencies, leading to a loss of stability in the amplifier. At higher frequencies, the inductive reactance of L1 provides an opposing force to any capacitive reactance that may exist in the speaker load, thereby maintaining the stability of the amplifier under adverse loading conditions. The inductance value of L1 is too low to have any effect on frequencies in the audio bandwidth. R26, which is in parallel with L1, is called a "damping" resistor. R26 reduces "ringing oscillations" that can occur at resonant frequencies of the speaker and L1 (resonance will be discussed in a later section).

Short-Circuit Protection

The circuit consisting of Q12, D2, R16, R15, and R19 provides short-circuit protection during positive signal excursions for output transistor Q16. Likewise, Q13, D3, R17, R18, and R20 provide short-circuit protection for Q17 during negative signal excursions. Since these two circuits operate in an identical fashion but in opposite polarities, I will discuss only the positive protection circuit, and the same operational principles will apply to the negative circuitry.

Referring to FIG. 11, imagine that some sort of adverse condition arose, creating a direct "short" from the amplifier's output to circuit common.

Note that all of the speaker load current flowing through Q16's emitter must also flow through R22. As the Q16 emitter current tries to increase above 3 amps, the voltage drop across R22 (which is applied to the base of Q12 by R19) exceeds the typical 0.7-volt base-emitter voltage of Q12, causing Q12 to turn on. When Q12 turns on, it begins to divert the drive cur rent away from the base of Q14, which, in turn, decreases the base drive current to Q16. The more the emitter current of Q16 tries to increase, the harder Q12 is turned on, thereby limiting the output current to a maxi mum of about 3 amps, even under a worst-case short-circuit condition. In other words, under an output short-circuit condition, this protection circuit behaves identically to the other current-limit protection circuits you have examined thus far. The only difference is the inclusion of Q14, which serves only to "beta-enhance" the output transistor Q16.

In addition to protecting the amplifier of FIG. 11 from short-circuit conditions at the output, the protection circuit also provides a more complex action than does simple current limiting. The voltage divider of R15 and R16, in conjunction with the dynamic action of the output rail (which is connected to the emitter of Q12), causes the current-limit action of the protection network to be "sloped." In other words, it will limit the maximum current to about 3 amps when the output rail voltage (i.e., the amplifier's output voltage) is 0 volt. However, as the output rail voltage increases in the positive direction, the protection circuitry will allow a higher maximum current flow. For example, if the amplifier's output voltage is at 20 volts, the protection circuit may allow a maximum of 5 amps of current flow. This type of sloped protection response allows the maximum output power to be delivered to a speaker system, while still providing complete protection for the output transistors. And finally, diode D2 is placed in the collector circuit of Q12 to keep Q12 from going into a state of conduction during negative signal excursions.

Now that the output protection circuitry has been explained, it becomes easier to understand the function of Q7. Q7 provides current limit protection for transistor Q8. If you refer back to Fig. 8d, (section 6) you will note that the Q7/Q8 combination of FIG. 11 is nearly identical; the only difference is in the value of the emitter resistors. To understand why Q8 in FIG. 11 needs to be current-limited, imagine that the amplifier out put is short-circuited. If a positive signal excursion occurs, Q16 will be protected by Q12 turning on and essentially shorting (short-circuiting) Q14's base current to the output rail. In this situation, the drive current to Q14's base is actually originating in the constant-current source of Q10/Q11. Since Q10 and Q11 are already in a state of current limit, Q12's shorting action is of no consequence. However, if a negative signal excursion occurs, Q13 will turn on to protect Q17, shorting Q15's drive cur rent to the output rail. During negative signal excursions, the drive current for Q15 is the collector current of Q8 through the negative power supply rail. Without the current-limit protection provided by Q7, Q8 would be destroyed as soon as Q13 turned on, because there would be nothing in the collector circuit of Q8 to limit the maximum current flow.

Negative Feedback Negative feedback in the FIG. 11 amplifier is pro vided by R8, R10, and C6. As you recall from previous studies of the common emitter amplifier, a "bypass" capacitor could be placed across the emitter resistor, causing the DC gain response to be different than the AC gain response. C6 serves the same purpose in this design of Fig. 11. The AC voltage gain response of this design is determined by the ratio of R10 and R8, since C6 will look like a short to circuit common to AC signal voltages. With the values shown, it will be about 31 [ (R10 _ R8) divided by R8 _ 31.3]. In contrast, C6 will look like an infinite impedance to DC voltages, placing the DC voltage gain at unity (i.e., 100% negative feedback). Simply stated, the negative-feedback arrangement of FIG. 11 provides the maximum negative DC feedback to maintain the maximum accuracy and stability of all DC quiescent voltages and currents. Regarding AC signal voltages, it provides the necessary volt age gain to provide the maximum output power to the speaker load.

Note that C6 is in parallel with D1. If a component failure happened to occur that caused a high negative DC voltage to appear on the output of the amplifier, a relatively high reverse-polarity voltage could be applied across C6, resulting in its destruction. Diode D1 will short any significantly high negative DC levels appearing across C6, thereby protecting it.

Other Concerns

Finally, a few loose ends regarding FIG. 11 have not been discussed. R24 and C10 make up the Zobel network, which serves the same purpose as the Zobel network previously explained in reference to FIG. 8. C11, C12, C4, and C5 are all decoupling capacitors for the power supply rails. You'll note that the "signal common" connection at the input of the amplifier is not connected to circuit common. This signal common point must be connected to circuit common, but it is better to run its own "dedicated" circuit common wire back to the power supply, thereby eliminating electrical noise that might exist on the "main" circuit common wire connecting to all of the other components.

Now that you have completed the theoretical analysis of the complex audio power amplifier of FIG. 11, as well as the other circuits illustrated in this section, you may be wondering how an in-depth understanding of these circuits will be beneficial in your continued involvement in electronics. Allow me to explain. If you happen to be interested in audio electronics, obviously this section is right down your alley. But if your interest lies with robotics or industrial control systems, don't despair! If you look at a schematic diagram of a servo motor controller, you'll say, "Hey! This is nothing but a customized audio power amplifier." If you look at the schematic of an analog proportional loop controller, as used extensively in industry, you'll say, "Hey! This is nothing but an adjustable gain amplifier with the means of obtaining negative feedback from an external source." If you look an the "internal" schematic of an operation amplifier IC, you'll say, "Hey! This is exactly like an audio power amplifier with external pins for customizing the feedback and compensation." (In reality, modern audio power amplifiers are nothing more than large "discrete" versions of operational amplifiers-you'll examine operational amplifiers in more detail in Section 12.) Expectedly, there will be times when you look at the schematic of an electronic circuit and say, "Hey! I have no idea what this is." However, if the circuit is functioning with analog voltages, currents, and/or signals, it will consist of some, or all, of the following circuit building blocks: single-stage or multistage amplifiers, differential amplifiers, current mirrors, constant-current sources, voltage reference sources, protection circuits, power supplies, and power supply regulation and protection circuits. It will also probably utilize some or all of the following techniques: beta enhancement, negative feedback, active loading, compensation, filtering, decoupling, and variations of gain manipulation. The only exception to these generalities will be in circumstances where the electronic circuitry consists primarily of linear integrated circuits performing the aforementioned functions and techniques. The main point is that if you have a pretty good understanding of the linear circuitry discussed thus far, you have a good foundation for comprehending the functions and purpose of most linear circuitry.

FIG. 12 Top-view silkscreen of the 50-watt professional-quality amplifier.

FIG. 13 Top-view silkscreen of the 50-watt amplifier showing top view of the PC board artwork.

FIG. 14 Bottom-view reflected artwork for the 50 watt professional quality amplifier.

Constructing the Professional-Quality Audio Power Amplifier

If you would like to construct the professional-quality audio power amplifier as described above, I have provided a professional layout design (Figs. 12 and 13) and the PC board artwork (FIG. 14) for you to copy.

See Section C for a full-size copy of FIG. 14, to be used in your project. I have constructed amplifiers more involved than this using the PC board "fabrication by hand" process described earlier, so it is not overly difficult to accomplish if you have a little time and patience. Also, I have constructed similar amplifiers on perfboard, but this process is really time-consuming, and you are more prone to make mistakes. Naturally, the preferred method of construction is to fabricate the PC board as illustrated using the photographic technique. Regardless of the method you choose, the amplifier circuit is reasonably forgiving of wire or trace size, and component placement. If you would like to avoid the task of PC board fabrication all together, a complete kit (including etched and drilled PC board) is available from SEAL Electronics (the contact information is provided in Section B).

The parts list for the professional-quality amplifier is provided in Table 2. Most of these components are available at almost any electronics supply store, but a few details need to be highlighted. If you want to go first class on this project, you can use 1% metal film resistors for all of the 1/2-watt resistors listed. Metal film resistors will provide a little better signal-to-noise performance, but it is probable that you will not be able to hear the difference-standard 5% carbon film resistors will pro vide excellent performance. Output inductor L1 is easily fabricated by winding about 18 turns of either 16 or 18 AWG "magnet wire" around a 1/2-inch form of some type (I use an old plastic ink pen that happens to be 1/2 inch in diameter). The exact inductance of L1 is not critical, so don't become unduly concerned with getting it perfect.

TABLE 2 Parts List for 50-Watt Professional-Quality Audio Amplifier

All of the transistors used in this amp are readily available from MCM Electronics or Parts Express (contact information is provided in Section B of this textbook), as well as many other electronic component suppliers.

The fuse clips used in the PC board design are GMA types (four needed) with two solder pins that insert through the holes in the PC board. Pay particular attention to the orientation of transistors Q14 and Q15 when soldering them into the PC board-it is easy to install them backward.

You should mount a small TO-220-type heatsink to Q14 and Q15. Almost any type of TO-220 heatsink will be adequate, since the power dissipation of Q14 and Q15 is very low. A common style of heatsink that will be ideal is U-shaped, measuring about ¾ x 1 inch (w x h) (width x height), constructed from a single piece of thin, stamped aluminum.

You will need one medium-sized heatsink for the output transistors (i.e., Q16 and Q17) and the Vbias transistor (Q9). Just to provide you with a rough idea of the size of this heatsink, the commercial models of this amplifier use a heatsink that measures 6 x 4 x 1 1/4 (l x w x d) (length x width x depth) (if you're lucky enough to find a heatsink accompanied with the manufacturer's thermal resistance rating, the specified rating is 0.7°C/watt). However, if this is all confusing to you, don't worry about it.

If you can find a heatsink that is close to the same dimensions and is designed for the mounting of two TO-3 devices, it should suffice very well. A few examples of larger heatsink styles are provided in FIG. 15; the square-shaped Wakefield type in the forefront of the illustration is a good choice for this amplifier project. If the heatsink is manufactured so that it can be mounted vertically, there is space provided for mounting it directly to the PC board. If not, you can run longer connection wires out to an externally mounted heatsink (the length of the transistor connection wiring in this design is not critical). Q9 should be mounted in close proximity to the output transistors on the same heatsink so that it will thermally track the temperature of the output transistors. The connection wiring to Q16, Q17, and Q9 can be made with ordinary 20- to 22-AWG stranded, insulated hookup wire. Be sure to double-check the accuracy of your wiring according to the silkscreen diagrams provided-it's very easy to get the transistor leads confused on the TO-3 devices.

FIG. 15 Some examples of larger heatsink styles.

Testing, Setup, and Applications of the Professional-Quality Audio Power Amplifier

Once you have finished the construction of the amplifier, I suggest that you use the schematic (FIG. 11) and accompanying illustrations (Figs. 12 to 14) to recheck and then double-recheck your work. The amplifier design of FIG. 11, like all modern high-quality audio power amplifier designs, is classified as direct-coupled. This term simply means that all of the components in the signal path are directly coupled to each other, without utilizing transformers or capacitors to couple the signal from one stage to the next. Direct-coupled amplifiers provide superior performance, but because of the direct-coupled nature of the interconnecting stages, an error in one circuit can cause component damage in another (a damaged electronic circuit causing subsequent damage in another electronic circuit is referred to as collateral damage). Therefore, it is extremely important to make sure that your construction is correct; one minor mistake is all that it takes to destroy a significant number of components.

The amplifier of FIG. 11 is designed to operate from a "raw" dual polarity power supply providing 42-volt rail potentials. Under these conditions, it is very conservatively rated at 60 watts rms of output power into typical 8-ohm speaker loads. The design also performs very well with dual-polarity rail voltages as low as 30 volts, but will provide proportionally lower output power. FIG. 16 illustrates a good power supply design for this amplifier. As shown, it will provide dual-polarity rail potentials of about 38 volts. This equates to a maximum power out put of a little over 50 watts rms.

FIG. 16 Recommended power supply for 50 watt audio amplifier.

There are a few details that bear mentioning in the power supply design of FIG. 16. The 0.01-uF capacitor connected across the power switch (SW1) is provided to eliminate "pops" from the amplifier when the switch is opened. Note that when SW1 is open, the full 120-volt AC line voltage will be dropped across the 0.01-uF capacitor, so the capacitor should be rated for about 250 volts (the peak voltage of ordinary 120-volt AC residential power is about 170 volts). R1 and R2 are called bleeder resistors, and they are incorporated for safety reasons. Referring back to the amplifier schematic of FIG. 11, imagine that a failure had occurred causing the rail fuses (F1 and F2) to blow. With the rail fuses blown, the filter capacitors of the power supply (i.e., C3 and C4 in FIG. 16) do not have a discharge path. Therefore, they could maintain dangerous electrical charges for weeks, or even months. If you attempted to service the amplifier without recognizing that the filter capacitors were charged, you could suffer physical injury from accidental discharges. R1 and R2 provide a safe discharge path for the filter capacitors to prevent such servicing accidents.

R1 and R2 can be 5.6-Kohm, 1/2-watt resistors for this power supply. C1 and C2 are typically 0.1-uF ceramic disk capacitors providing an extra measure of high-frequency noise filtering on the power supply lines. They are seldom required, and may be omitted if desired. The capacitance value for the filter capacitors (i.e., C3 and C4) should be about 5000 _F (or higher), with a voltage rating of at least 50 WVDC.

While on the subject of power supply designs, the power supply illustrated in Fig. 6 (sec. 6) will also function very well with this amplifier design.

You may want to consider this approach if you have difficulty finding a 50-volt center-tapped, 3-amp transformer (if you decide to use the Fig. 5-6 circuit for the amplifier power supply, it would be a good idea to install bleeder resistors as illustrated in the FIG. 16 design).

After connecting the FIG. 11 audio power amplifier to a suitable dual-polarity power supply, adjust P1 to the middle of its adjustment range and apply operational power. If either one of the rail fuses blows (i.e., F1 or F2), turn off the power immediately and correct the problem before attempting to reapply power. If everything looks good, and there are no obvious signs of malfunction, use a DVM to measure the DC voltage at the amplifier's speaker output (i.e., the right side of L1). If the amplifier is functioning properly, this output offset voltage should be very low, with typical values ranging between 10 and 20 mV.

If the output offset voltage looks good, measure the DC voltage from the emitter of Q16 to the emitter of Q17, and adjust P1 for a stable DC voltage of 47 mV. Once this procedure is accomplished, the amplifier is tested, adjusted, and ready for use.

From a performance perspective, the input sensitivity of this power amplifier design is approximately 0.9 volts rms. The rms power output is a little above 50 watts using the power supply design of FIG. 16. The signal-to-noise ratio (SNR) should be at or better than _90 dB, and the total harmonic distortion (% THD) should be better than 0.01%. The amplifier can drive either 8- or 4-ohm speaker load impedances. Its design is well suited for domestic hi-fi applications, and since it includes excellent short-circuit and overload protection, it is also well suited for a variety of professional applications, such as small public address systems, musical instrument amplifiers, and commercial sound systems.

Speaker Protection Circuit

I decided to include this project as a final entry in this section, because it is a good example of how various circuit building blocks can be put together to create a practical and functional design, and also because it represents the last electronic "block" in an audio chain starting from the preamplifier and ending at the speaker system.

There are several problems associated with high-performance direct-coupled amplifier designs, such as the design illustrated in Fig. 11. If one of the output transistors (Q16 or Q17) failed in the amplifier of FIG. 11, it is probable that one of the DC power supply rails could be shorted directly to the speaker load (bipolar transistors normally fail by developing a short between the collector and emitter).

DC currents are very destructive to typical speaker systems, so it is very possible that a failure in an audio power amplifier could also destroy the speaker system that it is driving. Another problem relating to audio power amplifiers is their power-up settling time (often called turn-on transients). When operational power is first applied to most higher-power audio amplifiers, they will exhibit a temporary period (possibly as long as 100 milliseconds) of radical output behavior while the various DC quiescent voltages and currents are "settling" (balancing out to typical levels). This settling action usually produces a loud "thump" from the speakers during power-up, which is appropriately referred to as "turn-on thumps."

FIG. 17 Speaker protection circuit.

FIG. 17 illustrates a very useful, practical, and easy-to-construct circuit that can be implemented into almost any audio system to eliminate the two aforementioned problems associated with direct-coupled audio power amplifiers; (1) it automatically disconnects the speaker system from the audio power amplifier if any significant level of DC current is detected at the amplifier's output; (2) it delays connection of the amplifier to the speaker for about 2 seconds on power-up, thereby eliminating any turn-on thumps; and (3) it provides a visual "status" indication of the amplifier's operation utilizing a typical LED indicator.

Operational power for the FIG. 17 protection circuit is normally obtained from the secondary of the power transformer used in the power supply for the audio power amplifier (the current drain of this circuit is very low, so it shouldn't represent any significant load to any large power transformer used to provide high currents to an audio power amplifier). The connection terminal extending from the negative side of C3 should be connected to the center-tap of the power trans former-the transformer center-tap will always be the circuit common point (see FIG. 16). The connection point extending from the anode of D1, labeled "ac in" in FIG. 17, is connected to either "hot" side of the transformer secondary (e.g., either 25-volt secondary wire coming out of the 50-volt center-tapped transformer of FIG. 16).

Referring to FIG. 17, D1 and C3 make up a simple half-wave rectifier circuit, providing a rectified and filtered DC voltage from the AC operational power obtained from the secondary of the power transformer.

For example, if you connected this circuit to the 50-volt transformer secondary of FIG. 16 (as described previously), the DC voltage at the cathode of D1 should be approximately 35 volts (25 volts _ 1.414 _ 35.35 volts DC). ZD1 and R1 form a simple zener voltage regulator circuit. Typically, R1 is a 220-ohm, 1-watt resistor and ZD1 is a 24-volt, 1-watt zener diode.

D3, R2, R3, C6, Q4, Q5, D9, and the relay form a time-delay relay circuit. When AC power is first applied to the AC input (i.e., the anode of D1), D3 is forward-biased only during the positive half-cycles. The positive pulses are applied to the RC circuit of R2 and C6. Because of the RC time constant, it takes several seconds for C6 to charge to a sufficiently high potential to turn on the Darlington pair (Q4 and Q5) and energize the relay. D9 is used to protect the circuit against inductive kickback spikes when the relay coil is deenergized. R3 is incorporated to "bleed" off C6's charge when the circuit power is turned off.

Transistors O2 and Q3 and their associated circuitry form the familiar astable multivibrator (discussed in previous sections of this textbook).

The values of C4 and C5 (typically 1 _f) are chosen to cause the circuit to oscillate at about 2 hertz.

The protection circuit of FIG. 17 is designed to accommodate "two" audio power amplifiers, since most domestic hi-fi applications are in stereo. The outputs of the audio power amplifiers are connected to the "right in" and "left in" connection points, while the right and left speaker systems are connected to the "right out" and "left out" connection points, respectively. If you desire to provide protection to only one speaker sys tem (obviously indicating that you will be using only one audio power amplifier), you can delete F2 and R5, and the output relay need only be a single-pole, double-throw (SPDT) type. Fuses F1 and F2 are speaker fuses.

Their current ratings will depend on the power capabilities of your audio power amplifier and speaker system.

Since the operational power for this protection circuit is obtained from the power transformer in the audio power amplifier's power supply, operational power to this circuit will be applied simultaneously to applying power to the amplifier(s). Therefore, when the operational power is first turned on, the speakers will not be connected to the power amplifier(s) because the relay will not be energized. The relay will not energize until C6 reaches a potential high enough to turn on Q4 and Q5. This will take several seconds. In the meantime, the audio power amplifier(s) will have had sufficient time to stabilize, thereby eliminating any turn-on thumps from being applied to the speaker systems.

While C6 is charging, before Q4 and Q5 have been turned on, the astable multivibrator is oscillating, causing LED1 to flash at about 2 hertz. This gives a visible indication that the protection circuit is working, and has not yet connected the speakers to the power amplifier(s).

When the time delay has ended and Q4 and Q5 turn on, the collector of Q4 pulls the collector of Q3 low, through D8, stopping the oscillation of the multivibrator, and causing LED1 to light continuously. The relay energizes simultaneously. By staying bright continuously, LED1 now gives a visual indication that the circuit is working, and that the speakers are connected to the power amplifier(s). At this point, the protection circuit of FIG. 17 has no further effect within the amplification system unless a DC voltage occurs on one or both of the audio power amplifier outputs.

Under normal conditions, when no DC voltage is present on the out put of either of the audio power amplifiers, the amplified audio AC signal from both power amplifiers is applied simultaneously to R4, R5, C1, and C2. Because the time constant of the RC circuit is relatively long, C1 and C2 cannot charge to either polarity. In effect, they charge to the aver age value of the AC waveshape, which is (hopefully) zero. This is the same principle as trying to measure an AC voltage level with your DVM (digital voltmeter) set to measure DC volts-pure AC will provide a zero reading.

If a significant DC voltage (i.e., higher than about 1.2 volts) appears on the output of either power amplifier, C1 and C2 will charge to that volt age regardless of the polarity (note that C1 and C2 are connected with both positive ends tied together, forming a nonpolarized electrolytic capacitor-this principle was discussed when describing the FIG. 11 amplifier circuit). This DC voltage will be applied to the bridge rectifier (D4 through D7). Although it might seem strange to apply DC to a bridge rectifier, its effect in this circuit is to convert the applied DC to the correct polarity for forward-biasing Q1 (a technique often called steering). When Q1 is forward-biased, it pulls the base of the Darlington pair (Q4 and Q5) low, deenergizing the relay, and disconnecting the speaker systems from the power amplifiers before any damage can result. At the same time, the astable multivibrator is enabled once more, causing LED1 to flash, which gives a visual indication that a malfunction has occurred. The circuit will remain in this condition as long as any significant DC level appears on either of the power amplifier out puts. On removal of the DC, the protection circuit will automatically resume normal operation.

Transistors Q1 through Q5 can be almost any general-purpose NPN types; common 2N3904 transistors work very well. The contact ratings of the relay must be chosen according to the maximum output power of the audio power amplifier(s). The voltage ratings for all of the electrolytic capacitors should be 50 WVDC for most applications. However, if the operational AC power obtained from the power amplifier's power sup- ply transformer is higher than 30 volts AC, you will have to adjust most of the component values in the circuit to accommodate the higher operational voltages. The diode bridge consisting of D4 through D7 can be constructed from almost any type of general-purpose discrete diodes (e.g., 1N4148 or 1N4001 diodes), or you can use a small "modular" diode bridge instead.

Top of Page PREV.   NEXT Index HOME