Interference

In any transducer system it's important that interference be minimized. There are numerous ways to do this; the most obvious and fundamental is good system design, employing optimum wiring techniques, shielding, and impedance matching. But there are certain cases in which interference may be a problem despite all of the commonly accepted layout and design practices. This section describes a few of the ways in which special interference problems can be dealt with.

WHAT IS INTERFERENCE?

We may say that, in general, interference is any energy that is present in a system and is not desired. Alternating-current “hum” is a good example of interference because it's almost never wanted. You have probably experienced this type of interference, for ex ample, in a public-address system having inadequate shielding of the microphone cord. Two radio stations may interfere with each other; which station is “desired” and which is “interference” depends on your preference.

The method to be used for minimizing interference depends on the type of interference, the spectral characteristics, the location of the interfering source with respect to the system, and many other factors.

AC HUM

Normally, if a system has good shielding and is wisely designed to eliminate excess wiring lead lengths or ground loops, hum is not a problem. But in high-gain circuits with low-level input, only a very tiny amount of power-frequency interference can result in objectionable hum. In general, the higher the gain of an amplifier, the more likely there will be some of this type of interference.

Figure 17-1 illustrates a block diagram of a multistage, audio- frequency amplifier such as might be used for detecting very weak noises (as in a “bugging” system for spying). If power-frequency interference is introduced between the transducer (a sensitive microphone) and the first amplifying stage, the interference will be amplified by all of the succeeding stages. The result will likely be a severe hum in the output if the interference is input at point A or point B. An equal amount of interference introduced at point C or point D would cause less trouble. Clearly, wiring and shielding precautions are most important in low-level parts of a circuit.

Suppose we have a sensitive amplifier chain, such as that shown in ill. 17-1, and have taken all normal precautions against hum, but nonetheless we have interference. There are various means we might use to alleviate the problem.

ill. 17-1. Block diagram of an amplifier chain, showing various points at which Interference might enter the system. The most severe problem would occur if the interference entered at point A; less and less severe trouble would occur at B, C, and D entry points.

One method of reducing hum interference is to use a low- impedance input circuit. We might, e.g., employ a bipolar transistor rather than a FET at the first stage; or we might connect an operational amplifier for a low-input-impedance condition. When the input impedance is low, electrostatic coupling to the out side environment is minimized.

Another method of reducing hum in an audio-frequency circuit is to use a band-rejection filter or high-pass filter. Filters may be installed at the input only, or between each stage, or between selected stages. it's almost always advantageous to place a filter at the input, because amplified hum may cause reduced gain or nonlinearity of later stages in an amplifier chain. Figure 17-2 shows simple passive LC band-rejection (A) and high-pass (B) filters. The band-rejection filter is resonant at 60 Hz; the high-pass device is designed for a cutoff frequency somewhat above 60 Hz.

ill. 17-2. Passive band rejection (A) and high-pass (B) hum filters.

Still another method of reducing hum is to apply a small alternating-current signal, equal in magnitude but opposite in phase to the interference, somewhere along the amplifier chain. This can prove extremely effective in such devices as very-low-frequency (VLF) receivers and modulated-light devices. We simply construct a small probe that picks up some hum from the environment, connect it to a circuit that provides either zero phase or phase opposition (selectable) and variable gain, and apply the resulting output to the input of the amplifier chain (ill. 17-3). This type of device has the advantage that it not only helps reduce 60-Hz hum but also the transients and other interfering frequencies that are often present on power lines.

ill. 17-3: Phase-cancellation circuit for reducing interference.

The problem of alternating-current hum in a direct-current circuit is dealt with in another way. A low-pass filter, which may consist of a large-value capacitor in parallel with the input, is commonly used to bypass hum. A series-connected choke, either alone or in combination with the capacitor, can also be employed (ill. 17-4). Phase cancellation can also be used, but this method is not often seen in direct-current applications.

ill. 17-4: Filter for eliminating alternating-current interference at the input of a direct-current circuit.

OTHER EXTERNAL INTERFERENCE

Alternating-current hum is an externally generated form of interference. This means that it does not originate within the system, but it gets into the system from outside. (The ultimate source of hum is, of course, the power-plant generators.) There are other forms of externally generated interference that can affect the performance of a transducer system. Examples include sferics (atmospheric static), radio signals, and even cosmic noise. In some cases electric or magnetic fields, even the earth’s magnetic field, can adversely affect a system. Let’s look at some examples.

Sferics are present at all frequencies from sub-audible to the visible-light and ultraviolet ranges. Most sferics are generated by lightning strokes, carrying momentary currents of hundreds of thousands of amperes. Sferics, unlike alternating-current hum, are a random form of electromagnetic energy. Sferics exhibit no evident pattern, such as a repeating wave cycle. The randomness of sferics makes this type of interference more difficult to deal with than hum because there is no specific kind of filter that is unfailingly effective against it.

Meticulous attention to shielding of amplifier stages and interstage wiring is essential if we are to minimize the interference caused by sferics. But sometimes, as in a radio receiver, the sferics enter the system along with the signal. Then we have no recourse except to use a “brute force” approach. Noise limiters, noise blankers, and narrowband filters are the most often-used methods of reducing interference caused by sferics.

The noise limiter circuit simply prevents interfering signals from exceeding the amplitude of the desired signal (ill. 17-5). A noise blanker circuit shuts down the amplifier chain (at one stage) during a noise pulse (ill. 17-6). The noise blanker can be extremely effective when interference pulses have short duration, but it does not function as well for pulses of longer duration. Noise limiters are more generally effective against atmospheric static than noise blankers.

ill. 17-5. Action of noise limiter and noise blanker. At A, original signal and Interference; at B, action of noise limiter.

Narrowband filters operate by reducing the total amount of noise entering a system. The bandwidth of the filter depends on the occupied bandwidth of the desired signal. A typical amplitude-modulated (AM) radio signal occupies about 6 kHz of spectrum space for voice and 10 kHz for music; thus, we could not use a 1-kHz filter in an AM receiver. We could, however, use a 1-kHz filter for receiving Morse code (CW) signals. There is generally an optimum filter bandwidth in any situation. This bandwidth depends on the amount of noise, the strength of the signal, and the type of signal.

Another kind of external interference, similar to sferics but of a different origin, is impulse noise. Impulse noise is generated by man-made devices such as internal combustion engines, light dimmers, or anything that creates repeated arcs or sparks. Impulse noise is not random in nature and can be more effectively dealt with than sferics. If the pulses are short, a noise blanker can practically eliminate the interference. Noise limiters and narrowband filters are also useful. Sometimes impulse noise can be reduced by phase cancellation, in a manner similar to the hum-reduction method shown in ill. 17-3.

Electromagnetic fields can cause many different kinds of problems in amplifier systems. A strong radio signal can enter an audio amplifier and be rectified, resulting in the demodulated signal appearing at the output. This problem may occur despite RF bypassing of input leads, the installation of series RF chokes, and complete shielding; a transducer such as a magnetic tape-recording head can respond to a strong RF field all by itself. Radio-frequency interference to audio amplifiers can be one of the most difficult problems to solve. Audio home-entertainment systems must be well engineered to minimize the chances of interference from nearby broadcast, amateur, citizens’ band, or commercial radio stations.

A geomagnetic storm, caused by a solar flare, can disrupt radio communications and even wire communications throughout the world. A shielded-cable system is less likely to be affected than a radio link, obviously, but severe disturbances can affect even the most well-designed system. Fortunately, such occurrences are rare.

A peculiar kind of problem can result from the buildup of an electrostatic charge in a system. This is especially apt to happen in a circuit in which the impedance is very high, resulting in low rate of discharge. The electrostatic charge alters the biasing of the circuit, affecting the linearity and gain in an analog circuit and causing “latchup” in a digital system. The source of the electrostatic charge must be removed in order to solve this problem. Finding the source can be difficult. Electrostatic charge may be generated by friction, by a faulty solder joint or other electrical connection, by corrosion of switch or plug contacts, or by any of numerous other causes.

Electrostatic problems can result from internal as well as external causes. In low-level circuits the type of solder used may affect the performance because of thermally generated direct-current voltages. Some kinds of solder are more susceptible to thermal voltages than others; special solders are available for use in devices operating at microvolt or nanovolt levels.

Heat itself can be a source of interference. The random motion of atoms in any substance is a source of broadband noise. In general, the amount of noise generated in this way is directly proportional to the absolute temperature. There is very little that we can do to cool down the external environment, but internally, it's often quite practical to lower the temperature. This can be done by means of liquid gases, such as helium, which have absolute temperatures just a few degrees above zero Kelvin. In extremely low-level radio-frequency devices, such as radio telescopes, significant improvement can be had by the use of temperature-lowering apparatus.

INTERNAL INTERFERENCE

Direct-current and thermal interference can be generated either outside a system or within, as we have just seen. Certain types of interference are caused only internally, however.

Unwanted feedback can produce disastrous malfunction of an amplifier system. In the worst case it results in oscillation, sometimes at more than one frequency, in conjunction with degradation of gain and linearity. Unwanted feedback can even cause physical damage to circuit components; an example is thermal runaway in a power amplifier as a result of parasitic oscillation.

Proper circuit layout and design is the best protection against unwanted oscillation in an amplifier system. In audio-frequency circuits, shielding of inter-stage wiring is often sufficient. In radio frequency power amplifiers, neutralization is usually necessary. Two methods of neutralizing an RF amplifier are shown in ill. 17-7.

Negative feedback can also occur, causing nonlinearity and degradation of circuit gain, although this type of feedback will not result in oscillation. It should be noted that negative feedback is often employed deliberately in an amplifier, and this of course is not interference.

Internal interference can take the form of unwanted coupling between or among different parts of a circuit. e.g., in a stereo audio amplifier we might have poor isolation between the channels as a result of improper design or layout. Unwanted coupling can be practically eliminated by individually shielding different parts of a system. In a stereo amplifier, e.g., we might put the left-hand and right-hand channel circuits in entirely separate metal enclosures.

ill. 17-7: Two methods of amplifier neutralization. At A, output-circuit neutralization; at B, input-circuit neutralization.

ill. 17-8. Two acceptable grounding schemes. At A, single bus; at B, multiple bus.

Grounding schemes can literally “make or break” any amplifier system. The commonly accepted method of grounding is illustrated in ill. 17-8A. Each grounded (or common) point is connected to a single bus. A variation of this method, in which several buses are tied together at the common point, is shown in B. (Actually, the use of the word “ground” is not completely accurate in most cases. We don't connect all common circuit points to the earth itself but to a bus or chassis that may or may not ultimately have an electrical connection with the earth.)

Improper grounding can result in susceptibility to interference. A notorious problem-causer in this respect is the So-called “ground loop” (ill. 17-9). A ground loop can act as a loop antenna or inductor—magnetic coupling can take place between the ground loop and strong electromagnetic fields. In many instances, ground loops are the culprit in radio-frequency interference to an audio system. it's never necessary to have ground loops, so they should be avoided in the layout of any electronic system.

INTERFERENCE TO OTHER DEVICES

Occasionally you will find yourself the culprit, rather than the victim, in an interference case. Radio amateurs know this problem well; sometimes their signals are picked up by home-enter equipment such as high-fidelity systems or television receivers. Sometimes this kind of interference is the fault of the radio transmitter, but in many cases it occurs as a result of defective design in the home-entertainment equipment.

Proper shielding and circuit layout will minimize the amount of unwanted energy that escapes from as well as enters a system. This is intuitively obvious but often overlooked. In recent years, shielding of home microcomputer devices has been vastly improved because the high-speed digital signals can cause interference to radio and television receivers in the vicinity. But for a while this interference was a serious problem. Often a computer could not be interfaced with an amateur radio station, e.g., because the computer generated so much noise that the station receiver was rendered useless.

It is sometimes possible to compensate for generated interference, even though it's the fault of one particular device, by modifying the “victim” devices. e.g., in the case of microcomputer interference to a radio receiver, we might restrict receiver operation to those frequencies on which the interference is minimal; or we might use a noise blanker or a narrowband filter and put up with what interference remains. This is not by any means an optimal solution to the problem, but it may be all that can be done. Ideally, the problem should be corrected by attacking it at the source.

PREV: Infrared Techniques
NEXT: Pressure Transducers

Fundamentals of Transducers (all articles)