Stereo Multiplex Systems [Radio Service Training Manual (1966)]

Home | Articles | Forum | Glossary | Books

With recent approval by the FCC of a multiplexing system to permit broadcasting separate right and left channels, sales of FM stereo receiving equipment have increased rapidly. The troubleshooting technician must keep up with technological advances of this kind, and this involves the development of isolation techniques for symptoms, as well as the study of the theory of operation of these new circuits.

In order to afford space for development of isolation procedures for symptoms, the theory will be briefly covered in this section. If a more extensive coverage of theory is desired, reference should be made to texts devoted entirely to this topic.


An understanding of the receiving systems is facilitated by acquiring a working knowledge of the characteristics of the signal that is to be detected. The fact that FM transmitters are capable of modulation at supersonic frequencies has been explained before. It is this property which enables a second signal to be transmitted.

Definition of Multiplex

When a second signal is imposed on the FM carrier, the process is called multiplex. A second carrier at a frequency of 38 khz is amplitude modulated with the additional information.

The sidebands, which are in the supersonic range, are then used to frequency modulate the transmitter along with the regular program material.

Definition of Monaural or Monophonic

The regular program material containing both left and right channels is called the monaural ( or monophonic) signal and is designated as (L+R). As we shall see shortly, the (L+ R), or monophonic signal, is an important part of the stereo signal.

Definition of Compatibility

The FCC requires that stereo broadcasts must be receivable on regular FM receivers, which means that a monophonic signal must be present at all times. This also means that the regular modulation must contain both left- and right-channel sounds so that there will be no loss of program material to listeners not equipped to receive stereo. It would be impossible to satisfy this requirement if, for example, the right channel were transmitted in the usual fashion, and the left channel were multiplexed at supersonic frequencies.

The problem at the transmitter then becomes one of continuing to transmit (L+R) in the usual manner while adding a special modulation, receivable only on stereo adapters, which enables the left and right channels to be separated. It is important to recognize that at no point in the system does the left or the right channel exist separately; they are combined until they reach the dual audio channels leading to the separate speakers. When the receiver is not equipped to separate the signals, they remain combined and appear as monophonic, (L+R), sound in the single speaker.

Fig. 13-1. Formation of modulation at the transmitter.

Figs. 13-1 and 13-2 illustrate a method by which this can be done. If the left and right channels are fed to a switch at the transmitter so that when the switch is on the right-channel input it is also on the right-channel output, the waveform shown can be produced. In this waveform, the voltage at the top output terminal carries the right-channel information, and the voltage at the bottom terminal carries the left-channel information. The dotted lines represent the switching rate and indicate that the entire modulation envelope is "chopped up" at the switching frequency of 38 khz.

This chopping does not distort the sound, because the switch in the receiver is synchronized with the one at the transmitter.

It will switch to the upper terminals when the incoming wave form is positive and to the lower terminals when the waveform is negative. The output waveforms are shown in Fig. 13-2 without the 38- khz segments, because this component is removed by filters in the audio channels. If the receiver is not equipped with a multiplex detector, then the composite waveform containing (L+R) will be fed to the audio section, and both channels will appear as usual in the same speaker.

Fig. 13-2. Demodulation at the receiver.

The addition of the 38- khz synchronous detector following the FM detector is all that is necessary to convert any FM receiver into a stereo unit. Several adapters are available which can be connected between the output of the detector and the de-emphasis network. Two separate audio amplifiers are needed, of course.

This composite waveform is a bit more complicated than it appears in the drawing because it contains sidebands of the 38- khz switching frequency in addition to the (L+R) information. These sidebands will be from 23 khz to 53 khz, because they are the result of amplitude modulating a 38- khz carrier with audio ranging as high as 15 khz.

It will be recalled from the explanation of detectors in Section 2 that amplitude modulation results in sidebands which are the sum and difference of the audio and the RF frequencies. If 38 khz, for example, is modulated with 15,000 cycles, the resulting sidebands will be:

38 khz + 15 khz = 53 khz ( upper sideband)

38 khz - 15 khz= 23 khz (lower sideband)

Since the audio will range from about 50 cycles to 15 khz, the sidebands will cover the range of frequencies 15 khz above and below 38 khz. Fig. 13-3 illustrates the distribution of the signals as they appear in the modulator of the FM transmitter.

Fig. 13-3. Signal distribution in the modulator of an FM transmitter.

Notice that the regular monophonic (L+R) signals extend from 50 cycles to 15,000 cycles, and that there is another group of signals from 23 khz to 53 khz. The space between 15 khz and 23 khz is not used for program material, but the diagram shows that a 19- khz pilot signal is transmitted in this space.

This pilot signal is transmitted continuously for the purpose of synchronizing the switching rate at the receiver. It is necessary to use 19 khz for the sync instead of 38 khz because 38 khz falls within the multiplexed sidebands. The frequencies be tween 23 khz and 53 khz are labeled (L-R). Matrixing

Fig. 13-4. Creation of the (L-R) signal.

Since the (L+R) signal must be present at all times, and the left and right channels cannot be separated during trans mission, it is necessary to recover the separate channels by adding signals of opposite phase.

(L-R) is created at the transmitter, as illustrated in Fig. 13-4. This signal, which is actually a monophonic signal with the "R" voltage inverted, is used to modulate the 38- khz sub carrier, and the resulting sidebands carry the (L-R) information. Now it can be seen that the stereo system consists of sending (L+R) on the regular channel while sending (L-R) in the form of 38- khz sidebands which are inaudible with ordinary monophonic receiving equipment. The 38- khz subcarrier in the transmitter is suppressed by a balanced modulator, and only the 19- khz pilot signal is transmitted.

In the receiver, the original (L-R) signals are recovered by reinsertion of a 38- khz carrier generated locally in the receiver and controlled by the pilot. The matrix is a resistive network arranged to combine (L+R) and (L-R) as follows:

(L+R) + (L-R) = 2L and (L+R) - (L-R) = 2R


There are two popular types of stereo detectors. A third method that combines the circuitry of the other two is some times used by a few manufacturers. Because the alignment and servicing of the circuits varies with the type of demodulation employed, we shall discuss them separately and point out ways to recognize each.

The Sampling Demodulators

This method is also called wave-envelope demodulation, time-division demodulation, and synchronous switching. The block diagram in Fig. 13-5 shows the layout.

The operation of sampling demodulators depends on the principle of the synchronized switch depicted in Figs. 13-1 and 13-2. Since the signal is actually divided into segments, with the left and right channels being carried in alternate segments, it is possible to detect the channels separately.

The block diagram shows a wide-band input amplifier which takes the signal from the FM ratio detector or discriminator and amplifies all components; (L+ R), (L-R), and 19 khz. A filter feeds the 19- khz sync signal to the 38- khz generator, and the reconstituted carrier is supplied to the synchronous switch, which is merely a pair of diodes. (L+R) and (L-R) are also fed to the switch.

When the switch is perfectly synchronized with the original chopping frequency at the transmitter, the right channel will be fed to the right audio amplifier and the left channel to the left audio amplifier. The separation, however, is not perfect, and the right channel carries approximately 20 % of the left channel. This imperfection is due to slight unbalance in the switching diodes, coupled with the fact that the switching volt age is not a square wave.

Fig. 13-5. Time-division multiplex unit.

This results in less separation between the channels and destroys the illusion of stereo. This incomplete separation can be overcome by the use of negative feedback of a signal equivalent to:

- .20(L+R) or - .20L - .20R

Adding this negative-feedback voltage to the output from the left and right diodes in Fig. 13-5 gives: Output from left diode: L + .20R - .20L - .20R = .SOL

Output from right diode: R + .20L - .20L - .20R = .SOR

Note that the output from the right and left channels is now pure, but it has been reduced by 20 % . The reduction in voltage output is not particularly noticeable, however, and is well worth the sacrifice since it removes the presence of the un wanted channel and improves the separation. But this necessitates additional circuitry and adjustments.

Manufacturers have used many methods to obtain the needed 38- khz signal in the receiver. In some cases, an oscillator running at 38 khz is synchronized on every other cycle by the 19- khz pilot signal. In other models, the grid tank of the oscillator operates at 19 khz while the plate tank is tuned to 38 khz.

There are also several versions in which the 19- khz pilot is fed to an amplifier-doubler, and the resulting 38 khz is used instead of generating the signal in a local oscillator. Fig. 13-6 shows an example of this type.

The entire circuit uses only one tube, two diodes, and a transistor. The input signal divides at point ! 17 ! , and part of it is sent to the grid of the 19 -khz input amplifier via the tuned circuit, L12. The plate of this stage is also tuned to 19 khz by the resonant tank, L13. The other path from __ leads to the 38 -khz switch via two 67 -khz traps. 67 khz is the frequency of another subcarrier used for private broadcasts to special subscribers only. If this subcarrier is not removed, it may cause interference.

Amplified 19 khz appears at pin 7 of V7 and is doubled to 38 khz in the plate circuit of this second triode. Transformer L14 is the synchronous switch, having 38 khz fed into one end and (L+R) and (L-R) fed into the other end. As the diodes are switched on and off, they perform the job of separating the left and right channels. The diodes are followed by an extensive RC filter to remove the 38 -khz component from the output. This filtering is important if the output is fed to a tape recorder, because the beat between the bias oscillator in the recorder and signals near 38 khz causes chirps and squeals in the recording.

The transistor (Xl) is used to operate a lamp that glows when the 19 -khz pilot signal is present, indicating that a stereo station is tuned in. The transistor has no forward bias on the base and is therefore cut off when 19 khz is absent. Thus, the bulb has no current flowing through it. When the pilot signal appears (in the form of 38 khz in the output of the doubler), the negative half of this sine wave causes conduction of the transistor, and enough collector current flows to operate the #49 bulb.

Figure 13-6

Figure 13-7 shows an entirely different approach. V1A is a wide-band amplifier handling all incoming signals. V2 amplifies the 19 -khz pilot signal and feeds it to transformer L2. At the same time, (L+R) and (L-R) are taken off at the cathode of V2 and passed on to the detector, via R6 and C7.

The input circuit to V2B is very interesting. It consists of a full-wave rectifier with its output supplied to the grid of the tube. It will be recalled from Section 2 that a full-wave rectifier uses both halves of the input signal, and that its output contains a ripple frequency equal to twice the input frequency.

In this manner, V2B is driven by negative pulses at 38 khz.

The pulses are negative-going because of the manner in which the diodes are connected. The plate of V2B feeds the resonant circuit of L3, which operates as a switch in the usual manner.

Fig. 13-8. A transistorized switching type stereo detector.

The two diodes of V1B are supplied with the 38-khz switching pulses plus (L-R) and (L+R) to produce the left and right channels separately. The detector is followed by filters.

Fig. 13-8 shows a transistorized version of the time-division, or switching-type, multiplex detector. The input signal is di vided at the input terminal, and X2 handles the composite signal feeding the diodes. Xl has a 19 -khz tank in its collector circuit, the output of which is used to synchronize the 19 -khz oscillator formed by the collector-to-emitter feedback through the coils of L2. L3 is the switch that is tuned to 38 khz and drives the diodes to produce the left and right channels separately. The separation control adjusts the level of the composite signal to correspond with the level of the 38 khz for optimum separation.

Fig. 13-9. A bridge-type stereo detector.

The circuit includes an inverse feedback to improve the separation, as discussed earlier. L4 and its associated capacitors form a low-pass filter that passes only (L+ R). Also, it will be noted that the signal taken from the collector of X2 is in the opposite phase from the one taken at the emitter.

With the voltage divider formed by the capacitors, the resulting signal is -.2(L+R), which is correct for cancelling the error inherent in the separation with time-division circuits.

The main signal is taken from the emitter because of the better frequency response obtainable with the emitter-follower circuitry.

The final time-division circuit to be discussed is shown in Fig. 13-9. It is essentially the same as the others, except for the bridge-type detector using four diodes. This is commonly done to eliminate the 38 -khz carrier from the output. Part A of Fig. 13-10 shows the bridge circuit redrawn to clarify the input and output connections. Part B shows the left side only for the time when there is no composite stereo signal applied.

In this drawing, the 38 -khz generator is represented as a center tapped battery because we are considering only instantaneous voltages. M6 and M7 are shown as 100K resistors. The channel-I output is shown as RL, and it can be seen that there will be no voltage across it at this time, since it is connected be tween two points of equal potential. This is the balanced condition showing how the 38 -khz carrier is kept out of the output.

This balanced condition will be changed when a voltage is applied to the composite-signal input. At the instant depicted in part A, a negative voltage is applied to the cathode of M6 and to the anode of M7. The resistance of M7 is therefore increased, and the resistance of M6 is reduced. The drawing in part C shows the new relationships where the resistance of M6 is reduced to 50K, and the resistance of M7 is increased to 150K. The resultant change in voltages across the diodes leaves a net voltage of 2.5 volts across RL. Similar changes occur on the other side of the bridge, resulting in an output signal in either channel which consists of the original stereo information without any of the 38 khz.

Fig. 13-10. The circuit of Fig. 13-9 redrawn for clarity.

Frequency-Division Type Multiplex Detectors

In contrast to the switching- or sampling-type detectors which we have been discussing, are the frequency-division detectors which handle the sum and difference signals separately. In Fig. 13-11 is shown the circuit of a frequency-division multiplex unit in which the input is applied to a wide-band amplifier having three outputs. Output 1 feeds (L+R) directly to the output of the diodes. Output 2 extracts (L-R) by means of the filter and feeds this to the input of the diodes where it will be combined with the oscillator signal. Output 3 places the composite signal on the grid of a tube, the output of which is tuned to 19 khz. This 19 -khz signal is used to control the 19-khz oscillator whose plate tank is tuned to 38 khz.

The (L-R) signal fed to the diodes consists of sidebands of the suppressed 38-khz carrier. These sidebands extend from 23 khz to 53 khz, and are the sum and difference of the 38 -khz and the audio frequencies, as explained before. When the side bands are recombined with 38 khz, the beat will be the original audio frequencies. It is this beat (the difference between the sideband frequency and 38 khz) which appears at the upper cathode and lower plate of the diodes in Fig. 13-11. The original (L-R) is at the cathode, and an inverted version, -(L-R), is at the plate. -(L-R) is the same as (-L+R). When these two diode outputs are combined with (L+R), the left and right channels will be separated, as shown in the figure. It is emphasized again that at no point before the diodes does the left or right signal exist separately; they are always combined into the monaural signals, (L+R) or (L-R).

Fig. 13-11. Frequency-division multiplex.

Fig. 13-12. Transistorized frequency-division multiplex.

A transistorized version of a frequency-division type adapter is shown in Fig. 13-12. X1 has three outputs-one to the 19 -khz tank, one to the (L+R) line leading to the output of the diodes, and one which passes (L-R) to the input of the diodes. An emitter follower is used for the (L-R) because of the better frequency response.


Despite the elaborate instructions furnished by manufacturers, the alignment of multiplex equipment is very simple. In general, the following must be accomplished, preferably in this order:

1. The 67-khz trap must be tuned to reject any 67-khz interference.

2. The (L-R) bandpass filter leading to the diodes must be tuned so that it has a flat response from 23 khz to 53 khz.

3. The (L+R) low-pass filter must be tuned to pass all signals from 50 hz to 15 khz, and to reject 19 khz and above.

4. The oscillator output tank must be tuned to 38 khz and synchronized with the pilot signal.

5. Adjustment of the separation control must be made.

Tuning the 67-khz Trap If this trap is included in the circuit, it may be identified from a study of the manufacturer's data and pictorial dia grams. Once the proper tuning slug is found, it is adjusted in the following manner:

Connect an audio oscillator, accurately tuned to 67 khz, to the input terminal of the wide-band amplifier. Set the oscillator signal for about 1 volt rms. Connect an oscilloscope or a high-quality AC VTVM to the output of the filter, and adjust the slug for minimum output.

Tuning the (L-R) Bandpass Filter

Tune an audio oscillator to 38 khz, and connect it to the input of the wide-band amplifier. Connect an oscilloscope or AC VTVM to the output of the filter, which is usually the input to the diodes. To avoid overloading the circuits, set the output attenuation of the audio oscillator for the minimum signal that gives a usable output indication.

First, adjust the tuning slug for maximum output at 38 khz.

Then, check the output at 23 khz and at 53 khz-it should be not less than one-half the amplitude at 38 khz. Finally, check the output at 15 khz-it should be minimum at this frequency.

Tuning the (L+R) Low-Pass Filter

Tune an audio oscillator to 19 khz, and connect it to the input of the wide-band amplifier. Connect an oscilloscope or AC VTVM to the output of the filter (this will be at the junction of the two resistors across the outputs of the diodes). To avoid overloading, set the attenuator of the audio oscillator for the minimum signal that gives a good output indication.

First, adjust the tuning slug for minimum at 19 khz. Then, check for an output at 15 khz, which should be much greater.

The output at frequencies from 15 khz down to about 50 cycles should be fairly constant.

The tuning of the filters just explained is required for frequency-division type units only. The more popular switching units do not separate (L-R) and (L+R), and are therefore simpler to set up. The following adjustments must be made on all types of multiplex units.

Synchronizing the 38 -khz Oscillator to the Pilot Signal

Tune in a station that is broadcasting stereo. Stereo signals can be identified by viewing the waveform at the output of the wideband amplifier, using a scope with the sweep set to lock in a 19 -khz sine wave. The audio program material will be seen as constantly changing variations superimposed on the 19 -khz sine wave, as shown in Fig. 13-13. During quiet moments in the program, only the sine wave will appear on the scope, as in Fig. 13-14.

Fig. 13-13. Audio program material super• F imposed on the 19 -khz pilot signal.

Fig. 13-14. Pilot sine wave viewed during quiet moment in the program.

Fig. 13-15. Signals in phase.

Fig. 13-16. Signals 180° out of phase.

Fig. 13-17. Signals approximately 90° out of phase.

Fig. 13-18. Horizontal input frequency is twice that of the vertical input.

Fig. 13-19. The 19 -khz and 38 -khz signals slightly out of phase.

Fig. 13-20. Figure "8" pattern with program material present.

When it has been determined that the station is broadcasting stereo, the scope can be used to set the phase of the 19 -khz circuits. Switch the scope to external horizontal input so that separate signals can be fed into the vertical and horizontal amplifiers of the scope. Feed the signal from the grid of the wide-band amplifier, or the 19 -khz amplifier if one is used in the unit, into the vertical amplifier, and feed the signal from the grid of the 19 -khz amplifier or oscillator into the horizontal amplifier. When signals of identical phase are fed to these two inputs, the resulting pattern on the scope will be a straight line sloping upward to the right. This type of pattern is shown in Fig. 13-15. Differences in the phase of the two signals affect the line as shown in Figs. 13-16 and 13-17. To make sure that the 19 -khz circuit is in phase with the pilot signal, it is only necessary to adjust the 19 -khz tank until a straight line appears on the scope.

The 38-khz tank can now be adjusted by connecting the horizontal-input scope lead to the 38-khz input to the diodes. When the signal on the horizontal deflection plates of the scope is exactly twice the frequency of that on the vertical deflection plates, a figure "8" will be displayed, as shown in Fig. 13-18.

When the 38-khz tank is slightly out of phase with the 19-khz pilot signal, the pattern shown in Fig. 13-19 will appear on the scope. The scope pattern with program material present is shown in Fig. 13-20. If the 38-khz adjustment seems to tune broadly, it should be adjusted for maximum horizontal width of the figure "8." Most manufacturers recommend that the oscillator signal should be at least 3 times the amplitude of the (L-R). This can be checked by switching the scope back to normal internal sweep and checking the peak-to-peak voltages of the two signals separately. (L-R) can be measured at the input to the diodes if the oscillator is disabled by shorting its grid to its cathode. The oscillator signal can be measured at the same point during a quiet moment in the program.

One precaution which must be taken in the above procedure is to listen to the program material to be sure that the right channel is appearing in the right speaker, and the left channel in the left speaker. Proper left and right channels can be identified by listening to the announcements made between musical selections, because these are nearly always made on the left channel only. If the speakers are reversed, this is an unavoidable error resulting from the fact that the 38 khz can by synchronized by the 19 khz in two different phases; the only difference on the scope display is that the figure "8" is upside down. The difficulty can be easily corrected if the inputs to the audio amplifiers can be reversed. Or, if necessary, the phase can be corrected by adjusting the 19 -khz tuning until the 38 khz is locked in the opposite phase.

It may be necessary to work back and forth between these adjustments several times to achieve perfect results. It is very important to note that the final adjustments must be made with the vertical input of the scope connected to the grid of the wide-band amplifier, and the horizontal input connected to the 38-khz input to the diodes. With these connections, un-desired phase shifts in the wide-band amplifier and 19-khz circuits will be compensated for, since the final oscillator output is being compared with the original 19-khz pilot signal. If the vertical input to the scope is connected after the grid of the wide-band amplifier, certain undesirable phase shifts may have already occurred, and the oscillator signal will not be compared to the true pilot signal.

The adjustment of the separation control, which will be described next, often causes a change in the phase of the 19 -khz signal because the control changes the load resistance of the wide-band amplifier. It is sometimes necessary to retouch the 19-khz adjustments after the separation control is adjusted.

Adjusting the Separation Control

Before making this adjustment, it is important to under stand the meanings of the terms balance and separation. "Balance" refers to the volume level of the two audio channels and, for good stereo, the gain controls should be adjusted so that equal volume appears in both channels while feeding the same signal (a monaural program) into both audio inputs. Likewise, the tone controls should be adjusted for identical tone quality in both channels.

The term "separation" refers to an adjustment of the amplitude of (L+R) applied to the demodulators so that it exactly equals the amount of (L-R) in the case of frequency-division units, or is equal to about one-third of the oscillator voltage in the switching type. The control to be adjusted usually consists of nothing more than a variable resistance in series with the (L+R). One way to adjust this control on the frequency-division units is to tune in a stereo station and arrange the speakers so they are equidistant from each ear of the operator at the time when he is in a position to operate the control. Turn the control fully clockwise (maximum resistance in the (L+R) line), which will render only (L-R) at the speakers. This is a good time to check for perfect synchronization of the 38 khz.

A slight error in the adjustment of the 19 -khz tank results in distortion of the (L-R), which can be heard when the (L+R) is removed.

The sound will be monaural-with the same signal in both speakers-and it will appear to come from a point directly overhead. As the control is turned to increase the (L+ R), the sound will appear to separate into left and right channels, and as the control is advanced still farther, it will again be come monaural and appear overhead. The control is set at the point that gives the best separation.

Another way to set the separation control, which is more accurate because it does not depend on the amount of separation in the program material, is to listen for an announcement which is made entirely on the left channel. At this time, turn the volume to minimum on the left-channel audio system ( or disconnect the wire from the input or from the speaker) and listen on the right channel only. If the separation is perfect, no sound will be heard from the right channel when the announcement is entirely on the left channel. Adjust the separation control to "null out" the sound as much as possible. Multiplex detectors do not have 100% separation, so some small part of the left channel will still be heard in the right channel, even with the best possible adjustment of the separation control.

Another way to adjust the separation control, preferred by some experts, employs the scope. The patterns shown in Fig. 13-21 were obtained with the vertical input connected to the left-channel output of the demodulators and the horizontal input connected to the right-channel output. The vertical and horizontal-gain controls were preset to give equal gain in both amplifiers of the scope.

(A) (L-R) only. (B) (L+R) only. (C) Separation control advanced to mix (L+R) with (L-R). (D) Maximum separation.

Fig. 13-21. Scope patterns showing the adjustments of the separation control.

The trace at Fig. 13-21A was obtained with the separation control fully clockwise-that is, with no (L+R) present. In this condition the signals appearing at the output terminals are (L-R) and -(L-R). Since these are 180° out of phase, the resultant on the scope is a straight line sloping upward to the left.

Fig. 13-21B shows the resultant when only (L+R) is present. In this condition, the output terminals will both have the same signal because (L+R) does not pass through the diodes but is applied directly to both output terminals. When the two inputs to the scope are equal and in phase, the resultant is a straight line which slopes upward to the right. It is important to realize that when either (L-R) or (L+R) is missing, the output is no longer stereo, but will consist of either the sum or difference signal rather than pure R or L. Fig. 13-21C shows better separation as the control is advanced, allowing equal amounts of (L-R) and (L+R) to appear at the diode outputs. The control is adjusted to give the least amount of "straight-line" deflection and the greatest "fuzz-ball" effect. This method of adjustment, like the first one described, depends on the amount of separation in the program material at the time the tests are being made. The pattern changes constantly with the program material. Fig. 13 21D shows an instant when there was great separation in the program with almost equal volume level in both channels.

Obtaining this trace on the scope is an important part of the troubleshooting procedures to be described. The display is used to indicate the presence or absence of either channel at the output terminals. The scope can often be used as an output indicator when repairing a multiplex unit that does not have the twin audio channels available. As mentioned earlier, adjustment of the separation control causes a phase shift in the 19 -khz pilot signal in some circuits, so it may be necessary to retouch the 19 -khz adjustments.

Except for adjustment of the traps and filters, which seldom require attention, alignment of multiplex units can usually be done successfully "by ear," with a stereo station tuned in.

Four simple steps are listed here:

1. Tune in a stereo broadcast, and turn the separation control fully clockwise to remove all (L+ R).

2. Adjust the 19 -khz tuning controls (usually the oscillator grid tank) until the sound is undistorted. If no such point can be found, move the 38 -khz adjustment in either direction about one-half turn, and try again.

3. When undistorted demodulation of (L-R) is accomplished in Step 2, turn the separation control to the middle of its range. If a low-frequency, growling oscillation is heard, it will be necessary to readjust the 19 -khz circuit again, because the change in the separation control has affected it.

4. While an announcement is being made on one channel only, turn the volume to zero on the channel being used (usually the left one), and adjust the separation control for a null in the other channel. Retouch the 19 -khz tuning if necessary.


Despite the complicated theory of multiplexing, the circuitry is relatively simple, and troubleshooting is straight forward. SERVICING CHART XV at the end of this section covers four symptoms:

1. Monaural sound only.

2. Warbling or gargling.

3. Squeals and birdies.

4. Hissing background noise.

Monaural Sound Only


This condition is the result of lack of separation, and the tests described are equally as effective when there is no separation as when there is partial separation. Remember that this symptom does not mean that only one speaker is operating--it means that the same sound appears in both speakers.

Failure of one audio channel is handled by the methods de scribed in Section 3.


The chart is divided according to the type of circuitry used. In the switching-type circuits, this symptom must be the result of a missing, or very weak, 38 -khz oscillator signal.

Since (L-R) and (L+R) are not separated in these units, neither one can be absent while the other is still present. The technician is therefore directed to the tests in the oscillator as shown on the chart.

In some units, no oscillator is used at all-the 19 -khz pilot signal is amplified and doubled, and applied to the switching diodes. When the 38 -khz signal is missing, distortion is usually not apparent because the (L+R) signal is still unimpaired, and the (L-R) remains in the inaudible range from 23 khz to 53 khz.


The situation in the frequency-division units is quite different, however, and a more extensive series of tests is used.

Loss of the 38 -khz subcarrier, with (L-R) present, results in severe distortion which can be heard if (L+R) is removed by grounding the center tap of the separation control. This is the purpose of TEST POINT 1.

Three possible results can occur by removing (L+R), and the following tests are chosen accordingly.

Further Tests When Removing (L+R) Results in Undistorted Monaural Sound


This indicates that (L-R) is being demodulated properly, so the lack of separation must be due to a loss of (L+R). The coupling capacitor leading to the separation control should be checked by substitution, and the resistance of the control itself should be verified. After this, the two load resistors in the output of the diodes should be measured. If these steps do not locate the failure, the (L+R) line should be opened at the junction of the two resistors mentioned above and the scope used to trace toward the separation control until the signal is found.

Further Tests When Removing (L+R) Results in No Sound


In this case, it is obvious that (L-R) is missing, so the tests mentioned on the chart are chosen to locate a failure which could remove (L-R) without affecting (L+R).



This distortion can result from an insufficient 38-khz subcarrier injection to the diodes, or a defect in the diodes themselves. After checking the diodes (if they are of the vacuum-type), the voltage at the grid of the 38-khz oscillator is measured. A few tests are listed on the chart under TEST POINT 2; these are to be made when the oscillator appears to be inoperative. But the technician must keep in mind that the 38-khz circuit may not be a self-sustaining oscillator. It may depend on amplification of the 19-khz pilot, in which case tests of the 19-khz circuits are in order.



A resistance check from the diode-input point to ground will determine the condition of the oscillator output secondary and any series resistors. If these are not defective, a check on the alignment of the 19-khz and 38-khz tuning should be made.

When the preceding preliminary checks do not locate the defect, it can be assumed that the lack of separation is caused by incorrect amplitude of either (L-R) or the 38 khz where the two are combined at the input to the diodes. TEST POINT 3, therefore, calls for a comparison of the peak-to-peak volt ages of these two signals. The ( L-R) is most easily checked by shorting the oscillator cathode to the grid to remove the oscillator signal so that (L-R) can be viewed alone. The oscillator voltage should be at least three times greater than the amplitude of (L-R). If it is not, the most likely cause is mal adjustment of the 38-khz output tank, or a failure of the tube itself. If the oscillator tank requires considerable retuning to increase the signal level, it is probable that some other defect has occurred in the oscillator, so it would be advisable to check voltages.

A weak (L-R) signal leads the technician to tests in the bandpass amplifier and bandpass filter circuits, and finally to alignment of the FM tuner.



This effect is caused by upper audio frequencies be tween the two channels beating together, and is the result of loss of synchronization of the carrier regenerating oscillator.

When the oscillator is only slightly off frequency, the effect is more like "warbling." When the error in oscillator frequency is greater, the sound is garbled or mushy, and may contain an audio howl. When this condition is suspected in the frequency division circuits, it is possible to listen to (L-R) only by turning the separation control down in the manner described for TEST POINT 1.

A scope connected to the input and output of the 19-khz amplifier is the best method of analysis. Once the point where the pilot or 38-khz signal disappears has been located, the VTVM can be used to examine the circuit components.

The possibility of misalignment always exists, of course, but the technician hesitates to perform alignment adjustments until he is certain that the trouble is not caused by some component whose defect will be temporarily compensated for.



These are always caused by unwanted signals being de modulated in the multiplex circuits and beating with the de sired audio frequencies. The trouble can often be traced to the presence of the 67-khz private-broadcast carrier which has not been filtered sufficiently. In some locations, the FM stations broadcasting stereo are also supplying this additional program material to private subscribers, and the 67-khz carrier is necessarily a part of the composite signal received by the stereo multiplex unit.

Squeals may also be the result of misalignment of the band pass filter, which permits 19 khz, or even lower frequencies, to enter the (L-R) channel. Squeals are a familiar complaint when attempting to record stereo broadcasts on a tape recorder. They result from the appearance of either 19 khz or 38 khz in the output of the multiplex detector which beats with the bias oscillator contained in the recorder. If this problem is not the result of unbalance in the diodes or misalignment of the filters in the unit, it can be cured by installing a 15 -khz low-pass filter at each output of the multiplex detector. Shielding and physical separation of the multiplex unit and the recorder may also prove helpful.



This is a common complaint with new stereo installations where only monophonic FM has been received previously.

The difficulty arises from the fact that a much stronger signal is required to produce noise-free stereo than for monophonic FM, and the user may find that his antenna system is inadequate or that the nearest station broadcasting stereo is simply not close enough. Sometimes nothing can be done, and at other times, improvement of the antenna solves the problem completely.

In cases where the user has already enjoyed good stereo from his equipment and suddenly finds a high noise level, the problem will nearly always be caused by tubes. The RF amplifier in the FM tuner is the most likely suspect, after which the IF tubes should be checked. The wide-band amplifier in the multiplexer is another possibility.

If the problem persists after tubes have been replaced, a check on the condition of the antenna should be made before extensive work begins on the chassis. If possible, a different tuner should be connected in place of the original to compare the noise level. This will allow you to gain evidence of the condition of the antenna without making a trip to the roof or the top of a tower.


1. Write a short paragraph to define each of the following:

a. Monaural, or monophonic sound.

b. Stereo sound.

c. (L+ R). Explain whether it is monophonic or stereo sound.

d. (L-R).

e. Switching (time-division) type synchronous detector.

f. Frequency-division type multiplex detector.

g. Balanced modulator.

h. Subcarrier.

i. Pilot signal.

j. Separation, as differentiated from balance.

2. Draw block diagrams to illustrate the difference between the two multiplex systems discussed in this section.

3. A stereo multiplex station is broadcasting a 5000-cycle audio note in the ( L-R) channel :

a. Will this sound appear in the left or right speaker, or both?

b. What are the frequencies of the sidebands present at the output of the bandpass filter in the receiver?

4. Sketch the scope waveform which will be seen with the vertical input of the scope connected to Test Point ~ and the horizontal input connected to ~ in Fig. 13-6, while receiving a stereo program.

5. If the tuner which feeds the circuit of Fig. 13-6 were changed to a monaural FM station, would the waveform in Question #4 change? If so, redraw the waveform.

6. Make up a servicing chart to cover the symptom of "No Separation" in the circuit of Fig. 13-11. Try to cover every possibility, and be specific about tests.

7. Look up the schematics for two multiplex units, one of each kind, and draw the circuits. Label all tuned circuits with the proper frequencies.

8. Why should the multiplex circuit be connected ahead of the de-emphasis network in the FM detector?

9. Describe one good way to determine if a station is transmitting stereo.

10. Explain how you would use a scope to determine the condition of the (L-R) signals.

11. In the case of the same channel reproduced in both speakers from a frequency-division type adapter, the following tests have been made:

a. All tubes found to be OK. b; All B+ voltages normal.

c. Oscillator running in sync.

d. (L-R) present at demodulators in sufficient amount.

What would you check next?

12. What steps would you take to remove the squeals that occur when you attempt to make a tape recording from a stereo adapter?

13. Suppose you are faced with the following difficulty in stereo reception: The program sounds very good from one station, but the program is distorted and full of squeals when another stereo station is tuned in. What would you suspect is wrong, and what would you do to correct the trouble?

----------- Servicing Chart XV: Stereo Multiplex Receiver Symptoms.

Top of Page