The Phase Discriminator
Fig. 7-26. Schematic of the phase discriminator, in which the two tuned
secondary circuits of Fig. 7-22 have been replaced by a single center
tapped tuned circuit.
Almost exclusively, f-m receivers employing limiter-discriminator networks use the Foster-Seeley type of circuit, which is also known as the phase discriminator, the center-tapped secondary discriminator, and the center-tuned type of discriminator. This circuit appears in Fig. 7-26. What is the difference between this circuit and that shown in Fig. 7-22? A tuned primary circuit appears in both. However, the two tuned secondary circuits used in Fig. 7-22 have been replaced by a single tuned circuit with a center tap, Two diodes and two load resistors are used in both. The diode circuits of Fig. 7-26 are completed through coil L and the respective half of the center-tapped coil associated with each circuit. Besides providing the d-c paths be tween the diode plates and their associated cathodes, this common coil has another function which will be shown later.
Voltage E1 is the i-f signal voltage developed across the tuned primary circuit. Examining the secondary of this i-f transformer, we note certain significant details. It consists of two windings L2 and L3 in series, resonated to the i-f peak by means of C2. The center tap on the secondary winding is connected to a coupling capacitor C and, also, to an r-f choke L. Associated with the two circuits and the r-f choke L are three volt ages, designated as E2 , E3, and E1 respectively, the latter being virtually identical to E1 across the i-f transformer primary. To explain these designations, it is necessary to discuss the coupling between the primary and secondary circuits of this transformer, as well as what happens in a transformer when the secondary is tapped at the mid point.
How does this type of discriminator operate? In brief the operation can be divided into three major actions, although more conditions than just three are actually involved. However, for clearer comparison with the double-tuned discriminator shown in Fig. 7-22, it will be explained in terms of three major actions.
In the first place, although a single tuned winding is used for the secondary circuit, the center tap on this winding causes a division of the signal voltage developed in the tuned circuit across the two halves of the secondary winding, that is across L2 and L3. The signal voltages across these two halves are always equal to each other, irrespective of the frequency of the signal voltage fed into this circuit from the primary.
The second major consideration is that the signal voltage present across the primary winding L1 is also present across winding L, which is common to both halves of the secondary circuit with respect to the signal voltages eventually applied to the two diodes D1 and D2. The final major action is the phase relation which exists between the signal voltage across L2, which we can call E2, and the signal voltage across L which, because it is the same as that across L1, is also identified as E1 ; also the phase relation between the signal voltage across L8 , or E3, and the signal voltage across L, or E1. The function of this discriminator network with particular reference to these three actions will now be discussed in detail.
Two methods of coupling the signal from the primary to the secondary circuit are used in this system. The resonant primary is inductively coupled to the resonant secondary winding; at the same time the signal voltage E1 across the primary is fed to the r-f winding L via the coupling capacitor C. If the circuit of C, L, and C5 is traced, it will be seen that L is in shunt with the tuned primary, the latter being grounded through C3. Neither C, L, C3 , or C5 is of a magnitude to alter the resonant conditions of C1 and L1 , the resonant primary. Thus, with respect to magnitude and phase, whatever signal voltage exists across C1-L1, also exists across L. The direct connection between the coupling capacitor C and the mid-point of the secondary winding is of no con sequence to the signal transfer between the primary and the secondary tuned circuits; however, it is the point to which the choke L must be connected to complete the differential rectifier circuit. Thus, the secondary system receives signal voltages in two ways: the resonant secondary receives its signal voltage by inductive coupling, and the r-f choke derives its signal voltage by means of capacitive coupling through the fixed capacitor C. The equal voltages across each half of the secondary winding are obtained in the following manner. When a winding is tapped at the mid-point and a voltage is induced in that winding by means of a varying magnetic field, the total voltage developed across the entire winding divides between the two halves. This is logical in view of the fact that half the total number of turns exists between the center tap and one end, and half between the center tap and the other end. So, whatever the nature of the signal voltage which will be developed across the tuned secondary circuit C2-L1-L3, it is possible to show that this voltage divides into two parts, that is, across each half of the winding. These voltages are designated as E, and E3, Resonance Conditions in the Phase Discriminator Let us examine this circuit at resonance. At resonance, the frequency of the applied signal and the resonant frequencies of the tuned circuits are both the same. Since the inductances and capacitances of a tuned circuit effectively cancel each other at resonance, the circuit behaves like a resistance. In a resistive circuit the current is in phase with the voltage, so in the secondary tuned circuit the induced current, call it I, caused to flow by the induced voltage, call it E, is in phase with this induced voltage. It should be remembered that this induced voltage is effectively in series with the inductance and capacitance of the secondary tuned circuit.
The in-phase relationships between E and I are indicated in the vector diagram of Fig. 7-27 (A) where vectors 01 and OE, the respective induced voltage and current vectors, are seen to be in phase with each other. The voltage across the primary circuit, designated as E1 is 180° out of phase with the voltage induced into the secondary circuit.
This voltage E1 is the main voltage upon which all the other voltages are based. Consequently, this voltage is drawn as vector OE1 along the 0° reference line.
In vector diagram, Fig. 7-27 (A), vector OE1 is 180° out of phase with the induced voltage vector OE. Since the voltage across a pure inductance leads the current through it by 90°, the voltage drops E2 and E3 across the secondary coil (called reactive voltage drops be cause the inductance is considered to be a pure inductance containing negligible resistance) lead the current I flowing through it. This is indicated where vectors OE1 and OE3 are leading the induced current by 90°. This also means that the reactive voltage drop across the secondary coil is lagging the primary voltage E1 by 90°, and vector OE1 is seen to be leading vectors OE2 and OE8 by 90°. The 90° phase difference between these voltages is very important to the operation of the discriminator. It should still be remembered that this voltage E1 also exists across coil L of Fig. 7-26 in the same phase and magnitude as that existing across the primary tuned circuit.
Fig. 7-27 (A), left, (B) right. Vector diagrams of the current and voltage
relationships existing in the phase discriminator. Note that in (B) vector
OE3 is shifted in phase by 180° from its position in (A).
That the secondary is center tapped means that it is in a push-pull arrangement, and hence voltages E2 and E3 are equal in magnitude but 180° out of phase with each other, as referred to the center tap.
However, the same current flows through both parts of the secondary coil, so that a 90° phase relation must still exist between each voltage and the current; but in one case one of the voltages is effectively leading the current and in the other case the voltage is effectively lagging the current by 90°. This means too that one half of the secondary voltage drop is leading voltage E1 by 90° and the other half lagging voltage E1 by 90°. All of this is indicated in the revised vector diagram of Fig. 7-27 (A) as shown in Fig. 7-27 (B), in which vector OE8 of the previous vector diagram has been shifted 180°. To demonstrate how all these voltages affect the duo-diode circuit we have redrawn that part of Fig. 7-26 appearing to the right of the secondary of the transformer in simple form in Fig. 7-28. In this figure we have made two separate circuits of the diodes, showing the respective voltages that act upon each diode. These two simple circuits then are combined to show how they actually work together.
Fig. 7-28. Simplified schematics of the diodes of Fig. 7-26 and their
respective load resistors. These two schematics are combined on the right,
which is permissible in as much as the voltage E 1 is common to both
diodes.
This figure and Fig. 7-26, show that voltage E1 is common to both diodes since it exists across the inductance L. Also, since voltage E1 is active on diode D1 and voltage E3 is active on diode D,, it is readily seen from Fig. 7-28 that voltages Es and E1 are active on diode D1 and voltages E8 and E1 are active on diode Ds, Further examination of this simplified circuit reveals that the rectified current flows through the individual diode circuits put certain polarities on their load resistors. Since the external current in a diode rectifier circuit flows from plate to cathode, the currents in the diode load resistors will be flowing in opposite directions, and the polarities across the individual load resistors will be bucking each other. Thus between points A and B a voltage will exist which will be the difference between the voltage drops across resistors R1 and R1. If the voltages Es and E3 have the same phase angle with respect to the voltage E1, both diode currents will be equal in value, and the same voltage drop will appear across each load resistor R1 and R,.
Fig. 7-29. Vector diagram of the current and voltages relationships
of the circuit of Fig. 7-26 at resonance and when both secondary voltages
E2 and E3 have the same phase difference (90°) from E1.
Since each resistance voltage drop is opposite in polarity to the other but equal in value, the total voltage measured between points A and B will be zero. Under these circumstances the output of the differential rectifier circuit is zero. If, however, the phase relationships between Es and E1 and between E8 and E1 differ, a differential voltage will exist between points A and B, because the voltage drops across R1 and Rs no longer will be equal to each other as different currents flow through each diode.
Zero voltage exists across points A and B when the resonant frequency of the tuned discriminator i-f transformer is exactly equal to the applied frequency. This is simply indicated by the vector diagram of Fig. 7-29 where both secondary voltages have the same phase difference, namely 90°, from voltage E1. This diagram is nothing more than a duplicate of that diagram of Fig. 7-27 (B) with the exception that the two voltages active on each diode are added vectorially. Thus, in Fig, 7-29, vector OE4 represents the resultant vector of the vector addition of voltages E3 and E1 across diode D2 and vector OE,; represents the resultant vector of the vector addition of E2 and E1 across diode D1. Resultant vectors OE4 and OE5 are shown to be equal in magnitude, causing the same current to flow in each diode circuit.
Thus, equal but opposite voltages are developed across diode load resistors R1 and R2, producing zero voltage between points A and B. Applied Frequency Higher Than Resonant Frequency When the instantaneous value of the FM signal is equal to its center frequency component, we have the frequency applied to the discriminator transformer equaling the resonant frequency. The situation for this was discussed in the preceding section. At either side of the center frequency component of the FM signal the instantaneous frequency is different from the resonant frequency of the i-f transformers.
Under these conditions the discriminator transformer is tuned below or above the incoming i-f signal.
Let us now consider an instantaneous value of the FM signal greater than the center i.f. The discriminator transformer then is tuned below the incoming i.f. The circuit is still the same as in Fig. 7-26, and the nonresonant conditions do not alter the fundamental rules of the action of the primary circuit, so that the voltage E1 that exists across this circuit also exists across Lin both the same phase and magnitude.
Induced voltage E in the secondary remains 180° out-of-phase with the primary signal, for this too is a fundamental condition which is not altered by non-resonance conditions. However, the phase relationship between the induced voltage E and the current I, which it causes to appear in the secondary circuit, is affected by the state of resonance, and in turn alters related conditions.
When the applied frequency is higher than the resonant frequency, the reactance of the secondary coil becomes greater than the reactance of the capacitor. This accords with the fundamental law that inductive reactance varies directly with frequency, and capacitive reactance varies inversely with frequency. Accordingly, a portion of the inductive reactance will be offset by the capacitive reactance, but a certain amount of inductive reactance will remain to exert a control on the induced current. The circuit as a whole now appears as an inductance and resistance in series, rather than as a resistance alone, which is the case at resonance.
Under this circumstance the induced current I no longer will be in phase with the induced voltage E but rather will lag this voltage by a certain amount, depending upon the extent to which the instantaneous FM signal is greater than the tuned frequency of the trans former. This is all indicated in the vector dia gram of Fig. 7-30 for the off-resonance condition now being discussed. Voltages E and E1 are still seen to be 180° out of phase but the phase relationships of the other component voltages differ somewhat from those of the vector diagram of Fig. 7-29.
For the sake of argument let us say that the difference in frequency between the instantaneous frequency of the FM signal and the tuned i-f transformer is such that the amount of inductive reactance remaining is sufficient to cause the induced current I to lag the induced voltage E by 35°. No matter what the phase relationship between the induced voltage and induced current, the two voltages E2 and E3 across the individual halves of the secondary are still 180° out of phase with each other and equal in magnitude. The induced current flowing through this secondary still bears the same phase relationship to these secondary voltages. Regardless of the phase difference between E and I, secondary voltage E 3 will still lag current I by 90°, and secondary volt age E2 will still lead current I by 90°. This is indicated in the vector diagram of Fig. 7-30; and if this vector diagram and that of Fig. 7-29 are compared, these phase relations will be seen to hold.
Fig. 7-30. Vector diagram of the current and voltages relationships
in the discriminator when the instantaneous frequency of the FM signal
is higher than the transformer resonant frequency.
Let us further compare these two vector diagrams. To keep constant the 90° phase relations between voltages E2, E3 , and current I, then when current I lags induced voltage E by 35°, voltage vectors OE2 and OE3 are both shifted 35° clockwise to keep these 90° relationships intact. The complete vector line E2-0-E3 is shifted 35° in a negative direction. When the respective voltages applied to the individual diodes are added under these circumstances, it will be seen from the vector diagram of Fig. 7-30 that the resultant vector OE_. representing that voltage across diode D2 and resultant vector OE6 representing that across diode D1 are no longer equal, but that vector OE_, is greater than vector OE5. In this instance diode Ds will draw the greater current, and in Fig. 7-28 load resistor R, will have a greater voltage drop than resistor R1; hence a differential voltage will exist across point A to B, with point B being more positive than point A. This is the same as saying point A is negative with respect to point B.
Applied Frequency Lower Than Resonant Frequency
When the instantaneous value of the FM signal input to the discriminator circuit is such that it is less than the resonant frequency of the discriminator transformer, the differential voltage will still exist across the diode loads, but the polarities will be reversed. Let us see how this happens.
Fig. 7-31. When the instantaneous frequency of the FM signal is -- lower
than the transformer resonant frequency, the current I will lead the
induced voltage E. Compare with Figs. 7-29 and 7-30.
We still are at off-resonance conditions, even though we are on the lower side of the resonant frequency, and the same 180° phase relationship between E and E1 exists. When the applied frequency is lower than that of the resonant frequency of the i-f transformer, the impedance of the secondary of the i-f transformer is such that the capacitive reactance more than balances out the inductive reactance, and the secondary is primarily capacitive. Since this circuit is capacitive, the induced current I leads the induced voltage E, If the off-resonance conditions are such that a phase angle of 35° again exists between I and E, I will be leading E by 35°, as seen in the vector dia gram of Fig. 7-31. Since the 90° phase relations between voltage E8 and current 1, and voltage E3 and current 1 must still exist, these two voltages are effectively shifted in phase 35° in a counterclockwise or positive direction. This is indicated in Fig. 7-31 where vectors OE1 and OE3 are still 180° out of phase with each other, but no longer 90° out of phase with vector OE1. Now when the individual i-f voltages across the diode circuits are combined vectorially, it will be seen that resultant vector OE5 applied to diode D1 is greater in magnitude than resultant vector OE_. applied to diode Di. Therefore, the current in the circuit of diode D1 is greater than the other diode current. This means a greater voltage drop exists across R1 , the load resistor of diode D11 than across R2, the load of D2, and a differential voltage exists between points A and B of the diode circuit of Fig. 7-28. However, under these conditions the polarity of point A will be more positive with respect to the junction of R1 and Rs than point B. That is, point B is negative with respect to point A. Summarizing the action described, it is evident that if a varying frequency input signal ( one which varies in frequency around a mean) is applied to the discriminator network - provided that the range of frequencies covered is not beyond the acceptance bandwidth of the discriminator transformer -- an output signal which changes in amplitude and polarity will be obtained. The output signal is deter mined by the frequency deviation; the less the frequency deviation, the less the departure from a 90° phase relationship between the re active voltages E2 and E1, and also E3 and E1. The greater the frequency deviation, the greater is the difference in angular displacement between E2 and E1, and E3 and Ei, so that the differential voltage obtained from the diodes is greater. When viewed from the angle of audio intensity, the greater the differential voltage from the rectifiers, the louder the audio signal, since the extent of deviation at the transmitter is a function of modulating voltage level. The greater the modulating voltage level within prescribed limits, the greater the frequency deviation.
In brief, then, the differential output voltage is a .function of the rate of deviation of the FM signal as well as the amount of frequency deviation. Since the amplitude of the audio modulating signal deter mines the amount of deviation and since the frequency of the audio determines the rate of change of the deviation of the FM signal, it becomes readily apparent that the differential voltage across the two diodes will be an audio signal equal in frequency and proportional in amplitude to the audio modulation signal.
The output characteristic curve of this discriminator network is the same S-shaped curve discussed in relation to the previous circuits and shown in Fig. 7-25. It is desired that this curve be linear for at least 75 khz on either side of its center point. That is, it should be linear over the maximum 75- khz peak deviation of the FM signal in order to avoid distortion. In fact, the greater the linearity (within reasonable limits) the better, and an over-all 200- khz linearity is considered very desirable. This accomplishes two things: Firstly it means that the receiver does not have to be tuned very accurately for the resting frequency to fall in the middle of the discriminator characteristic. If the receiver is mistuned somewhat, no distortion will result provided the receiver is not so badly mistuned that the frequency variations in the signal extend into the nonlinear or curved portions of the discriminator characteristic. Secondly, the fact that the characteristic of the discriminator is linear over a greater range than that actually required means that the linearity will be more nearly perfect over the center portion which is actually used in reception. The high degree of linearity obtained in this way makes demodulation of the signal possible with practically no distortion.
In Fig. 7-26 a resistor capacitor combination consisting of R3 and C6 appears in the output circuit of which no mention has yet been made. As will be recalled, in Section 4 we discussed pre-emphasis and de-emphasis networks in conjunction with the proper level for all audio frequencies. This R3 and C6 combination is the standard form of de-emphasis network that usually appears at the output of the detector network. The potentiometer R4 controls the amount of audio signal to be coupled to the audio amplifiers through coupling capacitor C7.
Fig. 7-32. Schematic of a modification of the phase discriminator circuit
is shown in (A) and simplified schematics of this circuit are shown in
(B) and (C). Note that coil L which appeared in Fig. 7-26 is omitted.
Modification of the Phase Discriminator
While we are on the topic of circuit construction, it might be well to state that there are some modifications of the discriminator detector circuit which appears in Fig. 7-26. Some of the more common types of such modified circuits do not use any choke in coupling the voltage across the primary inductance to the diodes. A typical circuit of this sort appears in Fig. 7-32 (A). From this drawing it is not readily apparent how voltage E1 across the primary inductance L1 is coupled to the diode circuits. Comparing this diagram with Fig. 7-26, it will be noted that the choke L does not exist in the circuit now under discussion. Regardless of this fact, the voltage E1 is introduced into the diode circuits.
If we trace the circuit from the top of L1 we can take two paths: through C1 and L2 or through Cs and L3. Since the circuit is balanced, Ls is equal to L3, and C1 is equal to Cs, and it is of no importance which path we take. For the sake of illustration we will choose C2 and Ls and the path ending with resistor R2 to ground. The C 1 and L2 path could likewise end with resistor R8 to ground. Consequently we can redraw these circuits and see how voltage E1 is injected into this part of the network.
This simplified drawing is shown in Fig. 7-32 (B), where C1 and L8 are in series as well as Cs and L3 and the two series circuits are in parallel with each other. This complete parallel circuit is in series with R2. Therefore, the voltage E1 also appears across this circuit as shown since the circuit is shunted across L1. One end of L1 is grounded for the i.f. through a bypass capacitor C6. However, at the frequency of the i-f input signal, L2 and C1 are in series resonance, and so are L3 and Cs, so that each series resonance combination is a very small resistance. Therefore, the impedance of the complete parallel circuit is so low compared with the resistance of R2, which is typically 100,000 ohms, that practically all of E 1 is considered to be across R2 in the same phase and magnitude as that existing across L1. Since R, is the load resistor of diode Ds this resistor, besides serving as the diode load, also serves to apply voltage E 1 to diode D2. So much for E1 being applied to D2. But how about also getting E1 across D1? This also can be easily shown by taking the path of R1 and C 4 to ground instead of only R2 to ground. Since the rest of the traced circuit remains as just discussed, E1 also appears across the series combination of R1 and C4. However, since the reactance of C4 is much smaller than R1, which is equal to R2 , all of E1 appears across R1 as well as across R5. Consequently, R1 serves the same purpose as Rs in being used as a load resistor and for applying E1 to D1. The complete picture of how E1 appears across both resistors R1 and R2 is illustrated in the over-all simplified drawing of Fig. 7-32 (C). In this drawing the components that have negligible reactance compared with the resistance of R1 and R, are drawn so that E1 appears to be present across both R1 and R2. Another type of modified discriminator uses the diode sections of a duo-diode high-mu triode tube. The 6AQ7-GT which is used by General Electric in some of its FM receivers is such a tube. The diodes in this tube can be used as a discriminator, because a separate cathode is used for the diodes and another for the triode section. A schematic of the tube appears in Fig. 7-33 (A), and a circuit diagram of how the diode sections are used as a discriminator is shown in Fig. 7-33.
Fig. 7-33. The 6AQ7-GT duo-diode triode tube (A) can be used in the
discriminator circuit (B) as it has one cathode for the diode plates
and another cathode for the triode. A simplified schematic of the parallel
circuits which appear across the transformer primary circuit is shown
in (C).
(B). The chief difference between this circuit and the conventional discriminator is the method of applying reference voltage E1 to both diodes D1 and D2. Coil L2, and thus voltage E2, are common to the upper diode Dz, and coil L3 and voltage E3 are common to the lower diode D2. Capacitor C 4 connecting L2 and L3 is of high enough capacitance, so that both coils are effectively in series to the i.f. As with the conventional discriminator, capacitor C2 is shunted across these two coils and with them forms the secondary tuned circuit. This analysis reveals how the respective induced voltages E2 and E3 are applied across the individual diodes Dz and D2, but the method of obtaining Ez across both diodes is not readily evident.
Tracing the d-c path for each diode, we find that resistor R1 is the load for D1 and resistor R2 is the load for D2. The reference voltage E1 is capacitance coupled through C3 to the common cathode of the diodes. Capacitor Cs is a bypass capacitor for the primary tuned circuit and completes the r-f path to ground. The reactance of Cs at the i.f. is so small that negligible i-f voltage appears across it.
If we trace the i-f path from the top of L1, we find that there are essentially two parallel paths which appear across the primary circuit to ground. This circuit is shown in simplified form in Fig. 7-33 (C). Going from the top of Lz we pass through Cs, and find two paths available: one through resistor Rz to ground and the other through resistor R2 and capacitor C4 to ground. The capacitances of Cs and C4 are so chosen that they will offer a low reactance at the i.f. compared with the resistance of R1 and Rs, This means that practically all of E1 also appears across R1 and R2. So far as the high-frequency i.f. is concerned, R1 and R2 are both effectively in parallel with L1 , and the reference voltage E 1 also appears across the load resistors R1 and R2. Since R_1 is the load resistor for D1, both voltages E1 and E2 act on diode D1 ; and, since R2 is the load resistor for D2, both E1 and E3 act on diode D2. The on and off resonance conditions function as in the conventional discriminator. The audio output appears across C4 or between the high side of R2 and ground. In this circuit, resistor R3 and capacitor C6 represent a de-emphasis network, and the de-emphasized audio is taken across C6. The triode section of the tube is used as the first audio voltage amplifier.
An FM Tuning Indicator
Fig. 7-34. The schematic (A) of the 6AL7-GT visual tuning indicator
tube and how it is connected in an FM receiver (B) employing a limiter.
In (C) are shown various target patterns under different conditions of discriminator output.
["F. M. Bailey, "An Electron-Ray Tuning Indicator for Frequency Modulation," Proc., IRE, p. 1158, vol. 35, October 1947. ]
A while back we discussed how the limiter grid voltage was used as a source of tuning indication in FM receivers employing limiters, even though certain disadvantages were prevalent. Ordinary indicator tubes, such as the 6E5 or 6U5 commonly used for a-m receivers, served this purpose. In late 1946 the General Electric Company re leased a new tuning indicator3 which worked on cathode-ray principles and was ideally suited for f-m receivers, although it can also be used on a-m receivers. This type tube is known as a 6AL7-GT, a schematic of which appears in Fig. 7-34 (A). This tube can make use of both the discriminator output voltage and the limiter grid voltage. It can be used satisfactorily with just the discriminator voltage alone, but it then is not able to distinguish whether the receiver is on tune or off channel. For, when the discriminator output voltage is used alone, the pattern for on tune or off channel is the same, because the discriminator output voltage is zero in both cases.
The tube is made sensitive enough to respond to a voltage difference between plus and minus 0.2 volt with respect to ground; this is within 2 khz of the discriminator tuning for distortionless signals.
This discriminator output is connected to one side of a center divided deflector [deflector No. 2 in Fig. 7-34 (A)]. A space-charge grid is used to increase the sensitivity of the tube. The deflector is so divided to form a method whereby the deflection of one half of the pattern can be compared with the other half. The other half of the divided deflector is usually grounded [deflector No. 1 in Fig. 7-34 (A)], providing a reference pattern with which the pattern due to deflector No. 2 may be compared.
Since one side of the discriminator load is usually grounded, the changing discriminator output voltage appears across both these deflector electrodes. On one side of the cathode is placed this divided deflector, and on the other side another deflector is placed to form a fixed boundary for one side of the pattern [deflector No. 3 in Fig. 7-34 (A)]. The first limiter grid (if more than one limiter is used) is connected to this electrode, and the limiter voltage therefore appears on this deflector and is used as a voltage which determines one boundary of the target pattern. This deflector enables us to distinguish, by different patterns, when the receiver is on tune or off channel.
A circuit diagram of how the tube is usually wired for operation in conjunction with a limiter and discriminator is shown in Fig. 7-34 (B). In Fig. 7-34 (C) are illustrated the target patterns under different conditions of discriminator output. In pattern 1 four squares are shown where P 1 , P 2 , and P 3 are produced and controlled by deflection electrodes No. 1, 2, and 3 respectively; there are two P3 squares. Patterns 2 and 6 are for minus and plus off-channel conditions respectively.
These patterns are identical, because there is zero voltage output from the discriminator and no limiter voltage. The size of squares P 8 is controlled by the amount of limiter voltage; the higher this negative voltage the smaller the squares, and the lower the negative voltage the larger the squares. Consequently, in patterns 2 and 6 the bottom half is of maximum depth due to the absence of limiter voltage.
In pattern 4 the receiver is on tune, which means a maximum negative limiter voltage and a zero discriminator output voltage. Thus the pattern is much smaller than those off-channel patterns of 2 and 6.
The decrease in pattern No. 4 is seen to be in the bottom half due to the limiter voltage.
In patterns 3 and 5 where the discriminator is off tune, limiter volt age, although perhaps not a negative maximum, is still present, limiting the bottom half of the patterns to a smaller size as compared with the patterns of 2 and 6. In pattern 3 the discriminator is off tune and presents a negative signal to the No. 2 deflector grid, making square Ps reduced in size. In pattern 5 the discriminator is also off tune and presents a positive signal to the No. 2 deflector grid, making square Pe increased. You will note that the deflection due to the positive off tune signal is greater than the deflection for the negative off tune signal under the same amount of positive and negative discriminator voltages. This is a result of the use of space-charge operation of the deflecting system (due to the space-charge grid). Deflection in the positive region would be much greater if a cathode bias resistor [3300 ohms as seen in Fig. 7-34 (B) ] were not used. This resistor places a positive voltage on the cathode with respect to the deflectors and the space-charge grid, so that with a positive discriminator signal the deflectors do not draw appreciable current.
The negative bias on the space-charge grid reduces the brightness of the pattern and increases the deflection sensitivity.
THE RATIO DETECTOR
In previous discussion on the discriminator detector circuits it was indicated that a limiter was necessary to produce an input FM signal to the discriminator that was constant in amplitude. The limiter is needed because the discriminator responds to amplitude changes in the signal as well as to frequency changes. To dispense with a separate limiter tube, a new FM detector system, the ratio detector, was developed that responded to frequency changes only and not to amplitude changes in the input system. Although its operation is different from that of the discriminator detector, it is nevertheless somewhat similar in circuit analysis. This will be seen as we progress with the analysis of the detector.
Simplified Ratio Detector
To understand fully the operation of a typical ratio detector circuit let us first study a simplified version of such a circuit as illustrated in Fig. 7-35. From this diagram we notice one thing that is common to all ratio detector circuits, namely, that the two diodes used are wired in series aiding with respect to the load instead of series opposing as in the discriminator detector circuit. By tracing the circuit it will be found that the plate of one diode is connected to the cathode of the other diode through the secondary of the transformer. The plate of the latter in turn is connected to the cathode of the former diode through a battery. Compare these connections with that of the discriminator detector in Fig. 7-26, and the difference will be immediately apparent. This difference in circuit arrangement of these two detectors is one quick method of telling them apart.
Fig. 7-35. Simplified schematic of the ratio detector.
In this circuit the diodes are series aiding with respect to the load instead of series opposing as in the discriminator detector circuit.
Coming back to Fig. 7-35, we find that the transformer network in conjunction with C and L is similar to the discriminator detector arrangement in that the voltage across the primary, L1, is also across L. This voltage in conjunction with the individual voltages across L9 and L3 is effectively on each respective diode. Across the output of this circuit appear two capacitors C2 and C3 and in parallel with them a battery of fixed voltage, EB. In the discriminator detector, besides two capacitors, two resistors also existed, across which the differential out put voltage was developed. However, when any amplitude changes occurred in the input signal to the discriminator detector, the output voltages across each resistor changed, making the differential output voltage different, indicating that the detector was responsive to a-m signals as well as to FM.
In the simplified ratio detector circuit of Fig. 7-35, the voltage from diode to diode in the output side of the circuit is maintained constant at EB by the voltage of the battery. Consequently, the total voltage across capacitors Cs and C3 will always be equal to EB. According to the polarity of the battery connection, no current will flow in the circuit until a signal is applied. The d-c path of this output circuit is through the battery, and the a-c path through the two capacitors C9 and C3. When an FM signal is detected by this arrangement, the individual voltages Ex and Ey across capacitors Cs and C3 respectively, will be constantly changing due to the change of deviation of the FM signal, but their sum will be constant due to the battery voltage EB. At all times EB= E£ + Ey. Since the values but not the sum of Ex and Ey can change, it is their ratio which will be constantly changing, and by placing a potentiometer across C3 the audio modulation of the FM signal can be tapped off this resistance. This will all be clearer when a typical ratio detector circuit is analyzed.
Any amplitude variations in the input signal will not appear across either C2 or C8 as changes in voltage due to the constant voltage across the output due to the battery. It is only frequency changes which appear across both capacitors. This is different from the discriminator detector circuit, where both amplitude and frequency changes were recorded as voltage changes across the load resistors. This simplified form of the ratio detector was analyzed first because through the use of the battery we were easily able to show the fixed voltage across the output and why this detector responds only to frequency variations.
Practical Ratio Detector
Fig. 7-36. A typical ratio detector circuit. Note the similarity to
the discriminator detector.
The use of a battery for constant voltage output in the ratio detector for FM receivers is not practical because, due to the nature of the incoming FM signals, we desire to have the output voltage constant only at the average strength of the incoming signal. Since the carriers of the different FM stations are not all of the same strength and also since the effective strengths of the individual carriers at the receiver change, it was found that the best thing possible was to have a relatively constant voltage across the output determined by the average value of the incoming signals. This was accomplished by placing a parallel RC network in the output of the circuit instead of the battery, as shown in the typical ratio detector circuit of Fig. 7-36. Let us now study this circuit in conjunction with the new RC output circuit, and the over-all action of this detector will become apparent.
This new circuit is very similar to the discriminator detector arrangement in many ways. For instance, the voltage E1 can be considered as effectively the same, even though it is not obtained from the limiter. In both circuits the signal is of the i.f. This voltage E1 is coupled to the rest of the circuit in two ways: by induction into L8 and L3 and by capacitive coupling across C1 to L and C4. The reactances of C1 and C4 at the i.f. are negligible compared with the reactance of the choke L. Consequently, the voltage drop across L is likewise E 1 as seen in Fig. 7-36. Depending on the degree of coupling between the primary tuned circuit and the secondary tuned circuit, a certain amount of voltage is coupled across each half of the center tapped secondary. So far a& voltages E2, E3, and voltage E1 across L are concerned, they all function in exactly the same manner as the corresponding voltages in the discriminator detector circuit of Fig. 7-26. This is true, too, for the vector diagrams of Fig. 27 (A), (B) and Figs. 7-29 through 7-31 with respect to the applied voltages to the individual diodes during the constant frequency changing of the in coming FM signal. Thus, the same phase shifting process is employed in the ratio detector circuit.
Examining Fig. 7-36 once more, it is evident that since the diodes are connected in series aiding they draw current in the same direction relative to R, which is also in series with them. Consequently, using the convention for the flow of electrons from cathode to plate, the current I will follow the path indicated by the arrow and the top part of resistor R will become negative with respect to its bottom or grounded end. If the primary and secondary tuned circuits are both resonant to the i.f. and if an unmodulated i-f carrier signal ( of the same frequency) is injected into the circuit from the i-f amplifier, the two capacitors C3 and C4 both will be charged to the same voltage due to the symmetry of the circuit.
Now if the i-f carrier were frequency modulated, the voltages appearing across capacitors C1 and C4 would vary according to the modulation of the i-f carrier. This happens as follows: It was mentioned that rectified current would flow in the direction shown in Fig. 7-36 and that the top portion of the resistor R would have a negative potential on it. The values of the resistor R and capacitor C are so chosen that they represent a long time-constant network.
Usually the value of this time-constant network can vary anywhere between one-tenth and one-quarter second and still be effective to the desired degree. (The value of the resistance in megohms multiplied by the capacitance in microfarads will indicate directly the value of the time constant in seconds). Consequently, with a long time constant, it will take the capacitor C quite some time to discharge fully through R. Therefore, the negative voltage at the top of resistor R will remain practically constant over the range of the lowest audio frequency desired to be reproduced in the output of the set. In other words, a time constant of one-tenth of a second corresponds to the period of a frequency of 10 cycles per second; therefore, for frequencies above 10 cycles per second, the duration of one cycle would be shorter than the time constant, and so the voltage across the R-C combination will remain practically constant. (The higher the audio frequency, the shorter the duration of one cycle.) Since the voltage across R and C is constant, the sum of the volt ages across Cs and C4 must remain constant. However, if the carrier frequency falls below, or rises above, the i.f., the voltages appearing across Cs and C4 will differ in value according to the degree of off resonance condition of the i-f signal. No matter what the difference between these voltages, their sum always remains the same, but their ratio will be varying at the rate of the deviation of the FM signal, and it is this change in ratio which is detected. If the i-f signal is frequency modulated, the i.f. will vary above and below its resonant frequency according to the degree of FM This accordingly will vary the voltages appearing across C5 and C4, but in a certain proportion determined by the potential across the RC combination. Consequently, it can be said that the voltage across C4 varies at an audio rate (due to the degree of FM) Therefore, the a-f output may conveniently be taken off across C 4, because one side is grounded, and applied to the audio section of the set. The instant when the incoming FM i-f signal is at the exact resonant frequency of the tuned circuits of Fig. 7-36, the a-f voltage across C4 will be zero. At instantaneous frequencies of the incoming FM i-f signal above and below the resonant frequency of the tuned circuits, the voltage across C4 will vary at a rate deter mined by the changing frequency of the FM signal. Since the FM signal is changing, or is being deviated, at an audio rate (that is, at the rate of its modulating signal), the output voltage across C4 will be varying at an audio rate. In this type of circuit the voltage appearing across Cs will be larger than that across C4 at frequencies below i.f., and above the i.f., the voltage across C 4 will be larger than that across C5.
The basic part of the ratio detector that removes any AM appearing in the input is the R-C time constant network of Fig. 7-36. It is the constant voltage across resistor R and capacitor C that plays the primary role in the removal of AM Let us suppose that an a-m signal appears at the input of the ratio detector and see what happens: Any a-m signal will tend to increase the voltages across capacitors C8 and C4. However, the voltage across the RC network cannot change rapidly enough to follow the AM, due to the nature of the long time constant, and the AM therefore, cannot change the voltage across C8 and C4. In other words, the capacitor C charges or discharges so slowly through R that the potential at the top of resistor R ( or the plate of diode D1 ) remains nearly constant and any AM cannot change the voltage across capacitor C in step with this AM Consequently, sudden increases in amplitude of the FM carrier will not have any effect in the output audio circuit, because these sudden increases of amplitude cannot appear across either C3 or C4 as a change in voltage.
AVC From Ratio Detectors
In the limiter discriminator arrangement, ave voltage was available in the grid circuit of the limiter due to grid rectification action. In the ratio detector system, since limiters are not employed, the ave is obtained from some other place. In the ratio detector circuit of Fig. 7-36 it is noticed that the voltage across resistor R serves as a means of obtaining ave voltage. Since the time constant network of RC is made to produce a constant output voltage at the average strength of the incoming signal, it is readily evident that this output voltage will change in accordance with the varying average strength of the in coming signal.
What this means is that the capacitor C in conjunction with resistor R averages these signal strength changes appearing across R. The time constant is not considered large in this instance as compared with the length of time required for changes in average signal strength. How ever, the effect of the time constant is sufficiently large to produce effective removal of sudden changes in AM, including that brought about by the response characteristics of the i-f stages. This is possible because the input signal does not change in strength as rapidly as these other amplitude variations, and the RC combination permits slow changes in voltage in accordance with slow changes in the received signal. Therefore, the negative voltage at the top part of resistor R serves as a source of ave voltage.
Other Ratio Detector Circuits
Although the ratio detector circuit of Fig. 7-36 is very typical of those in use today, there are enough variations in this type of circuit to warrant separate consideration. The most difficult part in the de sign of this type of circuit is the ratio detector transformer, namely that comprising coils L1, L2, and L3 in Fig. 7-36. It is beyond the scope of this guide to go into such design work, but it should be remembered that such factors as the proper coupling between windings, the respective Q's of the coils both during diode unloaded and loaded conditions, and the gain of the last i-f stage are important in this transformer design.
Fig. 7-37. A modified ratio detector circuit in which the coil L receives
its voltage by transformer action from the coil L1.
Although it may not be immediately apparent, Fig. 7-37 illustrates a type of ratio detector circuit which is a modification of that in Fig. 7-36. As in all the detector circuits discussed so far, the primary problem in this new circuit is in getting voltage E1 which is across L, to also appear across both diodes. This voltage across the diodes must be either in phase or 180° out of phase with that across L,. Unless something is known about the construction of the transformer comprising coils L, L,, L2, and L3, it is somewhat difficult to understand this circuit.
It should be known first that, due to inductive coupling, voltages E, and E3 appear across L, and L3, respectively, as in other types of detectors. The special thing that should be known is that coil L, which has only a few turns, is a separate winding usually closely wound around or near the bottom or B+ side of coil L,. In this manner the coupling between these two coils is a maximum, and practically all of voltage E1 appears across this coil and series capacitor C5 to ground.
Since L is untuned, the voltage induced into it from L1 is 180° out of phase with E1. The voltages Es and E3, as just pointed out are in quadrature with E1. Therefore, these two voltages are also in quadrature with the voltage across L, as required. Resistor R3 and capacitor C6 represent the conventional de-emphasis network.
The basic operation of this circuit can now be understood, especially if we consider the ground side of capacitor C6 being connected to the junction of Cs and C8 which is also grounded. In this manner it can be easily seen that both diode i-f currents flow through L and C6 but in opposite directions. The current through diode D1 represented by I1 flows from its cathode to the plate, then through C2 to ground, then through C5, L, L2, and back to the cathode of D1. The current through diode D2, represented by I2, flows from its cathode to plate, then through Ls, L and C5 to ground, then through ground to C.,, and back to the cathode of D2. All of which is indicated by the arrows of current flow. It will be seen that the net diode current flow through L and Cs will be zero, and no voltage drop will be present across C 5 when the instantaneous i.f. is at the resonant frequency of the transformer.
When the instantaneous frequency of the incoming signal is different from the resonant frequency of the tuned circuits, currents 11 and 12 will differ from each other, and a potential drop will exist across C5.
The value of the voltage drop depends upon the difference in diode currents 11 and 12 , which in turn depends upon the amount of difference between the instantaneous .frequency of the incoming signal and the resonant frequency of the tuned circuit. Since the input signal is frequency modulated and the amount of frequency deviation on either side of the center frequency and the rate of deviation are determined by the amplitude and frequency of the audio modulating signal, the voltage drop across C5 will vary in accordance with this audio modulating signal. The varying difference between 11 and 12 thus produces a proportionally varying voltage drop across Cs, and since this varying difference in current is determined by the rate and amount of deviation of the FM i-f signal, the voltage drop across C5 varies at the audio modulating signal. Consequently, the voltage across Cs is used as the source of audio signal for the following audio amplifiers.
Fig. 7-38. A modification of the circuit of Fig. 7-37: C2 and C8 have
been removed from the circuit and the cathode of the diode D2 is now
grounded. The coil L is again inductively coupled to L_1.
Another ratio detector modification is illustrated in Fig. 7-38. Comparison of this circuit with Fig. 7-37 discloses two differences: C2 and C3 of Fig, 7-37 have been eliminated and the cathode of diode D2 is now grounded. With this arrangement it is still possible to apply the voltage across L to both D1 and D2. The coil L in Fig. 7-38 is still inductively coupled to L1, and consequently a voltage 180° out of phase with E1 appears across it as in Fig. 7-37.
However, in Fig. 7-38 the voltage is applied to D1 by taking the path from the plate of D1 through L8, L, and Cs to ground, then through ground and back to the cathode of D2. For D1 , the path is from the plate of D1 through C4 to ground, and then through ground to C5, L, and L2, and back to the cathode of D1. From this analysis it is easily seen how a voltage 180° out of phase with E1 is applied to both diodes as well as voltages E2 for diode D1 and E3 for diode D2. To indicate how the audio voltage output appears, the current paths of each diode are illustrated; the audio voltage appears across Cs in a similar manner to that of Fig. 7-37. The electrolytic capacitor C4, even though it is in the path of the current flow of diode D1 , offers negligible impedance, so negligible voltage drop appears across it.
THE OSCILLATOR DETECTOR
We now come to the third type of FM detector system -- one based on the principle of the locked-in-oscillator. Two methods have been developed for adapting this locked-in-oscillator for detection of FM signals. One of them is applied in the Beers receiver, developed by G. L. Beers of the Radio Corporation of America, in which a locked in oscillator performs limiting action and a discriminator circuit the actual detection. This circuit essentially employs two tubes, one for the locked-in oscillator and the other a duo-diode for the discriminator. As compared with conventional discriminators, the discriminator operates over a reduced range of frequency deviation, and therefore is called a reduced-range discriminator. Since this circuit is not available commercially, it will not be discussed here.
The other method uses a single tube that operates on the locked-in oscillator principle and accomplishes the detection of the FM signal.
This latter method has been applied commercially, and the locked-in oscillator principle will be analyzed in conjunction with this single stage detector. Both of these locked-in oscillator detector system& respond only to frequency changes in the incoming signal and not to amplitude changes.
The Single-Stage Locked-in Oscillator Detector
Philco receiver Model 46-1213 incorporates the locked-in oscillator and FM detection in a single tube which is a special pentagrid construction. A simplified schematic diagram of this single stage FM detector appears in Fig. 7-39. This network comprises three different tuned circuits including components L1C1, L,C,, and L1C8. It is noteworthy that all three circuits are resonant to the same frequency, namely the i.f.
Fig. 7-39. The locked-in oscillator detector circuit of the Philco model
46-1213 receiver. The detector circuit of Philco model 48-482 is essentially
the same.
The cathode and first two grids of the unit form a Colpitts oscillator, from which the signal of oscillator frequency is electron coupled to the plate of the tube. The oscillatory tank circuit consists of L2 and C2 in conjunction with the 33 uuf and 68 uuf capacitors. Resistor R1 in parallel with C4 forms the grid leak bias arrangement of the oscillator section. Grid number 2 serves the purpose of the oscillator anode. As grid 4 is connected to grid 2, both receive the same supply voltage and are at i-f ground potential through a bypass capacitor.
Capacitor C3 and inductor L8 in parallel compose the tuned plate circuit. As mentioned, this circuit is also resonant to the i.f., but the parallel 6800-ohm resistor R3 lowers the Q of the circuit and thus increases the bandwidth. The lower the parallel resistance, the lower will be the Q, but the resistance cannot be too low or it may completely damp out the oscillations of the tank. However, the 6800 ohms are low enough to cause the bandwidth of the plate tank circuit to increase to about five times that of the FM signal. The bandwidth of this circuit is increased by a great amount, so that the impedance of this circuit will not change over the frequency range of the incoming signal. This plate tank circuit, by the nature of its action, which will be described later, is called a quadrature circuit.
The oscillator is so designed that its grid is driven positive over a small portion of its positive half cycle of signal, and by its class C operation the r-f current flow in the tube, due to the oscillator, is in pulses of short time duration. With no input signal applied these pulses will also flow unchanged in the plate tank circuit. The amount of this current flow to the plate is controlled by the polarity of voltage ap plied to grid number 3. If grid number 3 is made negative with respect to its potential before a signal is applied, less current will flow. Conversely, if the grid is made positive more current will flow. Thus grid number 3 is a controlling factor in the magnitude of the current flow in the tube.
The vector diagram of Fig. 7-40 (A) will make this clearer. Vector e1 is the oscillator voltage that exists on grid number 1, and quadrature vector e8a, is the signal voltage that exists on grid number 3 when the incoming frequency is equal to the center i.f. Vector e_sa is drawn in quadrature because it is known that, when there is no signal input or when the signal is exactly in tune (that is, equal to the center i.f.), the pulses of current flowing in the tube do not change. Vector e3a, thus should have no in-phase or 180° out-of-phase components, and it is therefore drawn in quadrature with vector e1. Under this circum stance the input signal passes through a zero value when the pulses of plate current are at a maximum, because the free frequency of the oscillator (that is, with no input to grid 3) is the same as the center i.f.
Fig. 7-40. Vector diagrams of the voltage and current relationships
in the oscillator detector circuit of Fig. 7-39.
If there is a phase change between the incoming signal and pulses of voltage, one component of the signal voltage will be either in phase or 180° out of phase with the pulse voltage. Consequently the amplitude of the pulse voltage will increase or decrease, which in turn will cause the pulse current to increase or decrease. When the magnitude of the input voltage vector remains constant while its frequency changes, the phase relationship between the signal voltage and oscillator pulse voltage also changes. This is shown in the vector diagram of Fig. 7-40 (A), where vector e3b indicates the signal voltage when its frequency has decreased from the center i.f. and vector e30 indicates the signal voltage when its frequency has increased from the center i.f. In the former case the phase relationship is seen to be decreasing, and in the latter is seen to be increasing. If these two off frequency signal vectors are resolved into their horizontal components, the latter will be either in phase or 180° out of phase with e1. From the vector diagram the horizontal voltage component e_b of vector e_3b is in phase with the oscillator pulse voltage vector e1 , thereby effectively increasing the magnitude of the pulse voltage and hence the current.
But horizontal component e_c of vector e_3c is 180° out of phase with e1, thereby decreasing its magnitude and likewise decreasing its current pulse. Consequently, it is seen that when the frequency of the in coming signal to the number 3 grid is increased above the center i.f., the magnitude of the current pulses decreases and when the incoming signal is decreased in frequency the magnitude of the current pulses is increased.
The quadrature and oscillator circuit are so coupled together that a certain amount of voltage is fed back to the oscillator circuit. The feedback voltage is proportional in amplitude to the pulses of current and is approximately in quadrature with the voltage that would exist across the oscillator if there were no feedback. This change in quadrature feedback voltage effectively changes the frequency of the oscillator circuit. The frequency change is such that the oscillator will lock-in at a frequency equal to that of the incoming signal. As the input frequency changes, the oscillator will follow because of the lock-in effect. Let us examine the vector diagram of Fig. 7-40 (B) which clarifies this preceding analysis.
Vector e0 is the voltage that would exist across the oscillator tank in the absence of feedback. Since this tank circuit is a resonant circuit, the current flowing through the circuit will be in phase with this volt age e0 , as indicated by the current vector I. In the presence of feed back the total effective voltage across the oscillator tank is equal to the reflected voltage plus the voltage e0 existing without feedback.
This reflected voltage is approximately in quadrature leading the oscillator voltage e0, and thus has the effect of introducing an effective inductance in series with the oscillator tank inductance. This increase in inductance establishes the operating frequency of the oscillator at the center i.f. When the incoming signal is exactly equal to the center i.f., a certain amount of voltage is reflected into the oscillator circuit from the quadrature circuit of Fig. 7-39. This voltage is designated as e1 u in the vector diagram of Fig. 7-40 (B). When vectorially added with e0, it produces the resultant oscillator voltage vector e01. From the vector diagram of Fig. 7-40 (A) we see that if the incoming signal increases in frequency, the pulses of plate current will be decreased, and since these current pulses flow through the quadrature circuit, the reflected voltage to the oscillator circuit will also decrease.
This voltage is shown as vector eR2 which has a smaller amplitude than eR1 and thus has a decreased inductive effect. Under this condition of reduced inductance the frequency is increased. Due to this increase in frequency the phase lead of eR2 is slightly less than eRt· When this voltage vector eR1 is vectorially added with e 0, a new resultant oscillator voltage e02 appears across the oscillator tank. Consequently, we have a locked-in effect when the input signal increases in frequency.
When the input signal decreases in frequency, the pulses of plate current will increase as shown previously. This increase in current will cause an increased reflected voltage into the oscillator tank circuit. This voltage is designated as vector eR3 in Fig. 7-40 (B). Due to its increased amplitude it produces an increased inductive effect in series with the oscillator inductance, thereby decreasing the frequency of the oscillator. When eR3 is vectorially combined with vector e 0, the resultant oscillator voltage vector :.s e03. Therefore, a lock-in effect also occurs when the input signal decreases in frequency. In either instance the circuit is so arranged that the plate current will change linearly with respect to frequency variations. Although the reflected voltage varies in amplitude, its phase relation with respect to the resultant oscillator voltage also varies in a manner to maintain the resultant oscillator voltage substantially constant. If the oscillator voltage were not constant, its variation would produce distortion of the output audio signal.
The rate of the plate current change is dependent upon the rate of deviation, which in turn is dependent upon the audio modulating frequency. The magnitude of plate current change is dependent upon the amount of frequency deviation, which in turn is dependent upon the magnitude of the audio. Therefore, the plate current varies in direct accordance with the audio modulating frequency. Since this current flows through the plate circuit of the tube, it is possible to obtain these audio variations across part of the plate load. In Fig. 7-39 resistor R, serves as the load across which these audio variations are taken off. The 1500 uuf capacitor C6 serves as a bypass for any i-f currents. Capacitor C6 and resistor R4 are used to couple the audio output to the succeeding audio stages.
The detector response curve is such that, with a minimum input signal maintained at all times, there will be a linear characteristic over a bandwidth of about 100 khz on either side of the center i.f. If the voltage input is too small, the lock-in effect which is necessary for the proper operation of the detector will not prevail.
Amplitude-modulation effects are suppressed in the following manner: A change in the signal amplitude tends to change the pulse amplitude. Any change in pulse amplitude will cause a change in the reflected voltage from the quadrature circuit into the oscillator circuit and thus cause a change in oscillator frequency. This is further illustrated in Fig. 7-40 (B) which also reveals that changes in oscillator frequency are accompanied by phase changes between the pulse and reflected voltage. Consequently, there is a phase change between the pulse and the input signal But the oscillator undergoes only a negligible frequency change until it once again locks-in frequency with that of the incoming signal.
Thus, the change in signal amplitude produces a negligible change in pulse current which, in turn, does not produce more than a negligible change in oscillator frequency because of the lock-in effect.
From this we see that the oscillator detector is highly insensitive to AM, since changes in signal amplitude have such little effect on the pulse current. Thus like other practical FM detector systems the slight sensitivity to a-m. is far less than the sensitivity to FM.
COMPARISON BETWEEN DETECTOR SYSTEMS
Of the three detector systems discussed, it would be difficult to say that one is 100 percent better than another. Each system has advantages and disadvantages, and, according to the performance desired by certain designers, a detector system may be considered good in one respect and bad in another. By a consideration of a number of these advantageous and disadvantageous characteristics a comparison of the individual detector system will be made.
It was shown that the ratio detector system has the advantage of operating more satisfactorily on weak input signals than the other two.
A certain threshold of input signal is required for the proper operation of the limiter-discriminator circuits and the oscillator detector, but not for the ratio detector. So the ratio detector is able to suppress a-m effects of the incoming FM signal at a lower signal level than the other two detector systems.
One of the main drawbacks of the ratio detector system is that it is very critical in its balance, and, consequently, in the linearity of its output characteristic. For this detector system to suppress AM properly, a careful balance between the two diode circuits must be maintained. One of the main factors in the balancing of this circuit is the proper design of the input transformer. To produce good balancing, a high Q bifilar wound secondary is used in many circuits with good results.
If the selectivity ahead of a detector is not wide enough, then on high-frequency swings of the FM signal a strong downward a-m effect is caused, which results in distortion at the output of the detector. When the. FM signal has AM introduced into it, the combined signal has the shape of an a-m signal, with the r-f part of it varying in frequency. The peaks of the AM are referred to as upward a-m and the valleys or troughs as downward AM If the selectivity ahead of the detector is sharp compared with the swing of the FM signal, some of the FM sidebands will be cut off and the output FM signal will have amplitude changes, with the downward modulation being great enough to cause distortion in the output. This effect is particularly evident in the ratio detector circuit. This disadvantage can be overcome if the FM receiver is designed carefully, so that the i-f selectivity can not vary enough to cause this downward a-m effect.
As previously mentioned concerning the limiter-discriminator and locked-in oscillator detector circuits, their main disadvantage arises from the fact that a high threshold input signal level is required for proper amplitude limitation. This requirement means that adequate r-f and i-f gain must be provided before the input to the detector circuit. Consequently, an r-f stage usually has to be incorporated, as well as at least two i-f stages. The single-stage oscillator detector circuit has an advantage over the limiter detector circuit in that detection and amplitude limitation are accomplished within a 1-tube circuit.
If a really good over-all performance of the limiter-discriminator detector is desired, it is best to use a cascade limiter arrangement instead of a single limiter stage. By this cascaded circuit the maximum possible range of linearity is obtainable, and yet the output level from the detector is kept fairly constant. By choice of the proper time constants of the limiter circuits, impulse noise will have the greatest amount of limiting in this detector circuit. The discriminator trans former also should have a careful design, so that the proper amount of coupling exists between the primary and secondary. The circuit should be well balanced, and adequate means be provided to correct any slight variation in the balancing. In practice, it has been found that the limiter-discriminator detector is the least critical so far as balance of the detector system and linearity of the output are concerned.
Many of the pros and cons of the limiter-discriminator detector also can be ascribed to the locked-in oscillator detector. The latter sys tem has the additional advantage that, by the intrinsic nature of its circuit, it is the one least affected by interference. However, this sys tem primarily depends upon the lock-in nature of the oscillator, and, if for any reason this cannot be adequately maintained, the system will not operate satisfactorily. Therefore the oscillator circuit must be well shielded from any effects other than those necessary for the shifting of the oscillator frequency to produce the lock-in effect. To insure this, the filament supply of the locked-in oscillator circuit, as well as that of the other electrodes, should be properly bypassed, so that no outside FM or p-m effects are introduced into the oscillator circuit by hum or other voltage fluctuations.
THE AUDIO SYSTEM
The audio stages of FM receivers are very similar to those of a-m receivers; the main differences between the two audio systems are in their response characteristics. The audio stages in both follow the detector stage, and the intelligence reproduced for injection into the audio systems is the same as that at the transmitting studio. Because of the FCC regulation on modulating frequencies, the normal maxi mum audio frequency that can be passed by most a-m stations is only 7.5 khz, but with FM, as was shown at the beginning of the guide, a maximum frequency of 15 khz can be passed.
This difference in possible frequency response is what causes the audio systems of a-m and FM receivers to differ in basic design. In many a-m receivers, the sets are designed with a response to about 3000 or 3500 cycles per second, although transmitters use higher audio frequencies; however, any system that does not allow full reproduction of all the audio modulating frequencies is not considered high fidelity.
In a-m systems, the audio frequency limitation starts at the transmitter where usually a maximum of only 7.5- khz audio can be passed.
Consequently, it is generally unnecessary for a-m receivers t0 have an audio response better than 7.5 khz, although the higher priced receivers may have such a range.
In FM the situation is different. By the nature of this type of modulation, the high frequencies involved, and the channel separation, FM transmitters can and are required by the FCC to have an audio frequency response to 15 khz. This means that in any FM program that is transmitted, especially non-recorded musical programs, the high frequency audio notes will be contained in the audio FM of the carrier signal. Therefore it is reasonable for the receiver used for FM reception to have an audio-frequency response greater than 7.5 khz and as high as 15 khz. The sales price of a receiver is a factor in determining the presence of the components which enables realization of the full 15- khz range of frequencies since such components increase the cost of the set.
As far as FM is concerned, high fidelity means that the system is capable of reproducing, with a minimum of distortion, the maximum range of audio frequencies that is considered necessary for good listening. For an FM receiver to have true high fidelity, the audio system should have a frequency response to 15,000 cycles. Such a system is very expensive compared with that needed in a-m receivers, because it entails the use of high-fidelity output transformers, speakers, and audio coupling networks.
Most audio systems incorporate one or two audio voltage amplifiers and an audio power output stage. These are the stages which should have a frequency response to 15,000 cycles for maximum high fidelity.
Since most FM receivers are combined with a-m receivers, the same audio network is used for both. In many instances the audio system is designed for the usual a-m reception, and the FM section must use this same audio system with a consequent loss in fidelity. A high fidelity audio system designed for use in an a-m, FM receiver would add greatly to the cost of the set, perhaps putting the price above that which the average consumer is willing, or able, to pay.
To attain proper high fidelity, the coupling between the audio stages, whether RC, impedance, transformer, or any combination of these, should have a frequency response that is flat up to 15,000 khz. Besides the coupling arrangements, the use of the tubes and other associated circuit components should be of the proper design, so that a minimum amount of distortion will be introduced. There are various means of doing this which are also employed in a-m receivers. One of them is to use a push-pull power output circuit; however, these circuits are not necessarily the criterion, as single-tube output circuits are in use which give as good a response.
Following the power output tube is the output transformer, an essential component in matching the impedance of the last stage to the speaker for a maximum transfer of energy. As mentioned, it is necessary to have the frequency response of this unit just as flat to 15,000 cycles as the preceding audio stages. A poor output transformer is one of the main causes for distortion in the output of a receiver.
The last unit in the audio system is the loudspeaker, which, if it is to be a high-fidelity component, is quite an expensive item compared with the rest of the FM set. True high-fidelity, distortionless loud speakers may cost as much as $200. Other types of good loudspeakers are not so expensive, but in general they cost much more than the conventional loudspeaker used in an a-m receiver. High-fidelity loud speakers generally are larger in size than the loudspeakers used in most conventional a-m receivers; the baffle used with these loud speakers, therefore, must be larger. The cabinet size is thus increased and this, too, adds to the price of an FM receiver.
The use of the necessary high-fidelity loudspeaker thus makes difficult the problem of producing an FM receiver within the economic limitation of most consumers. If a 100-percent high-fidelity FM reception system is desired, all the audio stages plus the loudspeaker have to be high fidelity and free from most distortion. Such a system, undoubtedly, would be priced too high for the average buyer. An alternative is to reduce the fidelity a little by using a good but not a true high -fidelity loudspeaker so that the cost is reduced. It is quite difficult, however, to reduce the over-all cost by very much, because most good loudspeakers are relatively expensive, even when they cost $50 as compared with the previously mentioned $200. The general public will not pay the considerable price necessary even for a set that employs a $50 loudspeaker.
Thus we see that the FM receiver manufacturer is in a dilemma.
He can produce a 100-percent high-fidelity FM receiver, but he knows that the loudspeaker, plus some of the other components, make the sales price too high for the average consumer. This has been proved by the fact that the public has bypassed the high-fidelity FM receivers that are on the market in favor of the lower priced FM, a-m combinations. Yet the manufacturer still wants to put FM on the market at a medium price. The easiest way to do this is to use a loud speaker that is within the price range of those used for AM This, of course, definitely eliminates the use of a perfect over-all high-fidelity system. Nevertheless, this is sometimes done, and as a consequence of the loudspeaker employed, many receivers do not have too good an audio response characteristic.
Even within the price range of the general public, the audio response of an FM receiver can be improved beyond the usual range of a-m receivers, although what is attained is not completely high fidelity.
Even though loudspeaker costs increase according to the increase in audio-frequency response and distortion-free reproducing ability, a compromise can be reached. The frequency response may be designed to be somewhere from 6 to 8 khz, and a loudspeaker that can handle these frequencies employed. In fact, the cost may be reduced some what more by not paying too strict attention to the distortion characteristics of the loudspeaker. This is done in some FM receivers.
Although these receivers may not be 100-percent high fidelity up to 15,000 cycles, at least they are consistent throughout their audio sys tem, including the loudspeaker, concerning the frequency response which is much better than the low audio response of a-m systems.
They are, nevertheless, in the price range of the average consumer.
It is hoped, and believed, that some day soon receiver and speaker designers and manufacturers will be able to produce high-fidelity loudspeakers and FM receivers in sufficient quantity, so that the radio buying public will be able to buy really high-fidelity FM sets within the price range of all. The beauty and clearness of true distortionless high-fidelity reproduction cannot be appreciated unless one has had the personal experience of listening to such audio reproduction.
FM TUNERS
Appearing on the market in recent years have been quite a few FM receivers that do not contain any, or contain only a part of, the audio system. These units in reality, are not true receivers, since they can not reproduce within themselves the necessary audio sound. Such units are called tuners or converters. Neither term has been given preference, but the former has been used widely during the past year or so.
The main reason such units are in demand is that they can be used with any audio system, whether part of a phonograph amplifier, the audio system of an a-m receiver, or a specially built high-fidelity audio amplifying system. To make these FM tuners available to the public, so that they can also receive FM stations through their a-m receiver (although the reception will not be high fidelity), the trend is toward producing tuners priced well within reach of everyone. In most of these tuners all the necessary design features of regular FM receivers are incorporated, so that nothing is lost as far as the operation of the tuner is concerned. In fact, many of these FM tuners contain designs that are unique and deserve careful consideration.
In this section on tuners we will analyze two new units that have just appeared on the market at the time of this writing. The units are priced about the same, and are about the same size, but their over-all designs are different. We will not discuss each unit completely but only those features that are considered unique in design and also that warrant special mention. Neither unit includes any audio stage at all.
One unit uses the limiter-discriminator method of detection, and the other unit uses the ratio detector. The two methods of tuning, as well as the types of i-f transformers employed, are quite different.
The Edwards Fidelotuner
This FM tuner employs the limiter-discriminator method of detection. The complete schematic for this tuner is illustrated in Fig. 7-41 (A), and a drawing of the top chassis layout is shown in Fig. 7-41 (B). Five tubes are employed in the unit, with three 6SH7 sharp-cutoff pentodes used as the first and second i-f amplifiers and also as the limiter. The discriminator uses the duo-diode 6H6 tube, and the 6J6 miniature duo-triode is used as the converter. Note that no separate r-f stage is employed, even though the tuner uses the limiter-discriminator detector network. The reason for this is that the signal-to-noise ratio of the set is high, and the signal input to the limiter is sufficiently great to give satisfactory performance.
A few interesting features about this tuner warrant special mention, and the most interesting is the method of tuning. The tuning in this unit is inductive, but does not involve any permeability tuning.
Open-wire parallel transmission lines are used for varying the inductance of the r-f input and oscillator section of the converter tube.
Consequently, this type of tuning is called transmission line tuning.
The variation in inductance is obtained by running a shorting bar along the parallel line, thus changing the inductance offered by the changing length of each line. A schematic representation of this tuning system is shown in the upper left part of the diagram of Fig. 7-41 (A), and its physical layout on the chassis is seen in Fig. 7-41 (B). To show exactly how this tuning unit works, a pictorial view of it, showing both the oscillator and r-f tuning lines is illustrated in Fig. 7-42 (A), and a schematic representation of how the tuning lines form the tank circuit is illustrated in Fig. 7-42 (B).
[Manufactured by FM Specialties Inc., of New York, N. Y., under the brand name of "Edwards Fidelotuner."]
Fig. 7-41. Schematic of the Fidelotuner, which employs the limiter-discriminator method of detection. The chassis layout of the set is shown in (B).
Since the frequencies involved are quite high, the inductive and capacitive values that form the r-f or oscillator tank circuits are very small. Thus, it is possible to use the difference in inductance in the varying length of transmission line as a means of high-frequency tuning. The actual formation of the oscillator and r-f tank circuits uses a fixed inductance and capacitance, which are connected together at one end, in conjunction with a transmission line in each case. From the other ends of these components, the two transmission line leads ...
Fig. 7-42. Pictorial layout of the tuning unit of the Fidelotuner (A)
and a schematic representation is shown in (B). The shorting block carries
two shorting loops, one for each pair of lines.
... are connected. This is indicated in Figs. 7-41 (A) and 7-42 (B). A metal shorting loop is rigidly placed across each set of lines, and the position of this loop on the lines determines how much extra inductance is added to the tank circuit because the shorting loop completes the tank circuit.
The schematic of Fig. 7-42 (B) will make this somewhat clearer.
The inductor L and capacitor C represent the fixed quantities in the circuit and are connected together as shown. Let us assume that the shorting loop is at the position indicated. To complete the tank circuit, the current must travel through one part of the line, then through the shorting loop, and finally through the other part of the line back to the circuit. It can travel from L, then to points W, X, Y, and Z and then to capacitor C; or it can go from C in the direction of Z, Y, X, W, and then to inductor L to complete the circuit. It is the added inductance of the parts of the line from W to X and from Y to Z, plus that of the shorting loop in conjunction with the distributed capacitance of the line parts, which determines the final resonant frequency of the tank circuit. The lengths W to X and Z to Y are equal due to the mechanical nature of the system. The shorting loops are very thin phosphor bronze springs rigidly mounted in a Lucite block, which in turn is connected to the dial cord and pulley arrangement for the proper tuning of the set. This is seen from Fig. 7-42 (A). Both the oscillator and r-f shorting loop are placed in the same shorting block, so that they are effectively ganged together and are variable as one upon tuning of the unit. The r-f and oscillator lines are spaced far enough apart, so that no serious interaction between these circuits is possible. The lines are all made of hardened brass tubing 5 ½ inches long and 1/8 inch in diameter, and they are covered with a thin layer of silver plate. The shorting loop makes a hair line contact with the lines, and this contact is maintained in a rigid state due to the high spring tension of the phosphor bronze and the shape of the loop. The lines themselves are fixed in position on two lucite mounting blocks fastened to a mounting plate.
Although there have been other types of inductive tuning circuits, we believe the type illustrated here to be of a unique yet very simple design, with which proper tuning can be obtained with the least amount of trouble.
There are two other interesting features to this tuner. One is that a duo-triode 6J6 miniature tube is used as a converter, in which one triode section serves as the mixer and the other as the oscillator.
Although it has not been mentioned previously, two triodes used for a system of frequency conversion have one of the highest, if not the highest, signal-to-noise ratios among converter systems. This high signal-to-noise ratio obviates the need for an r-f stage to increase the signal input.
The other interesting thing is that the oscillator plate current flows through the transmission line and shorting loop as seen in Fig. 7-41 (A). This current flow is a good reason why the loop contact to the lines must be tight.
The i-f and discriminator transformers are of the special kind that use magnetic shielding in conjunction with permeability tuning, as discussed in the section dealing with the i-f system and Fig. 7-12.
The terminal strip shown in Fig. 7-41 (A) is on the back of the chassis, and it affords a quick means of alignment as well as a fast and easy method of connecting the tuner to an a-m radio or amplifier.
The meter method of aligning the i.f.'s and discriminator is also indicated in Fig. 7-41. At the end of Section 8, the factory alignment procedure for this tuner will be found, using a combined meter and visual method.
Fig. 7-43. Schematic diagram of the Pilotuner with the alignment frequencies
and layout showing locations of the trimmers and variable core adjustments.
The Pilotuner
This FM tuner also employs five tubes but uses all miniature sizes with a 6AL5 duo-diode employed as a ratio detector. Three other tubes are 6BA6 pentodes used as an r-f amplifier and for the first and second i-f amplifiers. A 6BE6 pentagrid tube is used as a converter. The schematic diagram for this tuner appears in Fig. 7-43, and a top chassis view appears in Fig. 7-44.
From the schematic diagram, a few interesting things are noted.
First of all, the coils marked as P 8 and P10 are shown to be variable without any internal core adjustment similar to the i-f transformers.
Fig. 7-44. Top view of the Pilotuner chassis showing locations of the
tubes, transformers, and other components.
Looking at the ratio detector transformer, we note that the primary does not contain any fixed capacitance for tuning purposes. The core going through the primary coil is also extended to the coil directly underneath this primary, indicating that both coils are wound on the same coil form. These features plus some other interesting properties of the tuner will now be discussed in detail.
"Manufactured by the Pilot Radio Corp. of N.Y.C. under the brand name of "Pilotuner" and listed as model T-601.
In aligning the oscillator and r-f stages on the low end of the FM band, the tuner provides for inductive padding. The r-f section's inductive padder is designated P10 in Fig. 7-43 and also appears in the picture of Fig. 7-44. The interesting thing about this coil is that the variation of its inductance is simply made by changing the spacing between the coil windings. From Fig. 7-44 the r-f padder coil P10 is seen to consist of only two turns, and the space between these two windings is varied by means of a screw. The screw is kept in place by means of a piece of Bakelite tubing which is threaded on the inside.
Upon clockwise rotation of the screw, the coil spacing is decreased, thereby increasing the inductance, and upon counterclockwise rotation the spacing is increased, thereby decreasing the inductance.
Fig. 7-45. The construction of the oscillator padder used in the Pilotuner.
The threaded portion of the screw adjustment was omitted from the drawing
for reasons of clarity.
The oscillator padder P8 is located on the underside of the chassis and also consists of about two turns of wire. A drawing of this oscillator padder construction is shown in Fig. 7-45.
The variation of the inductance of this padder is somewhat different from that for the r-f padder P 10. The spacing between the coil turns of this padder P 8 are kept stationary, but the effective magnetic field about the coil is varied.
Between the spacing of the winding is a thin strip of Bakelite, which is mounted above the chassis by two brass spacers. This Bakelite strip runs through the spacing of the coil and contains a threaded hole where the center of the coil winding appears. Into this hole is inserted a screw which has a round metal plate attached to the underside of the screw head. This metal plate is about 7 /8 of an inch in diameter and 1/16 of an inch thick. Its diameter exceeds the diameter formed by the coil winding.
Inductance of this oscillator padder coil is varied by turning the screw in or out. This inductance variation is explained as follows: The metal plate acts as a short-circuited secondary winding to the oscillator padder coil, which effectively acts as the primary winding.
The short-circuited secondary reflects a reactance into the primary which is capacitive. This reflected capacitance is effectively in series with the primary inductance, and therefore reduces the effective value of the inductance. With the movement of the screw and, hence the metal plate, the coupling between the plate and coil changes. As a result the mutual reactance which exists between the plate and coils also varies, which in turn varies the reflected capacitance into the coil circuit. Turning the screw so that the plate moves toward the coil increases the coupling and also increases the reflected capacitance into the primary. This increased reflected capacitance decreases the effective inductance of the oscillator tank circuit, thereby increasing its frequency. Adjusting the screw so that the plate moves away from the coil increases the effective inductance and, hence, decreases the frequency of operation of the circuit.
Fig. 7-46. How the ratio detector transformer of the Pilotuner is constructed
(A) and its schematic dia gram is shown at (B). Note the bifilar wound
secondary.
The primary of the ratio detector transformer has no fixed capacitance across its coil, but it still forms a tuned circuit with the output capacitance of the last i-f tube, stray wiring capacitance, and the inherent capacitance between the coil windings. The amount of this capacitance is high enough at the 10.7 -mhz i.f. to form a tuned circuit with the primary coil. Besides this feature, the physical construction of the ratio detector transformer is quite interesting. A picture of the construction of the transformer is shown in Fig. 7-46 (A), and a de tailed schematic drawing is shown in Fig. 7-46 (B). This schematic drawing is somewhat different in appearance from the ratio detector circuit which is shown in Fig. 7-43, but they are effectively the same in circuit operation.
The tertiary coupling coil L is wound over one end of the primary coil L1 which appears on the top coil form of Fig. 7-46 (A). Because of this close coupling between L and L1 , the voltage appearing across L1 is effectively in series with L but 180° out of phase. The secondary is wound on the lower coil form. Both coil forms are mounted on a flat brass strip, and each coil form contains a variable core to change the effective inductance of the coils. The cores are adjustable by means of two screws which appear on one side of the brass strip as seen in Fig. 7-46 (A). These screws are about 1% of an inch apart.
Particularly noteworthy in the construction of this transformer is the bifilar winding of the secondary coil. This bifilar winding is obtained by using a very closely spaced twin lead, which is insulated with a transparent plastic and is wound around the bottom coil form with both ends open. The plastic insulator at the ends is split, ex posing two bare wires at each end. These two wires of the twin lead are shown in the drawing of Fig. 7-46 (B) as coil A to E and coil B to D. A trace of the coil circuit across capacitor C starting at point B, would travel from point B to D, then from F to A to E. Therefore, each part of the bifilar winding L2 contributes to the inductance for the secondary tuned circuit.
Coil L must be tapped to the center of the secondary inductance to maintain the balance for the detector circuit. This tap is obtained by connecting one end of L to point F, the junction of A and D, as seen in Fig. 7-46 (B). The complete inductance of the secondary is made up of the coil winding BDF AE, and since the length of BD is approximately equal to length AE, it is readily seen that by connecting to point F, the junction of AD we are effectively center-tapping the coil.
It is advisable that, if any trouble is suspected in the ratio detector circuit that lies within the transformer shield, the serviceman not attempt to take the circuit apart. It is suggested that the manufacturer or one of his representatives be contacted.
QUESTIONS
SECTION 7
7-1. What is one of the most important functions of an input transformer in an FM receiver?
7-2. What are two of the most common methods of making the r-f tuned circuits in FM receiver broad-band?
7-3. What are some of the advantages in using an r-f stage in an FM receiver?
7-4. Why is the signal-to-noise ratio low at the output of the converter (or mixer) stage when no r-f tube is employed?
7-5. a. Is image frequency interference possible within the FM band of today when the i.f. is 10.7 mhz? Why?
b. Is image frequency interference possible within the FM band of today when the i.f. is 4.5 mhz? Why?
7-6. What are some of the factors that determine the choice of a good r-f tube for an FM receiver?
7-7. What effect will appreciable lead inductance in the cathode circuit of the r-f tube have on the match between receiver and antenna?
7-8. Why is it preferable to use separate mixer and oscillator tubes in an FM receiver rather than a converter tube?
7-9. Some converters employ a 0.5 to 2-µµf capacitor between the signal grid and oscillator grid. What is its purpose?
7-10. How does humidity affect the stability of an oscillator?
7-11. a. Why are coils and especially capacitors with low temperature coefficients desired in high-frequency oscillator circuits? b. How does a capacitor with a low temperature coefficient help stabilize an oscillator to an increase in temperature?
7-12. Define sensitivity as applied to a radio receiver.
7-13. What is meant by the half-power points of a response curve?
7-14. a. What are some of the terms used to express the selectivity of a tuned circuit when the range of acceptable frequencies is small? b. Repeat part (a) for appropriate selectivity terms when the range of acceptable frequencies is large.
7-15. a. When the range of acceptable frequencies of a tuned circuit is small, is the Q of the circuit high or low?
7-16.
7-17.
b. When the range is large is the Q high or low? What is the minimum required bandwidth of i-f circuits of FM receivers? What are the two main purposes of i-f systems in superheterodyne receivers?
7-18. a. What is meant by a stagger-tuned i-f system?
b. What are two of the chief drawbacks to such a system when used in FM receivers?
7-19. Of the three main FM detectors, which perform the dual function of FM detection and a-m rejection?
7-20. In FM receivers employing the limiter-discriminator method of detection, between what two stages is the limiter system found?
7-21. What are the proper operating conditions for a limiter tube in an FM receiver?
7-22. Why do the harmonics introduced by the clipping action of the limiter circuit cause no trouble nor interference?
7-23. What causes grid current to flow in the limiter tube?
7-24. a. In Fig. 7-15 (A), if C equals 100 µµf and R equals 50,000 ohms, what is the time constant of the circuit?
7-25.
7-26.
7-27.
b. Is the time constant for part (a) considered appropriate for the proper operation of a limiter stage of an FM receiver with an i.f. of 5 mhz? Explain your answer.
In a limiter stage does the positive peak clipping of the input signal take place in the grid or plate circuit of the tube? Where does negative peak clipping take place? Neglecting interference as a cause of amplitude variations in the FM signal, what part of the FM receiver contributes appreciably to changing the amplitude of the FM wave before it enters the limiter stage? If after a few minutes of operation of a limiter tube, whose plate current-grid voltage curve is represented by the curve of Fig. 7-15 (B) on page 281, the input signal on the grid of the limiter has a peak-to-peak voltage of 6 volts with a peak-to-peak noise voltage of 4 volts super-imposed on that of the signal:
a. Would any of this noise appear in the output of the limiter? Give a complete explanation.
b. Would any of the noise be eliminated? Give a complete explanation.
7-28, a. In FM receivers having limiter-discriminator detectors, where is the ave voltage ( if any) usually taken from?
b. Is the signal from this point ever put to any further use?
7-29, a. In a two-stage cascade limiter system, what is the purpose of a large RC time constant for one stage and a smaller one for the other?
b. Does it matter which limiter stage has the high or low time constant?
7-30. If the required threshold voltage for the proper operation of a limiter tube is 3.3 volts and the over-all r-f and i-f gain equals 75,000, what is the minimum value of input signal voltage to the first r-f tube for proper operation of the set?
7-31, What is the polarity of the output signal at point A with respect to ground in Fig. 7-22, when the instantaneous frequency of the FM signal output from the limiter is:
a. Equal to 10.7 mhz, the center i.f.?
b. Above the center i.f.?
c. Below the center i.f.?
7-32. In Fig. 7-25, the curve at the lower left-hand side of the graph indicates the frequency deviation of the FM signal in put to the discriminator at different instants of time. The audio out put signal is seen to be changing in amplitude in direct proportion to the amount of frequency deviation of the input signal. Why must this be true in order for the discriminator to reproduce the true audio signal?
7-33. What other names are given to the Foster-Seeley discriminator circuit?
7-34. How is the signal from the plate circuit of the limiter coupled to the discriminator circuit of Fig. 7-26?
7-35. Explain how the voltage across L in Fig. 7-26 is approximately equal in magnitude and of the same phase as that existing across the primary circuit of the discriminator transformer?
7-36. In a double-tuned transformer circuit where the secondary is center tapped to ground and an input signal is applied across the primary, what is the phase relationship at resonance between: 7-37.
a. The induced secondary voltage and secondary current?
b. The individual voltages developed across each half of the secondary?
c. Each secondary voltage and the induced secondary current? In the circuit of Fig. 7-26, if the conditions are such that the induced voltage and current in the secondary are out-of phase with each other, what happens to the phase relationship between this induced current and the individual secondary volt ages of Ei and E,,"?
7-38. In the circuit of Fig. 7-26, what is the polarity at the cathode of diode D1 with respect to ground when the resonant frequency of the discriminator transformer is:
a. Lower than the instantaneous frequency of the incoming FM signal?
b. The same as that of the instantaneous frequency of the incoming FM signal?
c. Higher than the instantaneous frequency of the incoming FM signal?
7-39. In the circuit of Fig. 7-26, what is the polarity at the cathode of diode D1 with respect to ground, when the phase of the secondary current which flows as a result of the induced secondary voltage:
a. Lags this induced voltage?
b. Leads this induced voltage?
7-40. Which diode in Fig. 7-26 has the greater current flowing through it when:
a. The secondary current lags the induced voltage?
b. The secondary current leads the induced voltage?
7-41. In the circuit of Fig. 7-26:
a. If the output FM signal from the limiter is constant in amplitude at all instants of time, will the magnitude of voltages Ei, E1, or E: change?
b. Will the magnitudes of Ell and E 3 always be equal to each other?
7-42. Give two important reasons why it is best to have the linearity of the S-shaped characteristic curve of the discriminator at least 200 khz wide, even though for 100-percent modulation 150 khz is the minimum width of linearity?
7-43, Is the inductance Lin Fig. 7-26 a prerequisite for the proper operation of all types of discriminator circuits? Why?
7-44, In the circuit of Fig. 7-32, across what circuit component does practically all of voltage E1 essentially appear so that this voltage is:
a. Applied to diode D1 ?
b. Applied to diode D2?
7-45. In the circuit of Fig. 7-33, across what circuit component does practically all of voltage E1 essentially appear so that this voltage is:
a. Applied to diode D1 ?
b. Applied to diode D2 ?
7-46. What is a quick method of distinguishing between a discriminator detector circuit and a ratio detector circuit?
7-47. In Fig. 7-35 on page 314:
a. What is the purpose of the battery?
b. Why is it impractical to use the battery?
c. What is used in place of the battery in the circuits of today?
7-48. What point in the ratio detector circuit of Fig. 7-36 is suitable for tapping off an ave signal? Explain.
7-49. What is the reason for inserting resistor R3 in the quadrature circuit of Fig. 7-39? Why?
7-50. a. When the instantaneous frequency of the FM signal input to the third grid of the locked-in oscillator detector is higher than that of the center i.f. of the system, does the magnitude of the pulses of current in the tube increase, decrease, or remain the same?
b. What happens to these pulses when this incoming FM signal has an instantaneous frequency less than the center i.f. of the system.
7-51. In the vector diagram of Fig. 7-40 (A), vector e8a represents the incoming FM signal when its instantaneous frequency is equal to that of the center i.f. of the transformer of the locked-in oscillator detector. Is the instantaneous frequency rep resented by vector e3c higher or lower than the center i.f. of the circuit? What about the instantaneous frequency represented by vector e3b?
7-52. Why is it possible to obtain the audio output voltage from the locked-in oscillator detector across the plate-load resistor of the circuit?
7-53. Of the three principal FM detector systems-namely, the limiter discriminator circuit, the ratio detector circuit, and the locked-in oscillator circuit:
a. Which two circuits depend upon a certain threshold voltage of in put signal for proper operation?
b. Which system is very critical in its balance?
c. Which detector system is the least critical in its balance?
d. Which system requires more careful shielding?
7-54. What does high fidelity mean as far as FM is concerned?
7-55. a. What component in the audio system of an FM receiver, because of its high price, puts a practical limitation on the high-fidelity possibilities of the receiver?
b. What component other than the loudspeaker is a major cause of distorted output from an FM receiver?
7-56. a. If in the schematic of Fig. 7-42 (B), the inductance L equals 0.5 microhenry, the capacitance C equals 4.2 µ.µ.f (neglecting the distributed capacitances in the circuit), the inductance from that part of the line from W to X equals 0.1 microhenry, and the inductance of the shorting loop equals 0.05 microhenry, what is the resonant frequency of the circuit?
b. As the shorting loop is moved to the left of the schematic of Fig. 7-42 (B), what happens to the resonant frequency of the tuned circuit? Why?
7-57. In some FM receivers and tuners, an r-f stage is not employed when a duo-triode tube is used as a converter. What advantage does a converter system afford?
7-58. The oscillator padder coil, P 8, of the Pilotuner, Fig. 7-43, has its inductance effectively varied by changing the position of a metal plate a short distance away from it.
a. Does this metal plate change the physical spacing of the coil windings of the oscillator, P 8, to produce the effective inductance change?
b. How does moving the plate away from the oscillator padder coil affect the frequency of the oscillator? Explain.
7-59. In Fig. 7-43, the primary S2 of the ratio detector trans former does not appear to have any fixed capacitors across its coil, yet the coil is shown as a tunable unit. How is this tuned primary circuit formed?