# Y Low Frequency Oscillator

L. B. Arguimrau

RECENTLY the increased attention paid to the operation of - broadcast circuits at the lowest audible frequencies has created a demand for measuring equipment to cover this range. With this in view the General Radio Company has extended the frequency range of its Type 377 Low-Frequency Oscillator to include all frequencies lying between 25 cycles and 70,000 cycles.

At the same time that this change was contemplated it was thought desirable to design the new oscillator for use with one particular type of tube to operate under fixed battery conditions. In the Type 377 Low-Frequency Oscillator the choice of tubes and their operating points was left to the user for the added flexibility thereby obtained. In many cases, however, the general recommendations given were not followed; high-power tubes were used with incorrect operating conditions, and the oscillator was seriously overloaded with resultant distortion. It was believed that this situation could best be avoided by settling on definite operating conditions, providing an instrument with the power level best adapted to the average user, and employing additional amplification if needed in special cases.

Before going into a discussion of the characteristics of the new Type 377-B Low-Frequency Oscillator, it may be of interest to outline briefly the theory of its operation so that the necessity for certain adjustments may be better appreciated.

Consider the amplifier circuit shown in Figure 1. In accordance with the usual tube theory, a small sinusoidal voltage eg applied to the grid will be amplified in the customary way, giving rise to a slightly distorted plate-current wave. Most of the higher harmonics are filtered out in the tuned circuit so that the voltage es across the transformer secondary will be essentially sinusoidal. If the resistance R is properly chosen, this secondary voltage can be made exactly equal to the applied grid voltage. When this has been done, we may connect the circuit as shown in Figure 2, and oscillations will be sustained. The departure of the grid voltage from a sine wave depends upon the magnitude of the swing and operating point and also upon the selectivity of the tuned circuit. Just as in the usual amplifier theory, the least distortion will be present when the grid-voltage-plate-current characteristic of the tube is essentially straight.

It has been found that the grid current in an oscillator tube provides a remarkably simple and accurate means for estimating the amplitude of oscillation. For a given grid current, we can say that the grid swing has a definite value, regardless of the characteristic of the coil and the frequency chosen. Hence, if R is adjusted to give a predetermined grid current, the grid will be operating at a given swing, a consideration of prime importance when the signal is to supply an amplifier tube.

For those who may be interested, a brief sketch of the details is given in Appendix A. The important point to be noticed is the close analogy between an oscillator of this type and a tuned amplifier. F\irther application of this analogy can be made to account exactly for the waveform obtained and for the variation of frequency with operating point, but this is too involved to be of interest.

The circuit finally adopted for the Type 377-B Low-Frequency Oscillator is shown schematically in Figure 5 where it will be noticed that this follows the outline given above. An adjustable ioo,ooo-ohm rheostat R is used as a feed-back resistance and it should be adjusted to give a grid current of 30 micro-amperes (as indicated on a 0-200 micro-ampere meter mounted on the panel). The 50,000-ohm plate-supply resistance r provides a good coupling device and at the same time limits the oscillator plate current to about 2 milliamperes. In Figure 3 a tuned-plate oscillator is shown. In practice it was found desirable to use a Hartley circuit for the lower end of the range, using the tuned-plate circuit at higher frequencies.

It will be noticed that the output power is taken off across a slide-wire potentiometer in the plate circuit of the last tube. This was done for two reasons: to prevent interaction with the oscillator tube and to prevent the waveform from varying excessively with the output setting. It should be noticed that this arrangement makes the effective output impedance of the oscillator depend upon the potentiometer setting. In a few isolated cases (such as measurements on harmonic production in non-linear circuits) this is undesirable, but for all ordinary purposes it causes no difficulty. Measurements on harmonic production and

allied phenomena require special precautions, and no ordinary coupling device should be used without a careful consideration of its effect 011 the circuit being studied. In all measurements on linear circuits where the effect of harmonic flow is of no interest, the impedance of the generator tube need not be considered; by proper connections, any source impedance from zero to an arbitrarily large value can be simulated. This matter is treated at length in the November issue of the Experimenter.

Figure 3. Front of panel view of Type 377-B Low-Frequency Oscillator. It is mounted in a heavy hinged-back cabinet

Figure 3. Front of panel view of Type 377-B Low-Frequency Oscillator. It is mounted in a heavy hinged-back cabinet

Recognizing that the needs of different users will vary somewhat, provision has been made for using either one or two amplifier tubes. If only a small amount of power is needed and it is desired to reduce battery drain to a minimum, only one tube need be used; the total plate current will then be about 10 milliamperes. Without any change in the circuit, a second tube may be added in parallel; in this case, the plate current will then be about 16 milliamperes. Typical output characteristics for these two cases are shown in Figure 6.

A series of measurements has been made with a harmonic analyzer to determine the dependence of waveform on operating conditions. These measurements show that the voltage wave given by the oscillator tube itself contains no harmonic having an amplitude larger than 0.5 per cent, of the fundamental. On the other hand, the amplifier has been designed to deliver the maximum amount of power con

Figure 4. This photograph shows the internal construction of the Type 377-B Fow-Frequency Oscillator sistent with the usual requirements on waveform. Harmonics introduced by the amplifier may amount to 3 per cent, of the fundamental. If much better waveform is required, it is necessary to reduce the signal level at which the last stage is operating. This may be accomplished by substituting a 1.0 megohm resistance for the 0.5 megohm unit in the place marked A in Figure 5. This change reduces the harmonics to 1.0 per cent, of the fundamental, if the load resistance is not less than 8000 ohms.

The adjustment of the feed-back resistance for constant grid current helps to minimize the changes in frequency due to operating condition. Measurements on one particular oscillator showed that for frequencies in the neighborhood of 40 cycles, changes in plate battery amounting to 25 pet-cent. made a change in frequency of less than 0.1 per cent, when the feedback was readjusted to give 30 microamperes. If no adjustment was made, the departure was less than 0.3 per cent. If the grid current was allowed to depart by 10 micro-amperes from the rated value, the frequency drift was not more than 0.3 per cent. Changing the oscillator tube gave rise to similar differences. At cycles the changes were about twice as large as those mentioned (i.e., a maximum of about one-tenth of a cvcle).

These figures refer to frequency changes covered by tubes and operating points. Ageing effects may be larger but they should never exceed 3 per cent.

Appendix A

Consider the amplifier circuit shown in Figure 1. Assume a sinusoidal voltage eg impressed on the grid. When this swing is very small, we may treat the circuit in accordance with the usual tube theory, considering the tube as a sinusoidal generator neg in series with a resistance rp, the internal plate impedance of the tube. Neglecting the effect of the shunt resistance r and the condenser C* we may write for

* By the use of Thevenin's Theorem we may eliminate the effect of the parallel resistance r and the condenser C by making the substitutions,

using these new quantities in place of/x and rp. It will be noticed that this has little effect when

200 1000 iqooo

LOAD RESISTANCE (OHMS)

Figure 6. Power output of the oscillator as a function of load resistance

4-OpOO

200 1000 iqooo

LOAD RESISTANCE (OHMS)

Figure 6. Power output of the oscillator as a function of load resistance

4-OpOO

LOAD CHARACTERISTICS TYPE 377-B LOW-FREQUENCY ~ OSCILLATOR

the magnitude of the voltage across the secondary,

Z1 is the impedance of the parallel

circuit, and — is the turns ratio of the n i transformer. Even though some distortion is produced by variations in plate resistance introducing harmonics in the plate current, the voltage across the tuned circuit Zx will be very nearly sinusoidal. (Actually, the harmonics at this point can be kept below o.^ per cent.) It is to be noted that for a given tuned circuit and tubes (i.e., given Zy and rv), the secondary voltage can be regarded as a function of the coupling resistance R. In particular, if Zx is sufficiently high we can choose R in such a manner that the secondary voltage ¿*2 is equal to the applied grid voltage eg. When this has been done we may rewrite (1):

Actually, of course, this is not the whole story. Tn practice the amplitude will adjust itself until the dynamic plate impedance satisfies the above equation. By properly choosing R, however, we can select any such equilibrium positionj thereby fixing the amplitude. The flow- of harmonics out-of-phase with the fundamental will introduce a reactive component in the plate impedance and cause the frequency to assume such a value as to give an equal and opposite phase shift in the tuned circuit. This phase shift will be obtained in the case of a sharply resonant circuit by a much smaller percentage change in frequency than is the case in a broadly tuned circuit.

The commercial data for the Type 337-B Low-Frequency Oscillator are the same as for the old design. Price, also, remains unchanged.

Continue reading here: C o m m unica t ion circui t s

Was this article helpful?