US2168052A

US2168052A - Stabilized oscillating circuit

Info

Publication number: US2168052A
Application number: US146640A
Authority: US
Inventors: Richard L Snyder
Original assignee: Farnsworth Television and Radio Corp
Current assignee: Farnsworth Television and Radio Corp
Priority date: 1937-06-05
Filing date: 1937-06-05
Publication date: 1939-08-01
Anticipated expiration: 1956-08-01

Description

A g- 1, 1939- R. L. SNYDER 2,168,052

STABILIZED OSCILLATING CIRCUIT Filed June 5, 1957 u E XC/ TAT/0N OUTPUT F' ig.3.

INVENTORT R/CHA RD L SN YDER.

A TTORXF IF.

Patented Aug. 1, 1939 STABILIZED OSCILLATING CIRCUIT Richard L. Snyder, Glassboro, N. J., assignor, by mesne assignments, to Farnsworth Television & Radio Corporation, Dover,

of Delaware Del., a corporation Application we 5, 1931, Serial No. 146,640

5 Claims.

This invention relates to oscillators wherein frequency of oscillation is controlled by a piezoelectric crystal, and particularly to oscillators wherein the active element is an electronic discharge device of the multipactor type, i. e., a device in which the electrons which comprise the discharge are liberated by the phenomenon of secondary electronic emission and the potentials of impact causing such emission are supplied by the oscillation of the device.

It is now well known that multipactors or electron multipliers can be made to generate selfsustaining oscillations, methods of causingsuch oscillations in various modes having been described in the patents of Philo T. Farnsworth,

No. 2,071,516 granted February 23, 1937, and No. 2,128,580 granted August 30, 1938, among others. Circuits utilized in causing these tubes to generate such oscillations are, in general, of great simplicity, usually involving merely a parallel resonant circuit connecting two of the elements of the tube, together with a source of direct current and potential for supplying the primary energy. In general, within reasonable limits, ease of starting oscillation and stability of oscillation when started in tubes of this character increases with the size of the tube, with the voltage applied to the tube, and with the current drawn. These factors make crystal control of such oscil-- lators a problem, since the voltage applied to an oscillating crystal and the current carried thereby must be limited in order to prevent overheating of the crystal or its destruction by the mechanical forces set up by its oscillation. This has led to the custom, in crystal controlled oscillators of the audion type, of using the crystal in the grid circuit, where it need carry but a small proportion of the actual oscillating power delivered by the tube. This expedient is not, of course, possible with an oscillator'of the multipactor type, wherein a single circuit connecting the tube element must carry all of the power.

It is, accordingly, an object of this invention to provide a crystal stabilized circuit wherein the current carried by and potentials imposed upon the crystal are within the safe working limits thereof; to provide a circuit wherein an attempted overloading of the crystal will cause it automatically to drop the load attempted to be imposed upon it; to provide a crystal stabilized oscillating circuits wherein departures from optimum adjustment of the circuit constants will have little or no effect upon the oscillation frequency; to provide a crystal controlled multipactor oscillator which will deliver relatively large loads; and to provide a crystal controlled multipactor oscillator which is easy to start, has an exceptionally high degree of frequency stability, and is simple and positive in operation.

Other objects of my invention will be apparent d or will be specifically pointed out in the description forming a part of this specification, but I do not limit myself to the embodiment of the invention herein described, as various forms may be adopted within the scope of the claims.

Referring to the drawing:

Figurel is a circuit diagram of a crystal controlled multipactor oscillator embodying my invention.

Figure 2 is a drawing of a multipactor of a 15 type preferred for operation in the circuit of my invention, showing the method of capacitance coupling through the multipactor tube itself.

Figure 3 is a diagram showing the equivalent electrical circuit of a piezo-electric crystal.

Considered broadly, the circuit employed in the oscillator of my invention comprises a piezoelectric crystal whose resonant frequency is slightly different from the desired frequency of oscillation, shunted by a D. C. by-pass whose 25 impedance at the desired frequency is relative y low; and which allows the circuit comprising the crystal and its holder and the by-pass itself to oscillate at parallel resonance at the desired frequency. Connected in series with these two elements I prefer to'use a tank or parallel resonant circuit which "will time through the desired frequency of oscillation. In the prefered form of my invention, wherein this circuit is used in conjunction with a multipactor, the circuit thus 35 'described is connected in series with the customary source of D. C. potential and current between the anode and cathode of the tube. It

is characteristic of the circuit connected in the manner described that impedance changes of the 40 circuit due to small frequency shifts (in the neighborhood of the desired frequency) cause relatively large changes in the effective capacitance of the tube which tend to oppose the frequency changes.

The output power from the device may be taken from a circuit inductively coupled with the tank coil, but I prefer to derive the output from a condenser plate coupled with the multipactor cathode, as by a band which surrounds the tube 50 envelope, permitting the coupling to be varied by sliding this band along the tube. Means are also included in the preferred form of "the circuit for supplying a shock excitation to initiate the self oscillation of the circuit.

Describing in detail the preferred form of oscillator illustrated in Figure 1, the multipactor tube l is preferably of the type shown in Figure 2, comprising a cylindrical cathode 2 which surrounds a grid-like anode 3. For the present purpose the cathode is preferably one having a high ratio of secondary emission to impacting primary electrons, such as those having a caesia'ted surface which will give secondary-to-primary ratios of as high as 11 or 12 to l. The anode is connected to ground through a millimeter A and amtential source 4. The millimeter is desirable for convenience in operation, and is preferably shunted by a by-pass condenser 5, and an addi tional by-pass condenser 6, connected from the anode direct to the ground, also helps in exclud ing radio-frequency currents from the D. C. instrument A.

The oscillating circuit for the tube is connected to the cathode, and comprises an inductor 1 and variable condenser 8 forming a parallel resonant circuit 9 in series with the stabilizing crystal circuit. The latter is also a parallel circuit, comprising the crystal I l with its conventional holder shunted by the inductor [2. This inductor is preferably in the form of a low loss coil whose inductance is such as to require a capacitance of a larger order of magnitude than that of the crystal holder and crystal to tune it to the desired frequency of operation. Radio-frequency millimeters B and C are shown in series with the crystal and inductor respectively, and an additional radio-frequency millimeter D is also shown connecting between the crystal circuit and ground to indicate the combined currents of the two branches of the circuit. These meters are convenient for determining the operating characteristics of the circuit, but are not necessary. A variable condenser l3 may also be shunted across the crystal, if desired, to facilitate starting.

Means are shown for shock exciting the circuit to initiate oscillation, comprising a vibrator type inductor coil of the Ford spark coiltype. This coil l5 isenergized from the battery l6 by pressing a push-button IT. The high tension end of the coil is connected to ground through a small inductor l8, coupled to the coil 1 of the tank circuit.

-As above stated, the output of the oscillator may be taken from a coil which is also coupled with the tank'coil, but I prefer to derive the energy from the circuit capacitatively, by means of a band 20 of metal or other conductive material which encircles the multipactor tube and may be slid therealong in order to change its capacity with respect to the cathode of the tube. The lead 2| connected to this band may be connected to any circuit which it is desired to excite by means of the oscillator. i

The mode of operaiton of this circuit will be more readily understood if we first consider the operation of a multipactor oscillator in which only the simple tuned circuit 9 is employed, the crystal control circuit being omitted. Such circuits have been shown in Farnsworth applications above-mentioned, and may be made to oscillate with the resonant circuit tuned either to a frequency whose period approximates the transit time of the electrons across the tube, or at a frequency whose period is quite long in comparison with the transit time. In either case, the oscillation occurs at approximately the frequency of resonance of the tuned circuit as modified by the effective capacitance of the tube in parallel therewith. As this capacitance changes with the space charge conditions within the tube, the frequency varies with the supply voltage even in the short transit-time mode of oscillation, and frequency variations of as much as one per cent may be expected with changing voltages.

The potentials causing the electron impacts, which give rise to the secondary electron emission upon which the operation of the tube depends, are derived solely from the voltage drop through the external circuit. It will therefore be apparent that these potentials will be greatest, giving rise to the greatest secondary emission, when the impedance of the external circuit approaches a maximum. The addition of the crystal control circuit does not change this state of aflairs. With the crystal circuit included, the tube will oscillate at a frequency whereat the series impedance of the tank circuit 9, shunted by the crystal circuit, in series with the tube capacity, is a maximum.

The crystal in its holder ofiers an impedance which is equivalent to that of the circuit shown in Figure 3, wherein the motional capacitance of the crystal is represented by a very small condenser 22, the motional inductance by a very large inductance 23, and the motional resistance by a resistor 24, whose impedance is low in comparison with that of the inductance in the neighborhood of the resonant frequency, i. e., a crystal has a high Q. The capacitance of the crystal holder bridges the motional impedance of the crystal and this capacitance, plus that of the condenser l3 (where the latter is used) is represented by the condenser 25. The capacitance of the condenser 25 is usually at least one order of magnitude larger than the motional capacitance of the crystal as represented by the condenser 22. In actual'operation the condenser I3 is adjusted to minimum capacity, and eifectively merely increases slightly the capacitance of the holder. Hence it may be neglected, for the moment, in considering the operation of the circuit.

As viewed from the tank circuit, at frequencis below and considerably removed from the crystal resonant frequency, the crystal and its holder appear as a small capacitance in series with the capacitance of the tube, and changes in tube capacitance therefore have little effect upon the oscillating frequency under these conditions, increases in effective tube capacitance serving slightly to lower the oscillating frequency,

At frequencies higher than the crystal resonant frequency, the crystal itself appears as an inductance in series with the tube capacitance, and an increase in effective tube capacitance would, under these conditions, increase the inductive impedance shunted across the tank circuit and again tend to decrease the resonant frequency of the circuit as a whole and hence the frequency of oscillation. At frequencies closely adjacent the resonant frequency of the crystal, however, the reactance of the crystal changes very rapidly as the frequencies increase, becoming first that of an extremely large condenser, passing through zero and then becoming that of a very small but rapidly increasing inductance as the frequency rises. Since the impedance in series with the effective tube capacitance under this latter condition is low, changes escapee be points on the impedance curve where the effective motional inductance is in parallel resonance with the capacitance of the crystal holder; there will also be a point where the impedance of the crystal is effectively capacitive, and where this capacitance, in parallel with that of the holder, is in resonance with the inductance of the D. C., by-pass I2. The use of a resistive bypass in place of the inductive by-pass l2 simplifies the reactance curve, and it is sometimes desirable for this reason, but ordinarily the multiple parallel resonant points will not'lie sulficiently closely together to cause anomalous operation, even if the inductive by-pass be used. In practice I prefer to use a radio frequency choke having a natural period at about ten times the frequency of which it is designed to have the oscillator operate.

Under these conditions, the tube is put in operation by pressing the button H, with the condenser 8 backed off so that the frequency of oscillation will be well above the resonant frequency of the crystal. The circuit 9 is then tuned to decrease the frequency, and as the crystal resonance is approached the impedance of the crystal circuit changes very rapidly. This changes the current and space charge within the tube and thus changes its effective capacitance, increasing it and tending to decrease the frequency of oscillation more rapidly than the tuning of condenser 8 would account for. The circuit will tend to drop into crystal control very suddenly, at a frequency such that the crystal and its holder pass from one-half to one-third of the current carriedby the by-pass l2. The crystal and by-pass are nearly 180 out of phase so that the meter B reads approximately the difference between the readings of meters 0 and D. The ratio of the currents carried by the crystal and the by-pass may be adjusted by tuning condenser 8 very slightly, and it is usually advisable to back off the condenser to a slight degree after the crystal has assumed control of the circuit.

It is sometimes possible also to make the crystal assume control by approaching resonance from below, i. e., with the tank circuit originally tuned below crystal frequency. In this case, however, the increase in tube capacitance bucks the decrease in capacitance in the tank circuit, so that when the crystal tends to assume control the change in tube capacitance occurs very suddenly and carries the frequency of the circuit as a whole above that of the crystal. This tendency to jump may make tuning from below" too critical to employ.

It should be noted that the frequency upon which the crystal assumes control is neither the series resonant frequency of the crystal itself, nor the frequency at which the crystal and its holder and by-pass I! are in parallel resonance with the crystal. The operating point is actually between these two frequencies, where the impedance of the crystal control circuit changes most rapidly with frequency, and control appears to be exercised throughthe fact that this rapid change of frequency changes the effective tube capacitance in such manner as to hold the actual frequency of operation constant.

It will be clear that the constancy of operation will be greater as the Q" of the crystal cotrol circuit, including the crystal and the by-pass, becomes higher. In practice, however, where this circuit is made of too high a Q the range of crystal control (referred to by operators as the crystal slot) becomes so narrow that tuning the crystal control is unduly difilcult. Introducing resistance into the D. C., by-pass circuit broadens the crystal slot so that tuning for crystal control becomes very easy. Oscillators have been constructed, giving from ten to twenty-five watts output, wherein the D. C., by-pass was a non-inductive resistor of from fifty to one-hundred ohms. In such circuits tuning for crystal control is extremely easy, and the frequency stability is sufficiently good to meet the minimum government requirements. In general, however, it is preferred to use an inductive feed, and sacrifice a certain degree of ease in tuning for a more accurate frequency control.

Where the condenser I3 is used the starting operation is made somewhat simpler, although the general theory is much the same. Herethe condenser 8 may be tuned to the best operating point, where it remains fixed. Before starting the condenser I3 is set for a fairly large value of capacity, preferably to about the point where its impedance is equal in absolute value to the impedance of, the crystal at resonance, and the circuit is shock-excited as already described. After oscillation has started the condenser is set at minimum value, whereupon its effect on the circuit becomes negligible. Increase of its capacitance from its initial value will decrease the selectivity of the circuit, and where utmost selectivity is desired it may be omitted altogether, in spite of the additional ease of operation which it provides.

Both A-T-cut and X-cut crystals have been used in practice. The impedance of the crystal circuit is low in comparison with the total impedance of the tube circuit, and since the crystal is not operating at its series resonant frequency it carries a relatively small, proportion of the total current; if the frequency does approach the series resonance frequency of the crystal, so the current carried by it would normally rise to dangerous proportions, the change in tube capacity will cause the crystal to dump" its control and as a consequence no crystals have, in practice, been injured either by overheating or by mechanical damage due to excessive amplitude of oscillation, even though the crystal circuit as a whole (including by-pass) is directly in series with the tube and the tube is generating relatively high power.

The fact that the current may be divided between the crystal and the by-pass makes the arrangement described greatly to be preferred. It is, however, obvious that the by-pass may jump not only the crystal but the tank circuit as well; i. e., that the by-pass may take the form of a high impedance choke through which the tube is parallel-fed, instead of series fed, as is common practice with other types of vacuum tubes, without departing from the spirit of the invention.

I claim:

1. The combination with an oscillator comprising an electron multipler having a secondaryelectron emitting cathode and an anode and a source of exciting current, of a tuned circuit in series with said cathode, anode and source, and a control circuit comprising a piezo-electric crystal and a direct current by-pass connected in,

series with said multiplier and tuned circuit.

2. The combination with an oscillator comprising an electron multipler having a secondaryelectron emitting cathode and an anode and a source of exciting current, of a tuned circuit in series with said cathode, anode and source, and

a control circuit comprising a piezo-electric crystal and an inductor in parallel therewith connected in series with said multiplier and said tuned circuit.

3. The combination with an oscillator comprising an electron multiplier having a secondaryelectron emitting cathode and an anode and a source of exciting current, of a tuned circuit in series with said cathode, anode and source, and a control circuit comprising a piezo-electric crystal and an'inductor having a natural resonant frequencyfsubstantially one numerical order of 'magnitude greater than the: resonant frequency of said crystal connected in parallel with said crystal and in series with said tuned circuit and multiplier.

4. The combination with an oscillator comprising an electron multiplier having a secondary- .electron emitting cathode and an anode and a source of exciting current, of a tuned circuit in series with said cathode. anode and source, a resonant frequency controlling element having a series-resonant frequency slightly oil of the desired frequency of oscillation connected in series with said tuned circuit and said multiplier and a ries with said tuned circuit and said multiplier,

and a direct current by-pass connected around said frequency controlling element, said element and by-pass having an impedance at the desired frequency of oscillation which is low in comparison with the impedance of the timed circuit at said desired frequency.

RICHARD L. SNYDER.