October 1944 Radio-Craft
Wax nostalgic about and learn from the history of early electronics.
See articles from Radio-Craft,
published 1929 - 1953. All copyrights are hereby acknowledged.
So much time has passed since an average home garage mechanic could service his car or truck with standard tools - combination wrenches, screwdrivers, socket sets, timing light, and a multimeter - that asking "remember when?" is passé. That era pretty much ended in the late 1980s as computerized cars were becoming the industry norm. A good percentage of people nowadays have never and will never service their own vehicles. In the mid 1940s, the electronics world was lamenting a similar situation with diminishing ability to build and modify electronic components like coils and stacked plate capacitors because of the increasingly higher frequencies being used in communications (way up into the UHF band!). This article introduces the klystron tube, having been around for less than a decade at the time, as being one of the culprits that was enabling the disturbing trend ;-) . A layman's introduction to the physics behind its operation is provided. Those guys thought they had it bad, but knew nothing of the reduction of serviceability of electronics circuits and components compared with our present day multi-layered printed circuit boards stuffed to the gills with condiment-size surface mount components.
Klystron: Tube of the Future
The Klystron Will Fill Many Important Post-War U.H.F. Jobs
By Capt. Eugene E. Skinner Hq. A.A.F. Training Aids Division
Fig. 1 - Evolution of a resonant-cavity circuit from the coil-condenser combination may be traced with the help of this diagram.
Cut-away photo of a typical Klystron.
All Illustrations Courtesy Sperry Gyroscope Co.
The new fields of radio that have been, developed in recent years have so changed some radio components that it is now impossible for the amateur to wind a coil, change the number and size of the plates on a condenser, and follow the latest developments. Instead, as frequencies become higher, special, critical components are required. Examples of these are found in the resonant circuits and tubes.
For the ordinary range of frequencies with which radio technicians and amateurs were concerned a few years ago, a resonant circuit could be designed to fit any frequency or band of frequencies by changing the number of turns on a coil and the size of the condensers, or by changing crystals. With an increase in frequency, fewer turns in the coils and fewer condenser plates are used. As a very high range of frequencies is reached, a single small turn of wire connecting two condenser plates must be used and crystals become so thin as to be useless. Any further variation along conventional lines is very difficult. New methods must be adopted.
In addition to the problems offered by these lumped circuit constants, as inductance of turns of coils and capacitance of plates or condensers are called, the problems of the vacuum tubes themselves become evident. These problems, caused by characteristics which were not bothersome in ordinary tubes at lower frequencies, become of major importance in the ultra-high-frequencies. In these ranges, such small quantities of capacitance and inductance are needed that the capacitance between the electrodes of the tubes and the inductance of the loops formed by the electrical circuits through which the electrons flow in the tubes are of relatively large value. In addition to this, the electrodes are of such size distance apart, and relative position, that the time required for the electrons to travel through the tube becomes an appreciable part of a cycle. Finally a frequency is reached for each type of tube at which the starting and maintenance of steady oscillations becomes impossible. This feature is compensated for by the construction of smaller tubes of the "acorn" and other types, but the difficulties still exist above the ultra-high-frequencies, for eventually the point is reached at which the tubes are necessarily so small and possess such little power capacity as to be impractical.
One method of solving both these problems at the same time was devised by Russell and Sigurd Varian and W. W. Hansen at Stanford University in 1937, by the invention of "Klystron"* tubes. This type of tube embodies the principle of modulating the velocity of the electrons as they flow through it and which has at least a portion and possibly all of the oscillating circuit components included as integral parts of the tube.
The most important component of the Klystron is the resonant cavity. It is an outgrowth of the application of the lumped circuit constants in resonant circuits. When the point is reached at which a single small loop and a pair of condenser plates will no longer serve for the extremely high frequency desired, a method must be devised to apply the same principles in a different form. To decrease the inductance beyond that of a single loop, the principle that loops in parallel have less inductance than one loop is applied. Therefore, coil loops are placed between the same pair of condenser plates, until an infinite number of them have been added. See Fig. 1. The result is a doughnut-shaped ring slotted within the hole at the shortest diameter, with two plates placed parallel to the axis of the doughnut, one at each edge of the slot, so that it appears to a casual observer to be only a solid doughnut shaped object with a short plug in the center. If laid flat and cut a one would cut a pie, the cut edge would be like the outline of a dumbbell.
As the frequency is increased, this doughnut becomes smaller. In most Klystron tubes two of these resonant cavities are used, one as the "buncher" and one as the "catcher." It is obvious that an electron stream cannot pass through the solid plates 'in-the centers of these cavities. These plates are therefore replaced by meshes or grids. In actual applications, the cavities are not always tubular doughnut in shape but more often have distorted variations.
The action of the tube may be understood by referring to Fig. 2. A stream of electrons is beamed by a cathode through focusing electrodes and an accelerator grid toward the pair of bundler grids through which the stream must pass. The cathode and focusing electrodes have applied to them a high negative voltage with respect to the rest of the tube, which is grounded on the positive side. Voltage must necessarily be well regulated as fluctuations will affect the frequency of the tube. The buncher grids and their associated resonant cavity are excited by a radio frequency source in such a manner that as one is positively charged, the other is negatively charged as in coil-and-condenser action. These grids have between them, therefore, an electrostatic field which is parallel to the flow of electrons. The strength of this field is such that it appreciably changes the velocity of the electrons, but does not do so to the point that it will stop the flow completely at any time. As an electron comes into the field, assume that it enters such a portion of the cycle as to cause it to speed up. As the radio frequency excitation passes through the zero point of the cycle, the speed of the electron is not affected. Then, as the second half of the cycle is applied, or the charges are reversed, the electrostatic field opposes the flow of the electron, slowing it down. Now it can be seen that as the excitation passes through several cycles, those electrons which have been slowed down during a cycle are overtaken by those whose speeds were not affected, and both groups are overtaken by those electrons which were sped up, causing a bunching of the electrons, at a point past the "buncher" grids.
Fig. 2 - Schematic view of the Klystron.
Fig. 3 - How the tube is tuned. Coarse adjustments are made with the three large struts, fine tuning by varying one against the other two.
A second pair of grids, known as the "catcher" grids, is placed in such a position that the bunched electrons pass through them at the rate of one bunch per cycle. It is assumed that the first of the grids is in the negative half cycle of an oscillation. As the electrons approach the first of these two grids, which are part of a tuned circuit, the negative charges in the electrons induce a positive charge on the first grid in a condenser-like action, then pass on toward the second grid. The original negative charge on the first grid has a slowing-down effect on the electrons, and as the energy of motion cannot be lost by the electrons without being gained elsewhere, this loss of electrons induces energy in the tuned circuit.
The change of charge supports the oscillation in the tuned circuit, and by the time the bunch of electrons has reached the second catcher grid, the oscillations have changed the charge on it so that it is then negative. This negative charge, again in opposition to the charge of the electrons still further decreases the velocity of the electrons, and, as before, this loss of energy by the electrons induces energy in this tuned circuit. After the electrons have passed the catcher grids, they are picked up on a collector plate.
It can readily be seen that the energy considerations are very important. It has been shown how energy is taken from the catcher grids while none is supplied from sources other than the electron stream. In , the buncher, it is necessary to supply a source of energy to provide for the bunching, but as some of the electrons are speeded up, taking up energy from the buncher, so are some of them slowed down, giving up an almost equal amount of energy to the buncher. The overall result is that there is a very little additional energy needed. A short coaxial cable feedback from the catcher to the buncher provides the necessary source of radio frequency oscillations and energy. The excess energy from the catcher is the output of the tube and is also taken from the resonant cavity by means of a coaxial cable. Output and the feedback is accomplished by means of one-turn loops which are formed by the central conductor of the coaxial cables, which enter the hollow "doughnut," form a loop, and are grounded at the end. A part of the magnetic flux of the resonant cavities which exists within the hollow doughnut cavities, also passes through this loop, giving an inductive coupling which is the equivalent of an air-cored transformer.
The Klystron tubes are very versatile and have many applications. In addition to their electrical capabilities, they are reasonably rugged, lending great value under far-from-ideal combat conditions. As the frequencies increase, the tubes become smaller, the wave lengths being the controlling factor of practically all the physical dimensions. The general design is such as to provide satisfactorily for the necessary dissipation of heat.
Tuning the Klystrons may be accomplished by several different methods, depending upon their physical construction. If a tube is so constructed that the cavity is sealed within it, provision may be made whereby tuning can be accomplished by changing the spacing between the grids. Most Sperry Klystrons utilize a screw-type arrangement by which the spacing between the grids can be varied, made possible by having the grids mounted on a diaphragm which can be flexed by the screw control. See Fig. 3. This spacing is the most critical dimension as far as frequency is concerned.
If the cavity is outside the tube, tuning may be accomplished by screwing plugs into the cavity, effectively changing its volume. Several other methods of varying the dimensions of the resonant cavity or the flux inside it may be used, as well as variation of the applied voltage.
When no method of feedback is used with the Klystron tube, but a separate oscillation is used in the buncher, the tube is being applied in its simplest form as an amplifier. Since the beam of electrons gives off a great deal more energy than it absorbs, it is therefore a power amplifier (equivalent of a Class C type) and may also be used as a voltage amplifier for small radio-frequency voltages.
All that is necessary to convert such an amplifier into an oscillator is a method of feedback so as to sustain oscillations, and this is accomplished by means of the coaxial feedback coupling previously described. The Klystron is generally as efficient as any other type of ultra-high-frequency oscillator. While theory indicated that efficiencies of 58% are possible, in actual practice they are considerably less.
It is very improbable that electrons would be bunched in such a manner as to produce a pure sine wave. The wave produced is actually composed of a large number of harmonics. This being the case, it is only natural that the Klystron tube should find application as a frequency multiplier. In order to accomplish this, it is necessary that the relative phase of the two resonant cavities be such that bunches of electrons hit the catcher at a time which will cause the oscillations to build up, and the catcher cavity must be designed for the desired harmonic frequency.
Unfortunately a great many of the most important applications of the Klystron tubes cannot be discussed. In the meantime, more and more applications are becoming apparent daily in the fields opened by the discovery of this important electronic device.
One thing that the war, or rather, the peace which follows it, should be able to do for radio is to clean up and set in order the wavebands assigned to broadcasting. You remember the pre-war semi-chaos which conferences, plans and much hard work on the part of the U.I.R. (Radio International Union) had failed to straighten out? The end of the war will furnish a glorious opportunity of setting all this to rights and of forming a governing body for broadcasting in Europe as enlightened and as powerful as the Federal Communications Commission in the United States. One thing that most of us would like to see is a return to the original 10-kilocycle separation between stations - the 9-Kc. separation in operation when the war broke out was not sufficient to ensure decent quality or to prevent heterodynes.-(Wireless World, July 1944)
*Trade Mark Registered by Sperry Gyroscope Co.
Posted November 12, 2014