Letter from Gordon Hardman, WØRUN  

Vacuum Tubes:

Vacuum tubes have been the primary method of producing high power RF for almost a century. While semiconductors have made inroads in this arena, tubes still are dominant at the higher powers and higher frequencies. They are now manufactured at numerous facilities in the USA, Europe and Asia. With RFC introducing its own Alpha brand line of tubes, it is worth reviewing tube technology.

The origin of the vacuum tube dates back to the "Edison effect", where current flow was noticed between a heated wire and another unheated wire in an evacuated glass bulb. This was soon explained as being due to the motion of negatively charged electrons that had "boiled off" the heated wire. This lead to the development of the thermionic diode, which was patented in 1904, and soon thereafter to the De Forest "Audion", which was a triode. Triodes revolutionized radio receiver design over the next decade. The tetrode was invented by Walter Schottky in 1919. Fifty years after the first primitive tube, individual tubes were capable of RF power outputs in the hundreds of kilowatts. It is interesting to note that fifty years after the first primitive transistor, the largest RF transistors could produce roughly one kilowatt of power output per device. The thermionic vacuum tube was arguably the most important electrical technology development of the first half of the twentieth century.

A vacuum tube consists of a cathode capable of emitting electrodes when heated, an anode (or "plate") that attracts the electrons, and some means of controlling the flow of electrons. Tubes are classified as diodes, triodes, tetrodes, pentodes etc., according to whether they have two, three, four or five electrodes present. These electrodes are generally enclosed in a non-conducting space that is evacuated to produce the highest possible vacuum.  There are specialized tubes, such as travelling-wave tubes, klystrons, and magnetrons etc. that are used to produce very high powers at microwave frequencies. At the heart of almost every microwave oven is a magnetron, which is basically a diode and a very powerful magnet. The electrons interact with the anode voltage, the magnetic field, and the specially shaped anode to generate microwave power directly.

The tubes in the RFC lineup are conventional triodes and tetrodes. These can be categorized in a variety of ways, but the most important is the location, construction and function of the anode. Early tubes had all the electrodes contained inside a glass cylinder or bulb. The heat from the anode (which is where most of the waste heat is produced by normal tube operation) has to escape through the glass by radiation. Generally this occurs in the infrared portion of the spectrum, occasionally the anode will glow a dull red. Compared to conduction or convection, radiation is a relatively inefficient way to get rid of waste heat- the only way to increase heat transfer is to raise the anode temperature. There is a limit to this, of course eventually the anode material will soften or melt. Compare this with air, where the option exists to blow more air at higher pressure to carry away more heat. The most significant development in very high power tubes was to move the anode to the outside of the tube, so it is directly exposed. This is the so-called external-anode power tube, where the anode acts simultaneously as an electrode, pressure vessel and conductor of waste heat. The waste heat can be carried away using forced air through fins, a cooling liquid such as water, or even by direct conduction into a heat sink. This allowed anode dissipations to move from hundreds of watts into the hundreds of kilowatts region. On the other hand, radiation cooling can be totally silent, whereas forced-air cooling results in some amount of acoustic noise due to the fan and air movement.

The cathode also underwent improvement. Early cathodes were made of the metal Tungsten. While not a particularly good source of electrons, it can be operated at such high temperatures that it is a useful cathode material. The temperature at which emission occurs can be lowered by the process of "thoriation". Around 1.5% of thorium dioxide is added to the tungsten. By proper processing during evacuation of the tube the metallic thorium is brought to the surface of the cathode and emission increases about 1,000 times at a lower temperature than pure tungsten.

Even better performance can be obtained at a lower temperature using a so-called oxide emitter. A typical oxide cathode consists of a coating of barium and strontium oxides on a metal such as nickel. After "bake out" the cathode is said to be activated and will emit electrons copiously at a much lower temperature than even the thoriated tungsten cathode. A lower temperature means less power required to heat the cathode. Getting more emission per watt of heating power is one of the principal advantages of the oxide cathode. The heat is produced by a separate heater element, rather than by directly passing current through the emitting material, as in the case of tungsten cathodes.

Figure 1. Electron emission as a function of temperature for various cathode types.

Figure 1 is from the classic book "Radio Engineering" by F. Terman and shows the dramatic difference in emission with temperature for the different cathode types. Normal operating temperature for an oxide cathode is around 1,150K; for a thoriated tungsten cathode around 1,900K and for a pure tungsten one around 2,500K.

Comparing the two cathode types, there are some other differences. Apart from lower heating power requirements, oxide cathodes have higher current capability for short pulses (10 microseconds to 1 millisecond or so). For a thoriated tungsten cathode, peak and average emission are about the same. However, the oxide cathode is subject to degradation by bombardment of ions. Since there is always some small amount of residual gas in a tube, there are ions that flow in the opposite direction to electrons, and so strike the cathode. For very high voltages, this can lead to permanent damage of the oxide and loss of emission. Thoriated tungsten is not subject to the same kind of damage, so for higher voltage tubes thoriated tungsten is universally used. Another difference is that the life of a thoriated tungsten tube is a very strong function of the temperature of the cathode. It is recommended that the heater voltage be controllable, and that it be set to just above the lowest value that gives adequate emission, after the tube has burned in. Oxide cathodes are not nearly so critical in regards to the correct temperature, and so have greater latitude in terms of operating voltage. Finally, it is worth noting that oxide cathode tubes can be mounted in any orientation. Directly heated tungsten cathodes have a tendency to sag when hot, and so the tube must be mounted vertically, either up or down.

The Alpha lineup of tubes consists of all types- glass and ceramic, and directly and indirectly heated types.

Another important technology innovation was the switch from glass to ceramic for certain insulating parts of tubes. Apart from being less likely to crack, it is also possible to construct metal-to-ceramic connections which are vacuum tight and very strong. Leakage of gas past metal-glass seals is a factor in the lifetime of glass tubes.

Figure 2: Directly heated glass tube

Figure 3: External anode tube with an oxide cathode

One of the issues sometimes brought up is the lifetime of tubes compared to solid state devices. Life expectancy for a tube depends on a great many factors. In general, operation below the maximum ratings will increase the tube life. This is especially true for anode dissipation. The same, however, is true for solid state devices- the higher the junction temperature, the less the mean-time-to-failure (MTTF). Most power semiconductor manufacturers publish a graph so the designer can estimate the time to failure. Admittedly, typical MTTFs can be of the order of a hundred years, but at RFC we regularly see oxide cathode tubes that have been in amateur service for 25 years and look like they would probably meet the original manufacturers specifications. And this was in equipment that does not have all the protective features of modern computer-controlled amplifiers. So, while it is true that solid state devices are increasing in ruggedness all the time, it is still hard to beat the ability of a tube to withstand the vagaries of the real world. The occasional nearby lightning strike or power surge is unlikely to permanently damage a tube!