If one gives more than the most casual thought to the problem of how and why vacuum tubes fail, it becomes apparent that the word "failure" can mean more than one thing, depending upon who uses it. A failure may be a tube that fails to meet the requirements of some arbitrary test, as for instance, a tube checker's "good-bad" scale. Or it may be one which fails to meet some numerical value for a characteristic such as transconductance or plate current in a prescribed test. It may be a tube which, when used in a certain piece of equipment, reduces the performance of that equipment to below some arbitrary level. Finally, it may be a tube which simply fails to function at all, as for instance, one in which the heater doesn't even light.
It should be thoroughly understood that these situations are not merely degrees of the same simple phenomenon.
A tube that fails to meet any one of several tests may still perform with complete satisfaction in most equipment. A tube which meets all normal tests may fail to work satisfactorily in some equipment. Most tubes, new or old, contain some measurable defects; yet many tubes which are rejected as unusable will have no measurable defects. These are some of the confusing facts which must be dealt with in any serious study of the causes and corrections for the majority of tube failures. This is why there is no universal test that can be used to define a "good" or a "bad" tube, a subject that is discussed more fully in later sections.
In order not to confuse the reader, the term "failure," as used throughout this discussion, will be limited to those types of failures which, by their very nature, can be assumed to result in equipment failure or serious mal functioning most of the time. Where anything else is meant, it will be clearly stated and explained.
For purposes of analysis, it is convenient to divide failures into three broad categories. These are somewhat arbitrary and, without a doubt, overlap each other extensively. Nevertheless, they have become fairly standard among those who have made a study of the problem, and as a consequence, will be used here with some slight modifications.
Catastrophic failures are those which occur without warning and cause the equipment to become inoperable almost immediately. By far the vast majority of such failures occur within the first hundred hours of operation. They reach a very low incidence after 500 to 1000 hours and, with proper operation, become fixed at an extremely low level after 5000 hours of use. This normal curve or rate of failure (Fig. 1-1) can be interfered with by several factors which will be discussed. It should be carefully noted, therefore, that whenever the rate of catastrophic failures does not follow this normal curve, the secondary factors are usually to blame, rather than the tubes.
A second important inference can be drawn from this curve: As far as catastrophic failures are concerned, the longer a tube is in use, the better are its chances of continuing to remain in use. This will be referred to many times in this discussion, for there are many interrelated reasons why this apparent contradiction is true.
GLASS FAILURES
A not uncommon cause of the catastrophic failure is cracked glass. Although known to man since ancient times, glass is seldom recognized for what it really is.
Glass cold-flows; in so doing, it develops strains which eventually exceed its elastic limit until it breaks. If a sheet of glass is clamped on a bench so half of it extends over the edge and a weight is placed on the unsupported portion, a strain will be set up which will eventually cause the glass to break off. It may take weeks-or even months but it is only a question of time.
Fig. 1-1. Most catastrophic failures occur early in the life of a tube.
This curve shows the longer a tube is in service, the better its chances are against catastrophic failure.
A very common cause of glass breakage in vacuum tubes is bent pins. When they are inserted into sockets that tend to force them into alignment, tension is developed where the pin enters the glass. This tension in creases when the tube is heated and cooled. Depending on the severity of the strain, this force will eventually cause the glass button to break. Using a pin straightener before inserting the tube in the socket will help minimize the chances of this happening. Not disturbing tubes that are functioning properly will also help reduce the incidence of this type of failure, because it is a proven fact that merely removing and inserting the same tube in the same socket too many times will eventually lead to broken pins and cracked glass.
Incorrectly designed tube sockets can cause unnecessary glass breakage. The sockets are supposed to have floating contacts or inserts. These inserts may fit the pins snugly, but they must be free to move and align them selves with the pin. Sometimes an otherwise well-designed socket will be rendered dangerous to tube life by incorrect soldering techniques. Socket inserts should never be bent over and soldered directly to the chassis so there is no allowance for free movement of the insert. Stiff wires that restrict the movement of socket inserts can do the same thing. Naturally, allowing solder to run down the sides of the socket terminals, until it puddles at the bottom and "freezes" the insert, is to be avoided for the same reason.
Another source of trouble, as far as glass is concerned, are certain types of tube clamps that press upon the dome or sides of a tube and cause differential cooling. Glass has very low heat conductivity. If a metal ring is allowed to contact the envelope of a tube in some way, the sharp heat gradient which will develop at the point of contact, between the cold metal and the hot glass, will almost certainly result in a strain that will crack the glass sooner or later. In fact, a favorite method of removing the neck of a picture tube is simply to reverse this situation and wrap a hot wire around the cold glass. It will break off in a jiffy and leave a neat, clean edge.
Pin holes, or "star cracks" as they are called, are some times encountered in tubes located near very high-voltage sources, as in the high-voltage compartment of some TV sets. These cracks result from high-velocity electrons being beamed toward very small areas of the glass, resulting in local heating and consequent glass strain. Often the cause will be found in a high-voltage lead which passes too close to the affected bulb. Or these cracks may be caused by external arcing to the tube, as when inadequate corona shielding is employed.
Another little-known fact about glass is that its resistance changes with temperature. Where a very high voltage and a high temperature are encountered simultaneously, there is danger of sufficient leakage current developing to cause electrolysis. This is a chemical de composition of the glass which leads to surface damage similar to scratches, and this in turn leads to cracked glass. Plate caps that run very hot, or that carry high voltage or large amounts of RF current, should be cooled by using radiator caps with fins. A marked reduction in glass failure will result within the usable life span of the tubes if this precaution is taken.
HEATER FAILURES
Whenever one speaks of tube failures, the most likely cause that comes to mind is open heaters. Like cracked glass, this type of defect is both obvious and conclusive.
Even the nontechnical person can understand and frequently identify this type of defect. But just because it is so common, it is often taken for granted and accepted as a cause when it is far more apt to be a result or a symptom.
There are three methods by which a heater can become open. One is the result of long life and having been cycled "on" and "off" many thousands of times. This is characterized by a stretching of the tungsten heater wire which leads to eventual fracturing. A second means of producing an open heater is to run it excessively hot and literally vaporize the metal. When the wire finally parts, the ends are fused and balled, making this type of failure easily identifiable. Finally, a weld which is not properly made can open up. This, too, leaves a very characteristic tell-tale story. Tungsten wire does not weld to nickel, so it must be imbedded in the support wires. If this is done properly, it will break before it will pull out. But if it is done in correctly, it will come away intact.
Thus, in analyzing tubes which have failed because of open heaters, one can assign the cause of failure with a high degree of accuracy. Studies of tens of thousands of tubes returned for this cause show clearly that far more than half of them were "burnouts." The most common cause of heater burnout is excessive heater voltage. Tubes are designed to operate at a particular heater voltage. This is based upon many compromises in designing the tube. All of these compromises reach an optimum value when the tube is operated at precisely the voltage for which it is rated. When this voltage is exceeded by even a small percentage, the rate at which these compromises begin to deteriorate is very rapid. This fact is illustrated in Fig. 1-2. Equipments operated at ten percent over the rated line voltage have been shown to experience more than a one hundred per cent increase in breakdowns over a given period of time.
Tubes, when operated at ten percent above their rated heater voltage, will suffer up to a fifty percent decrease in heater life. A voltage under the rated value on the other hand, while having some disadvantages as far as certain other characteristics go, tends to increase heater life substantially.
Fig. 1-2. Curves showing how rate of tube failure rises as heater voltage
increases.
Whenever a heater is turned on from a cold start, there is a very high initial surge of current. This is because the cold resistance of the heater is many times less than its hot resistance. The surge current has a tendency to cause the heater to convulse and stretch, and it is this effect which leads to many premature heater failures.
Limiting Surge Current
Various means have been used to reduce this high initial surge current. Many recent TV sets have incorporated a surge-limiting resistor having a negative temperature coefficient. These resistors offer their maximum resistance when cold, and gradually decrease in resistance as they become hotter. In this way, they counteract the opposite characteristics of the heater. The use of these resistors is strongly recommended by all tube manufacturers. As much as a four-to-one decrease in heater burnout rate can be expected when surge-limiting resistors are used. Other advantages can be obtained from their use, but these will be discussed in a later section.
In the case of very large equipment using hundreds or even thousands of tubes, it has been proven economical, in terms of reduced maintenance costs, to allow heaters to be left on at all times, even when not in use. There are certain precautions that should be observed when resorting to this practice, because what one gains in increased heater life may be lost in other defects. (These are de scribed in the next section, under interface resistance.) However, it can safely be said that reducing heater volt age to one-half its normal value during long periods of standby will preserve heater life considerably, and will not cause any additional problems, other than the very brief delay required to bring them up to operating temperatures again.
Many circuits have several tube heaters connected in series. In quite a number of these circuits, the heater mix is quite drastic; that is, tubes having relatively low wattage heaters are in series with other tubes having relatively high-wattage heaters. Resistors are often shunted across one or more heaters to equalize the currents, as when 300-milliampere tubes are in series with 450- or 600 milliampere tubes.
Such combinations usually result in heater warm-up imbalance. The lower-wattage heaters will heat up most rapidly because their mass is frequently lower than the higher-wattage types and, hence, their thermal lag is less.
During this warm-up period, the heater voltages are not distributed uniformly; thus, the wattages may temporarily become very far out of balance. This is another form of the current surge problem, brought on by dissimilar warm-up times in a complex series-parallel heater arrangement.
Some tubes have the same warm-up time, regardless of their individual wattages. When these tubes are used in a series arrangement, there will be no surge current imbalance, because the resistance change with tempera ture in all tubes is the same. Thus, no one tube will light up brightly while all the others remain cold.
Filamentary Tubes
Tubes of the filamentary types, such as rectifiers and most battery or 1.4-volt types, have certain problems unique to themselves. Among these is the one called "filament sag." When heated, metal expands. Filaments are usually designed with provisions to compensate for this characteristic; however, certain precautions are required of the user in order to make these provisions effective.
Power rectifiers may develop shorts between filament and plate if the tube is operated for long periods of time in the wrong mounting position. Manufacturers specify that the tube, when mounted horizontally, shall be so oriented that the longer axis of the plate cylinder is vertical. This provides the filament with the maximum latitude for sag. Not all brands of identical tube types have the mount structure oriented in exactly the same position; consequently, some tubes could be oriented in such a way that even a small amount of sag will cause an arc or a short. Remounting a socket to favor the exact filament plane in the most commonly used tubes will result in a significant reduction of rectifier shorts and burnouts.
An interesting problem arises when equipment using vertically-mounted rectifier tubes is tipped up for bench servicing. Many rectifier tube failures occur at just such a time, especially if they have seen considerable service and have developed a fair amount of sag already. There is no simple solution to this problem except to recognize that it exists and to refrain from leaving sets running in such a position for long periods of time, as for instance, when waiting out intermittents.
Battery-type filamentary tubes are usually designed with a spring hook which takes up the filament slack.
In order for this spring to be most effective, it is best to operate these tubes in a vertical position, even though most tube manuals say they can be mounted in any position.
Fig. 1-3. Heater-to-cathode voltage versus life. Failure rate is mini
mum when heater-to-cathode voltage is kept within the 90-volt limit.
There are many heater failures that appear to be simply that and nothing more, but which in reality are the end result of heater insulation breakdown. This is particularly true where a relatively high heater-to-cathode voltage exists. Until a few years ago, all indirectly-heated cathode-type tubes had a heater-to-cathode rating of approximately 90 volts. This was based on the known limitations of the insulating material used on heaters. There has been no change in this material over the years, yet heater-to cathode voltage ratings have steadily increased. This doesn't make much sense-from an engineering point of view--even though it does permit the design of many circuits not possible under the older, more conservative ratings. As a result, we have far more breakdowns of this type in the field today than we had some years ago.
Life tests on a large number of tubes have shown a very definite relationship between the heater-to-cathode volt age and the length of time before a failure occurs (Fig. 1-3). For maximum reliability, this voltage should be kept below the 90 volts (between 50 and 90 volts being optimum) originally thought to be a safe maximum. Where excessively high heater-to-cathode voltages are encountered, reduced heater temperatures will help to minimize the damaging effect. In other words, high heater voltages and high heater-to-cathode voltages produce effects which are cumulative; when encountered simultaneously, heater life will be drastically foreshortened.
Some tube types are rated to withstand extremely high pulse voltages between heater and cathode. These tubes are made with a special ceramic insert between the inside of the cathode sleeve and the filament. Their voltage breakdown point can be raised to a considerably higher level than can be safely used for those types which use only the heater coating itself as the insulation. However, due to the large inert mass which this sleeve adds, these tubes are slow in reaching a suitable operating tempera ture at the cathode surf ace-leading into another problem, namely, that of arcing.
ARCING
Arcing in high-vacuum tubes can be caused by several things-among them lint, gas, and the drawing of large amounts of cathode current before the cathode tempera ture has reached its optimum working level.
Tubes which arc destructively are usually very gassy.
Arcing produces gassing, so if we examine some of the more common causes of arcing, we should find some methods for reducing the premature gassing of tubes. A very common cause of arcing is lint or other loose particles within the envelope. These become dislodged due to vibration or handling, and then fall through the tube structure where they become vaporized by the electron stream. This produces a localized gas cloud which, in effect momentarily reduces the spacing between elements separated by relatively high potentials. The electrostatic stresses plus additional electron collisions cause this gas to ionize, pro viding a lowered resistance path between the elements.
A heavy current follows; often, it is sufficiently strong to physically damage the grids of the cathode, and a metallic short develops. Gently tapping new tubes, base down, be fore they are installed will do much to insure that any loose particles within the envelope will drop to the bottom where they will do no harm. Refraining from removing satisfactorily operating tubes for the purpose of routine testing will also eliminate the possibility of stirring up loose particles that may cause arcing later.
A very common cause of destructive arcing is drawing maximum emission current before the cathode has reached its normal operating temperature. This is a frequent cause of damper and rectifier tube failure. What happens is this. An indirectly-heated cathode-type tube does not draw its plate current directly from its cathode, but from a reservoir of electrons in the space around the cathode. This reservoir (space charge) is supposed to stay ahead of the plate current demands. As long as it does, the plate current is drawn from the space charge. The cathode supplies the space charge with new electrons to replenish those drawn away.
The ability of the cathode to supply electrons, and hence to replenish the space charge, depends on temperature.
When the cathode is warming up from a cold start, its ability to supply electrons is fairly limited. This is what is known as its region of temperature-limited emission.
If very heavy plate current demands are made on a tube while it is temperature limited, the reservoir of electrons will be completely swept away, leaving the cathode ex posed to bombardment by the heavy negative gas ions that are always present. These negative gas ions are repulsed by the negative space charge when it exists, but when it has been drawn away, there is nothing to stop the ions from plunging violently into the cathode coating where they erupt the surface like miniature volcanoes.
The erupted cathode coating becomes vaporized in the electron stream, and more gas ions are formed. In far less time than it takes to describe, a gas arc has built up and serious pitting or stripping of the cathode has occurred. Quite frequently, the arc is sufficiently hot to burn a hole through the cathode sleeve and extend on into one of the heater folds where the arc current then finds a path to ground, opening the heater in the process.
The cure for this situation is a fundamental one; it will be referred to in greater detail, when specific types are discussed in later sections. In essence, all cathodes must reach their optimum operating temperature at the same time, or if this is impractical, to see that those tubes required to supply the current reach their optimum operating temperature before those tubes which draw upon the supply. In other words, be sure the load is not applied until after the rectifiers have reached their full operating temperature.
Rectifier tubes become the victims of another form of destruction which is frequently misunderstood. It has to do with the effects of peak and average currents and their relationship to such things as filter capacitors and their deterioration characteristics. The average current passed by a rectifier tube is the DC current drawn by the load.
Rarely does this current exceed the tube ratings, unless there is a short within the set. If so, rectifier arcing will very often occur, followed by cathode stripping or burn out.
It is beyond the scope of this discussion to go into the causes affecting breakdown in other areas of the equipment, except to point out that they are frequently the direct cause of tube failure. In the case of filter capacitors, there is that direct relationship, so they will be discussed in terms of how they affect rectifier tube life.
The amount of capacitance used in the input section of a filter has a very great effect on the peak current drawn from the rectifier tube cathode. This is the intermittent current which flows at the peak of each rectification cycle, replacing the current lost by the capacitor as a result of load and leakage currents. As filters age, their internal leakage current rises; if unused for long periods of time, their leakage current increases. When equipment with leaky electrolytics is turned on, the rectifier peak currents may exceed the tube ratings by quite a margin, and it is at these times that rectifiers are most apt to arc and burn out. This is why so many TV sets, for example, develop rectifier troubles immediately following a period of long disuse, as for instance, when the owner returns from vacation. When this happens, it would be in the customer's best interests if the service technician were to replace the leaky electrolytics with new ones, thereby preventing further tube failures the next time the set is left unused for a few days.
Simple circuit changes which will not interfere with the normal functions can be incorporated into the equipment to help minimize excessive peak currents that might otherwise destroy tubes. The principle requirement is that sufficient resistance be added in series with the rectifier plate to limit the peak current to a safe value. One place to add such resistance is in series with the center tap of the high-voltage transformer. Where no transformer is used, the resistance can be added directly in series with each plate.
FIXED VERSUS BIAS
While in this area of failures caused by other circuit breakdowns, it may be just as well if we touch upon two other prime sources of tube failures and their appropriate cures. Various types of equipment make use of some form of fixed bias, either for reasons of economy or because of some supposed improvement in performance. Whatever the reasons may be, this form of operation is regarded as very hazardous and unreliable by most tube people. The reasons have been copiously stated before, but they apparently require restating periodically.
Fixed-bias operation magnifies the inevitable minor differences between vacuum tubes. It provides no built-in margin of safety in the event of circuit or tube malfunctions. When trouble does develop in a fixed-bias circuit, due to a deficiency in an associated component or the tube itself, the results are usually quite destructive. The alter native - cathode bias - provides increased tolerance of characteristic variables and an automatic degenerative action which inhibits runaway conditions, thus helping to reduce destructive breakdowns.
One should not assume that fixed-bias operation is limited to only those applications that come under the heading of high-power equipment. It is becoming common in many of the latest hi-fi and stereo amplifiers, and several TV sets incorporate this type of circuit in their AGC circuits. These will be dealt with specifically in the section devoted to UHF and VHF tubes and their special problems.
Another form of fixed-potential operation which is a frequent cause of tube failures is fixed-screen operation.
Not only does this also tend to magnify small variations in tube characteristics, but it likewise provides no safety factor in the event of other component malfunctions. For example, if plate voltage is removed from a tube operated from a fixed-screen supply, the screen current will rise to the point where the screen will probably run red hot.
This will cause destructive gassing of the tube, vaporization of protective materials from the screen, and per haps warping and shorting of the elements. The use of screen-grid dropping resistors provides current degeneration with its many benefits, including greater tolerance to tube variables, built-in runaway protection, plus a variable gain characteristic that will lead to longer, more satisfactory tube life.
In the foregoing paragraphs, the many causes for sudden and disastrous tube failures have been pointed out. It has been shown that the great majority of them can be minimized-if not eliminated entirely--by thoughtful design considerations, or by simple maintenance and modification procedures. The incorporation of these principles into existing equipment has been demonstrated to have reduced catastrophic failures by several hundred percent. This means that far more tubes lived long enough to wear out.
The subject of why tubes do wear out eventually, and how their time of usefulness can be lengthened, is the subject of the next section.