SELECTED AND PREMIUM TUBES [Getting the Most Out of Vacuum Tubes]

Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting

In the previous section, we mentioned some of the normal variations that exist in vacuum tubes and the natural laws that govern them. In the early days of the electronics industry, when radio and communications were the principle uses to which vacuum tubes were put, these variations seemed to be of little consequence. More recently, during World War II, electronics played a vital role in many theaters of activity. The tubes used in the tremendous variety of military equipment ran the full range of all their normal characteristic variables. There is no evidence to indicate that they were any the less satisfactory for this reason.

Following the war, and with the advent of commercial television, a new crop of circuit engineers seems to have been born. Many of them had received at least a part of their training during their military service or in one of the military-supported electronics industries. Because of this training, they appear to have acquired a new engineering philosophy, based on the "cost-plus" unrealities of a war-time economy. This philosophy had as its main tenet the assumption that a circuit engineer's primary function was to write involved specifications, minutely describing each component in his hand-tinkered creation, and then leave it up to some other engineer-probably in his supplier's plant-to solve the problem of how they were to be made. Since very few significant advancements in any science or art have come about as a result of an edict, these specification writers didn't produce any miracles either. What they did produce was a practice which has now become so universally adopted that familiarity and indifference have begun to give it an air of respectability. This is the practice of an equipment manufacturer writing his own procurement specification for tubes which, though they bear a standard type number designation, are materially different from the tubes sold under this same designation to the replacement market or, for that matter, to some other equipment manufacturer.

These are what are known as "selected tubes."


There are many characteristics for which tubes are selected. Some of these are fairly harmless, like the tuner manufacturer who insists that the range of input capacity in the tubes he purchases should be held within such limits that his production workers have only to turn the trimmer screws, in or out, one turn. This saves him production time, he says. This is a harmless selection because the tubes his specification reject could be used in his tuners for replacements, and no one could tell the difference after the trimmers were readjusted.

But there are other selections that involve characteristics not so whimsical, as for instance, the cutoff characteristics of some IF tubes. In order to make an AGC circuit function over an extreme range of signal strengths, tube selection is held to those which have a sharper than normal cutoff characteristic. These are used in one of the controlled stages. When a nonselected tube is used in this same socket, "sync buzz" is apt to be heard, or perhaps poor vertical hold may result due to clipping in the final IF stages. It is fairly common practice among some set manufacturers to use colored paints to identify those tubes which have been selected for use in critical circuits.

This paint is often applied to the tip or sometimes to the base of the tube. Very often, there is no indication at all that these tubes are in any way different from others which bear the same type designation.

Selection of tubes, in order to make a marginal design work long enough to get it out of the factory, is a common resort of some in-plant engineering groups. Another fairly common practice is to select tubes that will cover up for some other component deficiency, as for example, a flyback transformer which is improperly designed. Many special selections have been set up to eliminate the picture interference known as "snivets," which is caused by a combination of tube characteristics and transformer de sign. Tubes that fail to work with one design are bound to work with another, and those that work well in one set may not work in another. The most undetectable forms of tube selection are those which deal with unregistered characteristics. In a particular equipment design, it will be found that certain tubes fail to work as well as others.

After much investigation, some obscure characteristic will be isolated and the cause assigned to it. A specification will be written, requiring that tubes having that characteristic be removed from the shipments going to that particular manufacturer. However, since it is an unregistered characteristic and no one else measures it in his product, the probability is that no other manufacturer's tubes will work satisfactorily in that socket. If the characteristic were a registered one, someone would discover, sooner or later, that one manufacturer's tubes were uniquely different in one respect, and standardization would begin to take place. But there are many unregistered characteristics and variations that can go unnoticed for years.


There is one aspect about tube selection which is almost invariably true. It almost never produces a better product as far as the user is concerned. The same problem could have been solved in a number of other ways, all of which would have lead to the same amount of customer satisfaction. Where the customer gets hurt is when the selected tube finally gets into trouble and has to be replaced. The question of whether or not a standard tube having standard tolerances will replace it is often a very expensive one for the customer to resolve.

Selected tubes are generally more prone to failure than nonselected ones; not because they have been selected, but because the circuit in which they are used is generally abnormally critical. That's why it requires a selected tube in the first place! Its design is such that all the normal tolerances have been left out, and as soon as the selected tube's characteristics begin to drift, it is in trouble. In fact, herein lies the reason for a great deal of the mis-leading evidence which has been built up concerning the unreliable nature of tubes in general. When normal tubes are used to replace tubes which were originally selected to make a critical circuit work, the results are apt to be somewhat less than satisfactory.

Selecting tubes on a performance basis, so that they will operate at their best in certain common applications, is not always a harmful practice in itself, provided some indication that they are in some manner unique is given.

Occasionally, an equipment manufacturer will put a special part number on such a tube. This tells anyone who later is faced with the problem of replacing such a tube that it is a special one. It will then be realized that the equipment may not function as well if a replacement tube is used, or that the circuit may require additional adjustment to make the new tube work properly.

A special form of this situation exists when tubes are sold as matched pairs for use in hi-fi and other audio applications. These tubes have been selected from the normal distribution of all the tubes in the product and matched so that they have similar characteristics. Usually, plate current is the characteristic that is matched. When these tubes are used as a pair, because their plate currents are nearly identical, cancellation in the output transformer will minimize any tendency toward DC saturation of the core, with its consequent increase in distortion. Of course, if the amplifier is equipped with a balancing potentiometer in the cathode circuit of the output tubes, the same thing can be accomplished with unmatched tubes, but the pretested and prematched tubes make it easier and don't require the use of any instruments when making the substitution.


We have talked about various forms of special selection as it applies to standard, commercial grade tubes. How ever, there is another large class of selected tubes. This class includes types which are known as industrial and military tubes. These are frequently referred to as premium tubes because they sell at somewhat higher prices than their prototypes. They are also referred to as reliable tubes. This latter reference is the one we want to explore at this time.

Immediately following World War II, a great deal of interest was expressed by the military in the development of specially designed tubes that would have a higher degree of survival in military applications than the then existent commercial-grade receiving tubes. The story was told about an extensive search of military maintenance files, which disclosed that about the only thing wrong with electronics was that vacuum tubes were involved, and they were just plain unreliable! All sorts of statistics were brought forth, tending to prove that with the currently available, low-reliability tubes going into military equipment at that time, the new giant bombers which used several thousand tubes in their electronic brains simply wouldn't be able to get off the ground. Theoretically, at least, enough tubes would fail just during the ground checkout period to render the mission a failure. The fact that there were many airplanes in the air at that moment, apparently none the worse for this situation, didn't alter the argument. Tubes were highly unreliable devices and had to be improved. It was amid such an atmosphere that the first reliable tubes were born. They were the out growth of a crash program to bring out tubes which could be immediately substituted for existing types in existing sockets. Their prototypes were the popular 6AL5, 6J6, 6AK5, 12AU7, and similar types.

These early "reliable" types proved to be little, if any, more reliable than their prototypes. This was a disappointment to some who immediately launched into a renewed program of specification writing aimed at tighter or more reliable specifications for those same tube types.

As each new wave of specifications was produced, new numbers were appended to the originals until at last there was no more room to put additional numbers on the bulb!


The 6AL5 is a typical example of what took place during this era of reliable-tube specification writing. Beginning as the EIA type 6AL5, it was a simple dual diode designed for second detector and AVC service in pre- World War II days. It was perhaps one of the most reliable tubes in the applications for which it was intended.

But, it found its way into many unique, and originally undreamed of, applications when it went to war. It was used as a clipper, clamper, bias rectifier, and as a low power switch. It was used at frequencies ranging from the very lowest to over 500 megacycles. It became a very common tube type in many very uncommon usages. In short, it became a universal component, having many and varied requirements.

Well, its first metamorphosis came when it was drafted into military service as the JAN-6AL5. A specification was written, describing its average characteristics; these were, in general, taken directly from the EIA registration data for this type. Many thousands of these JAN-6AL5's were purchased by the military services or their contractors, and most of them went into finished equipment where they proved normally satisfactory. After the war, when a general campaign got under way to "ruggedize" all military tubes, certain shock and vibration tests were added to the old JAN specification, and the tube became known as the JAN-6AL5W. "Ruggedization" meant more or less what its name implied. Tubes were being used in more and more places where they were subject to rough handling as well as heavy vibration. Trucks and jeeps bouncing over rough terrain, as well as multi-engine aircraft, and even the newer jets were subjecting tubes to additional stresses which they had not been required to withstand previously. Some tubes underwent large structural changes in order to qualify for the new "W" suffix. The 6AL5 wasn't changed in any recognizable form because it was pretty rugged to begin with.

The ruggedization program soon became lost in the larger one called the new "reliable tube" program. So, new specifications were written, adding a few more controls and test points, and the 6AL5 became the JAN-5726.

This hadn't been around too long before some additional changes were made in the style of specification being written, and the JAN-5726 became the JAN-6097.

There were now four specifications, including the EIA registration for the original 6AL5, and they all described the same tube. This was obviously not in the best interests of anyone, so some attempts were made to simplify the whole situation by combining the salient features of all of these specifications into one. The first attempt at this produced the tube type JAN-5726/6AL5W. Apparently, someone overlooked something because that was followed by a later specification, summing up all the former specifications and including the previous summations. That became known as the JAN-5726/6AL5W /6097. It was at this point that they ran out of space to print any more numbers on the bulb, and someone facetiously suggested that the tube be called the 6AL5 again which, of course, it really was anyhow.

How is it possible for such a comedy of well-intentioned, misdirected efforts to take place in this age of space probes, atomic energy, and wonder drugs? In order to understand the reasons, it is necessary for us to draw back the curtain on another area of popular misinformation-an area which was very much less understood when all this was happening than it is today. This has to do with the very heart of the problem--the nature of reliability and the standards or specifications that aim to de fine it.


Standardization is one of the fundamental requirements of any complex society such as that in which we live. Without it, we would have absolute chaos in our dealings with other people, and our vast industrial economy would be unable to function. Because of standards, we can buy the various products of a host of manufacturers, all with absolute confidence that they will do what we want them to do. There is hardly any article we come across in our daily lives that is not made in accordance with some standard or another. Some of these, like the food and drug standards, are government-enforced. Others-like auto mobile tires, electric light bulbs, or thread-are carefully controlled by mutual agreement between the manufacturers themselves, for their benefit as well as that of their customers.

Standardization is important to the manufacturer be cause it allows him to reduce his costs by having only one kind of machinery, one type of process, one set of tools, or by having to train his people in fewer operations. It reduces inventories and obsolescence, and thus further reduces costs.

To the user, standardization means a better product at lower prices. Standardization encourages competition, and this works to the user's advantage as far as available choice is concerned as well as being a further stimulus toward better quality and lower prices. The user can also feel confident that when he buys a standard article, it will satisfy his needs most effectively. Why? Because of the very method by which standards are developed.


Most standards are arrived at by collective study and compromise. The end use for the product is first studied very thoroughly. If there is more than one use, they must all be studied in turn. Next, with the requirements of each end use well known, the product is then measured to see how well it fits these needs. The variations in product are matched to the variations in end use, and an over-all specification is arrived at to describe the best possible compromise between the two. It is axiomatic that the fewer the end uses are, the more confining and detailed the specification can be. This is the secret of a good standard--a specific application and a specific product to meet it.

Take a very simple example-the electric light bulb.

There are many kinds of electric light bulbs, such as the various sizes made for house lighting, industrial lighting, and decorative lighting. Then there are automobile lights of all sizes, flashlights, instrument lights, various signal lights and outdoor advertising lights. Yes, the list is a long one and it is apparent that there is far from a single standard that could be applied to all these lamps. There are voltage differences, socket variations, environmental ranges, and a lot of other factors that aren't applicable to each and every bulb.

This problem is met in a simple, straight-forward manner. No one specification is attempted for all electric light bulbs. Instead, there is an individual specification for each specific application. Thus, all house lamps are made to work on one voltage and to fit one standard screw socket.

There are differences in wattage and bulb color, but that's about all. With only those few variables to deal with, a standard is possible and practical. Similarly, there are standards for all other sizes, shapes, and end uses for electric light bulbs-all tailored to their specific end use.

Thus, in order to have a standard at all, you must first know the end purpose for which it is intended, and the end uses for the product must be more than similar. They must be specifically alike. When this can be said to be true, then standardization-the making of all products intended to fill that end need in an identical manner, or with as few variables as possible-is a direct way of producing improved reliability. In short, specialization, which is only possible if you have standardization, is a very effective means of approaching perfection-the essence of reliability.


Tubes are sometimes thought of as being first cousins to electric lights, probably because they both use filaments of hot wire in performing their functions. They are largely constructed of glass and must be evacuated in order to perform at all. It was because of this common association that we used the electric light bulb in our earlier example of the standardization of a very common article of manufacture. But this assumed relationship is extremely remote if one becomes familiar with the facts.

In spite of the commonness of the electric light bulb, it is far from a universal product. If it were, a few types would be found doing just about all the lighting jobs there are. But this is not the case. Each lighting task has its own special shapes, bases, voltage ratings, wattages, etc.

No one would think of taking a lamp intended for one type of service and forcing it to do some other kind of service.

Vacuum tubes, on the other hand, are one of the most universally-applied products around us. This may not be apparent to many because the differences are not usually visible. But there is just as great a difference between the requirements of a tube intended for blocking-oscillator applications and one intended to be used as a DC amplifier as there is between a lamp intended for intermittent flash light use and one intended for use as a radio tower warning light.

Many applications for vacuum tubes are actually such that they require characteristics not mutually attainable in the same tube. Attempting to enforce both of these characteristics in the same design, to have a more widely applicable tube, usually results in excessive cost as well as a compromise of both of the primary objectives namely, reliability and efficiency. In order to see more clearly how this comes about, let us take another simple example from something outside the electronics field, but one which is familiar to everyone.


Water is perhaps one of the most universally-used commodities on our planet. In fact, we regard it as so essential that we can't even imagine life without it. Suppose we examine some of its many uses or applications and see whether we could write a simple, universal test that would insure that some strange new liquid, just discovered on a new planet, could be used wherever water was needed.

Water is used by both man and animal as a beverage.

For this purpose, we require that it be tasteless, as well as germicidally pure. It would have to be chemically inert as far as our systems were concerned. As a matter of preference, we like our beverages to be either nearly ice cold or almost boiling hot. We also use water for laundering and bathing. For this purpose, we like it to have a low mineral content, or to be "soft," and we want it to be clear and colorless. For our internal combustion engines, which use water as a coolant, we merely concern ourselves with its boiling and freezing points-the higher on one end and the lower on the other, the better.

Water is used in fire fighting. Here, it must be non flammable and plentiful. There are many industrial uses for water that require it to be free of all chemicals, as well as of all organic matter. Agriculture uses water, and is concerned principally with nature's distribution system. Water is also used for transportation; here, its specific gravity is of primary importance.

There are many other ways in which water serves as a universal commodity, but these examples are enough to illustrate the point. How would you set up a universal specification for the applicability of a sample of water for all of these applications? Well, you could simply test it for everything. This would be so costly as to render it too expensive for several of its primary applications.

But, suppose you overlooked this fact and decided to test it for everything-just to make sure it would work in any application. This approach would def eat your primary purpose because several of the more severe requirements are incompatible and at cross purposes with some of its other application requirements. For example, chemically and organically pure water is not suitable for use as a beverage. It doesn't even taste like water!


We come now to the subject of reliability itself. Al though we have been using the word throughout this section, we are quite sure that it has many different meanings in the minds of most readers. The least realistic of you will have been thinking in terms of a tube which, when plugged into any equipment suited to its general characteristics, will perform better and longer than the original tube. More realistic readers may have in mind only a longer useful life under conditions some what more grueling than normal. Others may think in terms of no better over-all performance or life-simply a freedom from early, or sudden and unpredictable, failures.

In varying degrees and amounts, these are the wishes of all those who seek reliable tubes. Some want better performance from existing equipment to be brought about by a simple tube change. Others want improved performance in new designs built around super tubes. Yet all, for the most part, want the same broad characteristic as the prototype had. In other words, it must still taste like water.

There are several methods by which an attempt is made to specify reliability. The favorite one is to narrow the range of characteristic spreads in the belief that a more restricted range (higher uniformity) can, of itself, de scribe reliability. This is largely wishful thinking unless something else is also done to control the bogey shift.

In other words, all tubes may fall within narrowed limits, but who is to say that they won't all be just within either the upper or lower limit of the specification, as would be the case if they were selected to meet the specification.

If the shape of the distribution is described and controlled by statistical methods, the uniformity of a given characteristic may be well maintained. Whether this has any correlation with reliability depends more or less upon what definition of reliability you are using.

Environmental and accelerated life testing are other common methods frequently employed to attempt to measure reliability. The success or failure of this approach depends very much upon the degree of correlation which exists between the method chosen and the end use. For in stance, merely vibrating tubes does not insure that they will not succumb to some specific mode or period of vibration in the equipment for which they are intended, unless the test conditions are identical to the field conditions.


A specification aimed at insuring the reliability of all tubes which pass it is the ultimate objective of many procurement people. The fact that so many specifications have been written, all describing the same few tube types in so many different ways, would tend to indicate that most of these objectives had failed. There were many who foresaw this and argued against the concept of the universal reliable tube. While their voices were drowned in the babble of those who usually had something to sell as each new reliable tube formula was advanced, the evidence available today preponderably supports their contentions.

The reliability of any product cannot be improved with out a complete knowledge of the precise application for which it is intended. When this is known, and suitable correlative tests have been established, it is quite possible that very significant improvements in performance, stability, life, and freedom from initial failures can all be accomplished by a standard or a specification which accurately describes that particular product-but only insofar as its use in that specific application is concerned.

Where the use of prototypes is involved in older equipment, the chances of making significant improvements are less than where new tubes are being designed into new equipments. The requirement that tubes fulfill more than one specific function, or that they work in existing equipment with no circuit alterations, minimizes the amount of improvement that can be expected.

Existing, so-called reliable tubes have only a small chance of being actually more reliable in replacing proto types in uncorrelated applications. In many instances, the exact reverse has been experienced, wherein so-called reliable types have proved to be less reliable than the original equipment types. This comes about quite logically, and is in no way a contradiction of logic. In one instance, it turned out to be the result of the extra micas used in the more reliable types to reduce vibration; this resulted in greater cathode cooling and, in a particular application, failure of a local oscillator to function at high frequencies. The tube had been made more reliable for certain pulse applications, but that didn't make it a better high frequency oscillator. Examples of this type are fairly common with many of the early reliable types, where a change was made to correct one specific trouble without enough knowledge of all the applications involved to make sure that none of them were adversely affected.

As a result of this, there have been developed more recently, several groups of tubes that are not known as reliable tubes. However, their use in the particular applications for which they are intended may materially improve their over-all reliability. These are not universal tubes and their use in some other function, other than their intended one, may even result in an over-all reduction in equipment reliability. Typical of these are certain unique low-noise audio tubes that may be used to replace older versions like the 12AX7. Then there are a number of computer tubes intended just for that type of service.

Even the early rugged tubes are definitely more rugged if physical shock is what you are talking about. More recently, instrument tubes have been introduced for use in such equipments as scopes and direct-coupled amplifiers.

In fact, the modern trend is not to attempt to make some one tube better for all applications, but instead to make specific tubes, or families of tubes, which are more precisely tailored to meet a limited number of applications.

This is leading to a lot more tube types, a trend that will continue as we enter the period of specialization.

In summary, it can be safely stated that there are not now, and there are not likely to be, any universally reliable tubes. We can have unreliable universal tubes, or we can have highly reliable special-purpose, limited-application tubes. But we can't have both in one tube. They are incompatible requirements. In our next section, we will examine some of these special tubes when we answer the question--"Why So Many Tube Types?"

Prev. | Next | Index

Top of Page   Home

Updated: Thursday, 2021-08-26 18:35 PST