Even though the three-element Audion had been invented by Lee De Forest in 1906, a decade later the quality of the tubes produced by his company still varied widely. The result, as this article acknowledged, was that "very seldom are two Audion bulbs exactly alike", and newly-produced Audions had to be individually tested and graded. The company had only a tenuous understanding of how to produce consistent results, and this article noted that within a given production batch "There may be one X grade bulb in 100 or there may be twenty--no one can tell". (Two years after this article appeared, the U.S. Navy placed an order with DeForest for 2,000 vacuum tubes, but had to reject 90% of them as substandard).

In retrospect, the general consensus is that De Forest did not have a very good grasp of how his invention worked. In particular, he believed in leaving too much gas in the tubes -- throughout this article there are numerous comments on the danger of having too high a vacuum. But later research showed that the electron flows which powered the Audion generally worked best in high vacuums, so later tubes were much more highly evacuated, and thus much more consistent and reliable.

 
QST, March, 1916, pages 41-44:

PRACTICAL  POINTERS  ON  THE  AUDION

By A. B. Cole, E. E.
Sales Manager -- De Forest Radio Tel. & Tel. Co.

A. B. ColeTHE purpose of the present article is not to sell Audions, but to assist those who use the justly famous Audion to obtain the best results, with a view to promoting the best interests of the amateur experimenter. So much has been written without authority of the patentees which is partially incorrect or serves to cause the experimenter to aspire to fields which cannot be achieved in a practical fashion by him and so many false directions have been given for making apparatus from the regular Audion Detector for which it is not adapted, that it seems only right to set the experimenter on the proper track, showing what can be done and how to do it, pointing out the fallacies in statements made by unauthorized parties.
    Be it known, first of all, that the writer is one of the pioneer amateurs, whose experimenting in wireless matters dates from 1904, and although engaged in commercial as well as amateur work, is still an amateur at heart and thoroughly in sympathy with amateur work, not as some whose first training was commercial and who look down on the amateurs as a small detail from whom revenues must be extracted because their superiors wish it, but strictly as an amateur of many years' experience, and, therefore, qualified to assist those working in this field in an intelligent way.
    In 1908, I used one of the very early types of Audion, so that it is not very new to me. As Sales Manager for the manufacturers, I have investigated and corrected probably every kind of trouble which can be had with the Audion. Yet, I am frank to admit that I do not know it all, being thereby different from outsiders who can write books on it from their limited experience with a few bulbs.
    Now, you should know that very seldom are two Audion bulbs exactly alike in their characteristics, but not much variation is permitted because too much in the way of instructions would have to be furnished the user, and so those not coming within the limits are discarded or furnished only to those of considerable experience.
    The Audion is a peculiar instrument in many ways. If used in connection with the proper circuits and accessories, it is the most sensitive and reliable detector ever invented, which fact cannot be denied. If, for example, the "grid condenser" is of the wrong capacity or inefficient, or the wiring of the circuits is not correct or the insulation of the wires of these circuits poor, or if they are placed too close together, trouble is very likely to occur which is difficult to locate and is invariably charged to the "bulb." It is remarkable how easy it is to have trouble if the circuits and parts of the detector as a whole are not correct, and this is one of the main reasons why the bulbs are not furnished separately and why all guarantees are immediately dissolved if an Audion bulb is used otherwise than in the Audion Detector made for it. One of the best amateurs in the country who has used the Audion for some time sent in his bulb on request, as he could not obtain good results. Investigation showed that he had taken the detector apart and assembled it in a complete receiving cabinet and after much loss of time, it was discovered that the "grid condenser" did not suit his taste and so he built another which consumed internally in losses nearly all of the incoming energy. The bulb, used in a regular Audion Detector, was excellent. This is only one representative case of many.

AUDION  BULBS

    It has often been asked, "How is an X grade Audion Bulb made, and what makes it different from the S grade?" Also, "How are the bulbs tested?" These are questions of interest to all users, and rightfully so.
    The process of testing Audion Bulbs is one of the most careful and expensive tests in the entire electrical industry. A bulb used in the proper detector may give extremely loud signals from stations twenty or thirty miles away, and yet not be really sensitive to weak signals from great distances, or it may give much weaker response to nearby stations than a crystal detector and still be extremely sensitive to weak signals over long distances.
    Theoretical methods of testing are, therefore, of no value. The only practical method is to compare with a standard, under actual working conditions, receiving weak signals. The standard is set by comparison with the best crystal detector. The unknown bulb is connected in circuit on a double detector, and the oscillations are tuned from one bulb to the other, reducing coupling and making the necessary changes in capacity to counterbalance change of mutual inductance, until the signals can be heard on one bulb and not on the other. Testing by throwing the circuits out of tune is not satisfactory because the Audion is a potentially operated detector.
    If the unknown bulb is equal to or better than the standard, it is passed, but otherwise discarded. If it is sufficiently more sensitive, it is passed as the "X" or extra sensitive grade. There may be one X grade bulb in 100 or there may be twenty--no one can tell. There may be ten S grade bulbs in 100 or there may be fifty. The Audion bulb looks simple, but is one of the most difficult instruments to make ever invented.
    The testing is done by expert, commercially licensed operators, of years of experience, and thoroughly familiar with the Audion. The employment of a beginner, or one only experienced in operating, for the work, would be fatal. The tuning apparatus is regular amateur equipment, not the specially designed and necessarily more efficient apparatus made for the Audion.

AUDION  TROUBLES  AND  HOW  TO  CORRECT  THEM

    (1) The most common cause of trouble is due to exhaustion of one or more of the "B" high voltage dry flashlight batteries. These batteries are at best unreliable, from their very nature. If only one becomes exhausted until it registers 3½ volts or less on a voltmeter, the efficiency of the entire receiving set is remarkably decreased. The great difficulty is to convince the operator of the necessity of testing every battery often with a voltmeter. As soon as one or more shows 3½ volts or less, it must be replaced, and all connections must be soldered. If not replaced at once, the operator always brings up the intensity of the filament, and, of course, burns it out before long--then says the bulb was defective.
    (2) The next most common cause of trouble occurs when the operator replaces one or more cells, and connects them backwards with respect to the others, connecting the carbon to the carbon of the next battery.
    Then again, when the operator replaces the entire set of batteries, he often connects the carbon to the filament of the bulb. The detector cannot operate unless the negative or zinc is connected to the filament. And again, the bulb was all right, but now has lost its sensitive qualities!
    (3) The third cause of complaint, due entirely to the operator, is the fact that, "The old bulb was very good, and the new one cannot compare with it." This is because of either of the above causes, or he does not try reversing the connections from the "A" or lighting battery. Some bulbs will not operate at all unless this is done to find which way is best.
    (4) Then, some operators use the Audion in a cold place. Consider the surface of glass of the bulb, and realize that the very small quantity of gas in it must necessarily be affected by change of temperature. The Audion works best when, the temperature of the room is 60 degrees or higher.
    (5) Then comes the complaint from a beginner who says, "The Audion brings in 600 meters and amateur stations better than any other detector I have ever tried, but Arlington on his 2500 meter wave is much stronger with a crystal detector." This is easy to explain. The Audion is selective to some extent as to spark frequency, but positively is not selective as to wave length. It is exactly as sensitive at 3000 meters as at 200. If you will consider the proportions of an amateur tuner, you will realize that it is most efficient from 200 to 1500 meters. They are all built in this way. When a loading coil is used in the primary circuit, the efficiency of the primary is decreased, although its period is increased.
    The secondary, if one is used, is seldom loaded. A variable condenser will help to some extent but unless the tuner can receive with high efficiency the long waves, it is impossible to operate the Audion to full efficiency. A crystal operates whether exactly in tune or not, but the Audion, being potentially operated, must have efficient tuning equipment if its advantages can be realized to the fullest extent.
    (6) Do not try to use a fixed condenser in series with an Audion Detector. Every one of these instruments has a special mica condenser within it, properly connected and built, and I do not believe it can be improved upon. Unauthorized parties give all kinds of advice, but it seems only reasonable that the manufacturers should know and install what is right.
    A variable condenser across the tuner posts is of great value in obtaining the exact point of resonance, but nothing in the way of a condenser should be used in series to reduce efficiency.
    (7) Then there is the operator who turns off his lighting battery and neglects to cut in more resistance by means of the rheostat. The battery recuperates while standing and when turned on again, often burns out the filament or seriously injures it. Remember that a sensitive and delicate instrument requires a little thought and care.
    (8) The use of a magnet near an Audion Bulb sometimes increases the intensity of signals, but we have seen filaments literally bent out of place due to this, and a common result is short life of the bulb. (Note below under Fallacies (3).
    (9) Remember that all regular Audion Bulbs have 3½ volt filaments and never connect more than 6 volts to it. The rheostat will handle effectively 6 volts, but no more should be employed.

ADJUSTMENT  OF  THE  AUDION

    (1) Nearly every Audion bulb has two critical points. One is found when the "A" and "B" batteries are adjusted to certain points. The other is found, generally with the "B" battery adjusted to a higher voltage than above and the filament operating at less brilliancy with the lighting battery connections reversed. Some bulbs have only one critical point, within the limits of the usual "B" battery voltage, and so it is important to TRY REVERSING THE LIGHTING BATTERY connections to the detector.
    An Audion should always be operated with the higher "B" voltage and lower filament intensity to obtain greatest life.
    (2) The "critical point" is reached at certain adjustments of both the "A" and "B" batteries. At this point, a hissing sound is generally heard in the receivers. If this is present, the filament brilliancy should be decreased until the sound is just audible or is just below the audible point.
    When strong signals are received, they may be much increased in intensity by increasing both the "B" battery voltage and the filament brilliancy, but while greater volume of sound is obtained, the bulb is not in its most sensitive condition at this adjustment. With some excellent bulbs, no hissing sound can be heard at any adjustment, or it may appear, but be very weak.
    (3) The "Blue Glow" appearing at certain adjustments of the "B" battery in the old style tantalum filament bulbs is often not found at all in the tungsten and Hudson filament types, and is not necessarily an indication of sensitive qualities.

FALLACIES

    (1) Making an Amplifier from a Detector. There are several reasons why this cannot be done with even fair efficiency. If the necessary three winding transformer is not exactly made and balanced, the results are poor. The Audion Amplifier Bulb is entirely different from the Audion Detector Bulb in construction and vacuum.
    The result of trying to make an amplifier from a detector is only a makeshift in which the efficiency is so low as to make it an expense out of proportion to the benefits obtained.
    The Detector bulbs cannot last long enough to make it worth while, and the efficiency is very low.
    (2) Receiving Continuous or Undamped Waves. Regular Audion Detector Bulbs are not adapted for the reception of continuous waves, because the vacuum is not correct for the purpose and because the filaments must be operated at such a high intensity that they give very short service, making them unnecessarily expensive.
    Then, their use in this way causes the vacuum to gradually increase until 75 to 150 volts are required for the "B" circuit.
    (3) "Amplifying" Circuits. This is the same old story, begun with the creation of the world, of obtaining something for nothing.
    If you will think, you will realize that if you pass through an Audion Bulb or any other apparatus, two or three times as much energy as it is designed to carry, the result will be a short life. This appears in "amplifying" circuits in operation of the filament at excessive brilliancy and in rapid, deterioration of the filament from carrying abnormal power. The vacuum increases to an extremely high value in most cases.
    The final result is expense for renewal bulbs far out of proportion to results obtained and general dissatisfaction--not due to any fault of the manufacturer, because the instruments and bulbs are sold and licensed for only one purpose, but because the operator is looking for "something for nothing," which is impossible in the long run.
    There are many operators of experience with the Audion who know of all the above points, and many more who do not, and these suggestions, the result of long and active experience of one who has corrected more Audion troubles than perhaps any other one individual, should be given that close attention and careful consideration which they deserve, that the very best of results may be obtained and the longest possible ranges covered.

BLUE  DISCHARGE  OF  GLOW

    This appears in some Audion Bulbs and not in others. If allowed to persist, the vacuum automatically increases. For this reason the glow should not be allowed to appear and certainly not to continue, as the vacuum may rise to a very high value, requiring very high voltage in the "B" battery. Many amateurs cannot seem to comprehend the value of instructions of this kind, and if they ruin bulbs in this way, always claim that the bulb "was defective," although it is their own negligence which caused the condition.
    A super-sensitive detector like the Audion should be handled and used with reasonable care and intelligence. It is not "fool-proof."

ADJUSTMENT

    One of the most essential points very seldom appreciated or even known by most operators is the fact that a very fine regulation of the "B" or high voltage battery potential is extremely important in securing utmost efficiency with the Audion. A difference of one volt may make a difference in range of as much as 25 per cent.
    The amateur types of Audion Detector manufactured until 1916 have a switch adjustment for this high voltage, providing at each step a difference of three to four volts. The Audion Bulbs were, of course, and still are tested on this type of instrument, so that they may operate properly with same.
    One kind of "trouble" is due to the fact that after an Audion Bulb has been used for a time, the vacuum changes slightly, so that the correct voltage required becomes slightly different in value. Now, it can be readily understood that the operator will have difficulty in obtaining best results when this occurs, because with the switch on one point, the voltage is too low, and on the next higher point, it is too high, and the hissing sound in the telephone receivers may drown out the incoming signals.
    The purpose of replacing the "step by step" switch by a potentiometer, which has been done in the new instruments, is to overcome this difficulty and provide any desired voltage within the limits of the battery.
    For the benefit of the large number of operators whose Audion Detectors are provided with "step by step" high voltage control switches, the following is recommended on account of the difficulties certain to be encountered, even if the potentiometers were furnished them, in installing same in their instruments.
    If a small box is made, large enough to hold one 3 cell flashlight battery, and a switch is mounted on the box, and arranged to cut in one cell at a time, and this device is connected in series with the telephone receivers, and the receiver binding posts on the detector, a very satisfactory adjustment will be provided, enabling the operator to accomplish excellent results with bulbs otherwise considered less sensitive than standard. Of course, this extra battery must be connected in circuit in the proper direction, so that it acts with and not against the regular high voltage batteries of the detector. The proper direction can be readily determined by trial.
    This arrangement is thoroughly recommended to every user of an Audion instrument. It is certain to produce results far superior to any ever previously obtained.
RJ6 receiver