Research & Development Center
---How To Books, Projects & Plans---

Subscribe to our FREE monthly newsletter, a Great source for tips, advice and updates. Plus! Free! Software Every Month...


How I Invented the Audion

By Dr. Lee de Forest

Written exclusively for the Electrical Experimenter

The first conception of a detector of Hertzian waves which should employ the medium of heated electrodes or heated gas came to me as follows: In the summer of 1900 I was experimenting on a new type of electrolytic detector, doing this work at night in my room in Chicago. The receiving apparatus was placed on a table beneath a Welsbach gas burner. A spark coil which I was using as my source of oscillations was located in a closet about ten feet distant. One night I noticed that whenever I closed the switch of the spark coil by means of a string running across the floor from my table to the coil, there was a decided change in the illumination from the Welsbach burner. The light from the gas mantel increased very perceptibly and resumed its normal low brilliance as soon as the sparking ceased.

This phenomena continued and imprest itself strongly upon my attention. My first thought was that I had discovered a new form of detector of Hertzian waves, of extraordinary sensitiveness, and was, naturally, much enthused, as any young investigator would have been under similar circumstances. But upon closer investigation of this novel phenomena I found that when the door of the closet was closed, or almost closed, the effect of the spark upon the gas burner ceased! This proved conclusively that I was dealing with sound waves coming upon a sensitive flame and not with electrical waves.

The delusion lasted, however, long enough to force upon my mind the conviction that heated gas molecules were sensitive to high frequency electrical operations, and I determined to investigate further at my first opportunity and actually discover evidence to substantiate my theory. I was unable to do this until the fall of 1902 or '03 when I returned to my gas mantel experiment. I first attempted to investigate the new detector phenomena by using two needles of steel, or platinum, placed close together in the incandescent Welsbach mantel. These two needles were connected to a dry battery and telephone receiver. I was, however, unable to obtain any appreciable current between the two electrodes in the mantel. I then investigated the flame of a Bunsen burner and soon found a point in the outside envelope of the flame w:here an appreciable current did pass between the two electrodes, making a soft fluttering sound in the telephone receiver. (See Fig. 1, Patent No. 824,638, issued in 1906.) Then, connecting one electrode to an antenna and the other to the earth, J heard for the first time signals in the telephone receiver; signals which represented clearly the sound of the transmitting spark. Here at last was actually demonstrated my earnest belief in the existence of this new detector principle.

My next step was to enrich the gas flame by putting a lump of potassium or sodium salt in the flame directly below the two platinum electrodes. This increased ionization caused increased flow of battery current, and a corresponding increase in sensitiveness of the new detector. I did considerable work then with various types of Bunsen burner arrangements for permanently enriching the gas flame, etc., and set up a laboratory type of flame detector which was actually used in 1903 for receiving signals from ships down the Harbor of New York.

The inconvenience of supplying a source of gas for the new detector was, of course, obvious, and I sought for other means of obtaining the necessary heated gas and heated electrodes. The electric arc first suggested itself. I anticipated that while this arc would be a detector, it would be exceedingly irregular and noisy in the telephone receiver. This was found to be the fact. The battery fed the arc thru the primary of the transformer, in the secondary of which was connected the telephone receiver, and, altho at times the looked for response to electric waves was thus obtained, the noise in the telephone receiver from the arc was so deafening that the idea was abandoned.

The next plan which suggested itself to me was to use incandescent filaments in an enclosed chamber. This arrangement as well as the gas detector, was illustrated in ray patent No. 979,275, which bears the first date of November 4, 1904. This patent application was not filed until the following February. The drawings, Figs. 2 and 6, of this patent show the incandescent or glow members both in the air, and sealed within a closed chamber.

It will be noted now that I approached the general problem of this new type of detector from an entirely different angle from that commonly supposed to-day. In the first place I always employed a battery, and this original battery was what is now universally called the "B battery." My source of electric current for heating purposes was second, and secondary, so that the vociferous contention of the advocates of the Fleming valve that the audion evolved from the Fleming valve, and was originally the Fleming valve with the "B battery" added as an afterthought, is entirely untrue.

It was now a very obvious development of the evolution thus far described, to partially exhaust the glass envelope containing the two incandescent electrodes, or filaments, so as to increase the conductivity of the space between them.

But Fig. 5 of this patent, showing the inclosed filament in a vacuum bulb, is interesting from another consideration because it contains the first embryonic germ of the later "grid" or third electrode. It was realized from the very first that a certain proportion of the high frequency energy from the antenna could be lost thru the bypath circuit supplied by the battery and telephone receiver. In order to prevent this the arrangement shown in Fig. 5 was tried out, wherein the local and high frequency circuits are kept separate. In this arrangement, as actually tried in the gas flame, no actual advantage was observed, because the high frequency electrodes were necessarily some slight distance from the path conducting the direct current, and hence the effect of the high frequency currents upon the ions carrying the local current was weakened.

During the years of 1904 and '05 my duties kept me almost continually away From my laboratory, in traveling about the country directing the installation of numerous radio stations; consequently there was little opportunity for prosecuting this development work, and carrying out the designs and sketches which I made from time to time. In fact, it was not until 1905 that a lamp manufacturer was found to undertake the construction of the various experimental forms of lamps which I had designed as a successor to the flame or arc detector. I was familiar in 1905 with Prof. Fleming's work on the subject of the "Edison Effect" as utilized for the rectification of high frequency oscillations, or currents.

This was interesting to me only as an evidence of growing activity along lines similar to those of my new detector. I was familiar, of course, with the phenomena originally discovered by Edison, and investigated and developed by Howell. Fleming, Wehnelt, et al; but from the very first of my work with radio detectors I always had in mind a relay, in the true sense of the word, not a rectifier; in other words, a detector in which the energy of the audible signal was supplied from a local battery. This local energy being merely controlled or released by the incoming signal. It had always been obvious to me that such a device should be more efficient than any form of rectifier where the signal indication is effected only by the energy actually received thru the antenna.

Fig. 1, Patent No. 824.637. is an obviously practical development of Fig. 5 of my first patent, No. 979,275. In this figure two incandescent filaments are sealed in the glass bulb, and each lighted from its own "A" battery. Here, as always, the original separate "B" battery is shown.

This figure is interesting to many, who until recently, if even now, have never heard that an audion of this type works equally well whether both electrodes are incandescent, or whether one is incandescent and the other cold. It might be well to point out now in this connection that during the patent trial - the Fleming valve vs. the Audion - a demonstration was made before the court where all three electrodes were in the form of filaments, where each could be heated to incandescence by a separate battery. The demonstration showed that when this device was connected up as an audion that either outside filament could be used as the "plate" electrode indifferently, and the signals were of the same intensity whether two electrodes were cold or all three hot. The purpose of this test was to demonstrate beyond all cavil the falseness of the assumption that rectification plays in any way an important or essential part in the operation of the three electrode audion.

Fig. 3 of the above mentioned patent No. 824,637 shows the next obvious step in the evolution of the audion i.e., doing away with the unnecessary battery for heating one of the electrodes. It was obvious, of course, that so long as the "B" battery was properly connected so that its positive pole led to the cold electrode, there was no necessity or advantage in heating this second electrode. The audion in this form - one hot and one cold electrode, and the "B" battery connected with this positive terminal to the cold electrode, was used for some months, and shows a sensitiveness as a detector, superior to that of the electrolytic detector, and far superior to that of the Fleming valve rectifier. We had, thanks to the "B" battery which was invariably employed, a genuine relay or trigger action of he high frequency oscillations upon the normal current-carrying ions, or electrons, passing between the two electrodes.

As everyone familiar with incandescent lamp or X-ray bulb phenomena knows, the proportions of current passing between two electrodes therein (one or both being hot) carried by ions or carried by thermions, depends chiefly upon the degree of exhaustion of the bull). The gradual preponderance of thermionic conduction over ionic conduction has been gradually increasing as the art has progrest with improvements in pumps, knowledge of the fine points of exhaustion, etc. Thus it has always been impossible to lay a finger upon a certain date or upon any audion type of device, and say, "This marks the distinction between an ionic, or gaseous detector, and a thermionic detector." It is in fact impossible to prove even to-day when audions or oscillions are exhausted to the highest degree that the conductivity is entirely electronic.

In the spring and summer of 1906 I had opportunity to spend a good deal of time on the audion problem, and was always seeking to improve it- efficiency. Keeping in mind then the disadvantage of directly connecting the high frequency circuit with the local circuits, and harking hack to the four electrode gas-flame detector above mentioned, I sought to keep one electrode of the high frequency circuit distinct from the two electrodes of the local current.

Obviously the most simple experiment was to wrap a piece of tinfoil around the outside of the glass bulb and connect this to one end of the secondary transformer of my receiver. The other end of the transformer was connected to the filament, it being obviously unnecessary to employ four electrodes to effect the end desired. Exactly this arrangement, with the third electrode around the outside of the bulb, is shown in Fig. 3 of patent No. 841.386, which was filed in August, 1906. It will probably be recalled that this outside electrode has been very recently "re-discovered'' with considerable eclat and acclaim ! I also at this time wrapt a coil of wire around the bulb, connecting one end of this to the antenna and the other to the ground, seeking thus to effect the ionic conductivity between the inner electrodes by electro-magnetic influence from the high frequency oscillations passing around this helix. See Fig. 4, patent No. 841,386.

The arrangement of the external tinfoil belt may therefore be called the parent of the third electrode. It showed a decided improvement in the sensitiveness of the detector, as I had anticipated. I recognized that by this arrangement I had in effect a condenser between the filament connection and a hypothetical third electrode, which consisted of the conducting layer of gas located on the interior walls of the bulb, the other arm of this condenser being the tinfoil belt outside the glass. I recognized also that this was a very inefficient and indirect way of impressing the effect of the high frequency oscillations upon the conducting medium between the filament and plate. The third electrode should therefore be placed inside the bulb. I immediately instructed McCandless & Company to make such a bulb. The first type of this third electrode was in the form of a plate, located on the opposite side of the filament from the "B" battery plate. This arrangement showed the increased efficiency and sensitiveness anticipated. It is shown in Figs. 2 and 4 of patent No. 841,387, filed October, 1906. This is the audion amplifier and telephone relay patent. Fig. 2 of this patent is interesting as showing also for the first time a third battery ("B") in the external circuit between the third electrode and the filament. T in this figure represents the high frequency transformer.

In Fig. 4, where this battery is omitted, is shown for the first time the grid stopping condenser C. In studying this type of three electrode bulb, I recognized that the third electrode was not yet in its most efficient position. It should be placed directly in the path of the ionic or thermionic stream, passing from filament to plate, where the high frequency electric charges imprest on the electrode could best affect this stream. But if placed directly between two electrodes, a solid plate, of course, would constitute practically a complete barrier. Hence I devised the grid or perforated screen structure. In fact, the first audion where the third electrode was placed between the filament and plate utilized the wire bent in grid form which is familiar to every amateur or user of the audion prior to 1914.

This type of third electrode so located was so marked an improvement over the preceding three-electrode bulb, that shortly thereafter a patent was applied for on it. This was issued in February, 1908, No. 879,532. See Fig. 1 where the complete receiving arrangement and the grid audion is clearly shown.

The audion remained in this form for six years. During that time its merits became gradually recognized in Europe as well as here, and it was not long before the little stranger was, like its predecessor, the two electrode brother with "B" battery, adopted into the Marconi family, and like its predecessor re-christened the "Fleming Valve." As soon as the audion amplifier had been developed for long distance telephone service by the engineers of the Western Electric Company, and installed on most of the long distance lines of the A. T. & T. Co., we find certain English publications adopting it also into the Fleming valve family; and now after the three-electrode device has demonstrated its utility as a radio transmitter of absolutely constant undamped waves and made possible transoceanic telephony, we learn that this big brother which I first named the Oscillion, is also the Fleming valve. Not even ''junior" or "senior" is used to distinguish one from the other in this rapidly growing Fleming valve family.

The art founded on the three-electrode audion has grown of late years with enormous strides. The great war has produced a tremendous intensity of development for various military purposes and it cannot be disputed that the engineers of the Western Electric Company have taken a foremost position, and much of the present-day efficiency of the detector and amplifier has been due to their efforts, spurred on as they were by the difficult demands and specifications of our Signal Corps officers and engineers.

It is estimated that there has been constructed for the U. S. Signal Corps during the war between 200,000 and 300,000 audion and amplifier bulbs, and at least 50,000 small oscillators. In Great Britain wartime production has probably equaled or exceeded the above ; while in France we are informed that during the last two years of the war, the audion production has averaged about 5,000 per day! The French bulb is particularly interesting as being efficient and suitable in all three uses, detector, amplifier and oscillator. For such purposes, of course, a compromise in efficiency was inevitable, and maximum efficiency in either of these three branches has been somewhat sacrificed.

Considerable discussion has lately arisen as to the first use of the audion as an oscillator or source of alternating current. This matter is now being thrashed out in a multiple interference procedure in the U. S. Patent Office. But the evidence so far indicates that the writer's application of this property of the audion in the spring of l912, marked the first use of the audion as a generator of undamped electrical currents, In view of recent developments, particularly the highly interesting announcements of President Vail of the A.T.&T. Co., regarding multiplex wire telephony and telegraphy over a single conductor pair, it may be prophesied that the application of the audion as a generator of alternating currents will be fully as useful as that of detector and amplifier.

There is, in the writer's opinion, no doubt but that if the development of radio is not now made a Government monopoly, it will not he long before commercial trans-oceanic wireless telephony will be effected. This work, whether the generator be a bank of oscillions or a high frequency alternator, will be made possible only thru the extraordinary amplifying properties of the audion, when used as telephone repeater or relay.

The simplicity of the oscillion transmitter in small sizes, coupled with the extraordinary- sensitiveness of the "zero beat" audion detector or amplifier of received high frequency energy, warrants the belief that before long the wireless telephone will be installed on thousands of vessels, supplementing, and in many cases, supplanting the wireless telegraph. In addition there is an enormous number of small vessels where a wireless telephone installation is more feasible.

As indicative of the growth of the Audion Art, the number of patents issued on various devices and circuits dependent thereon, gives a pretty fair key: Up to 1912 there had been issued about 20 patents, all filed subsequent to 1904. To-day there are over 100 United States patents on the Audion Art, and the number is very rapidly growing. Regardless of what name may be applied to the device patented, practically everyone of these patents since 1906 shows the three electrode bulb. They may all therefore be properly described as the outgrowth of the ideas first shown in the audion amplifier - patent No. 841,387.

Copyright © 2019 Alternative Technologies

All of this article content and photos is Copyright © 2019 Alternative Technologies. You may use this Article for content for websites, articles, or blogs, but must contain this Copyright Notice and All links.

Or you can use this HTML code to link to this page for your websites, articles, or blogs


Home    Aircraft    Security    Plans    Software    Energy    Tools    Mind Power    Health Science   
e-Books Library    Mail Order Form

Copyright © 2019 Alternative Technologies