Ever since electricians had experimented with voltaic currents, and especially since the introduction of the electric telegraph, it had been a familiar fact that a loose or imperfect contact in the circuit caused a resistance to the flow of the current and interrupted it more or less completely. To obviate the occurrence of loose or imperfect contacts, binding-screws were invented; and many were the precautions taken to make tight contacts at joints in the line, the resistance of which it was desirable to maintain at a minimum. Young telegraphists were particularly instructed to press their keys well down in signalling, because a light contact would offer some resistance which, on an increase of pressure, would disappear. In fact, it was generally well known that the resistance of two pieces of metal or other conducting material in contact with one another might be made to vary by varying the goodness or badness of the contact with the application of more or less force. This fact was known to apply to good conductors, such as copper and other metals, and it was known to apply also to non-metallic conductors, such as plumbago. Plumbago points were used by Varley for the contacts of relays; it having been found that points of platinum were liable to become fused together with the passage of the current, and by so sticking rendered the instrument useless. Since plumbago was known to be infusible, it was hoped that a plumbago contact would prove more reliable. In practice, however, the plumbago relay did not turn out so well. True it did not fuse, or stick, or rust; but it was even more liable than platinum to form imperfect contacts, the resistance of the light contact being so high that a sufficient current did not pass. It is not known whether other non-metallic substances were tried; probably not, because of non-metallic substances plumbago is one of the few that are good conductors.
According to Edison (British Patent, No. 792, 1882), compressed graphite is a substance of great conductivity. According to Faraday (‘Exp. Res.’ vol. i. p. 24), retort-carbon is an excellent conductor. Both graphite and retort-carbon agree with the metals in the property that the electric resistance offered at a point of contact between them varies when the pressure at the contact is varied. It is indeed remarkable through what wide ranges of resistance the contact between two good conductors may vary. The resistance of contact between two pieces of copper may be made to vary in a perfectly continuous manner by changes of pressure through a range, according to Sir W. Thomson, from a small fraction of one ohm, up to a resistance of many thousand ohms. The same is true of silver, brass, and many other good conductors, including graphite and retort-coke, though with the latter materials the range of resistances is not so great. With partial conductors, such as oxide of manganese, sulphide of copper, sulphide of molybdenum, &c., and with bad conductors, such as lamp-black and selenium, whose conductivity is millions of times less than that of graphite, copper, and other good conductors, it is impossible to get equally wide variations of resistance, as the amount of pressure at a point which will bring the bad conductors into intimacy of contact, will not turn them into good conductors. Platinum being in the category of good conductors, is amongst those substances which yield a very wide range of electrical resistances at the contact-points which are submitted to varying pressures.
With the very highest conductors, such as silver and copper, the electrical range of contact-resistance is higher than with those of lesser conductivity, such as lead, platinum, graphite, and retort-coke.
But though the range of variation in electrical resistance at contacts is highest for the best conductors, there comes in another element, namely, the range of distance through which the contact-pieces, or either of them, must be moved in order to pass through the range of variations of resistance. This is quite a different matter, for here the best conductors have the smallest range, and some that are not so good a greater range. In any case the available range of motion is very small—to be measured in minute fractions,—millionth-parts, perhaps,—of an inch. So far as experiments go, however, silver has the smallest range of all, then gold, then copper. Platinum and nickel have a considerably wider range, plumbago and retort-coke a still wider one.
It is an extremely difficult matter to decide what is the precise nature of that which goes on at a point of contact between two conductors when the pressure at the point is altered. The principal suggestions hitherto advanced have been that the change of resistance observed is due:—
It is admitted that this last suggestion, though it might account for a difference between different substances, in so far as they differ from one another in the effect of heat upon their specific resistance, implies as a preliminary fact that the amount of surface in contact shall be varied by the pressure. No convincing proof has yet been given that the alleged layer of air or other gases has any real part to play in the phenomena under discussion. Nor can the hypothesis, that minute voltaic arcs are formed at the contact be regarded as either proven or probable.
The only two theories that have really been investigated are (a) and (b) of the above series. Of these two (b) is certainly false, and (a) is probably, at least to a very large extent, true.
It is often said by persons imperfectly acquainted with the scientific facts of the case, that carbon is used in telephone-transmitters, because the resistance of that substance varies with the pressure brought to bear upon it, whilst with metals no such effect is observed. This statement, taken broadly, is simply false. Mr. Edison has, indeed, laid claim to the “discovery” (vide Prescott’s ‘Speaking Telephone,’ p. 223), that “semi-conductors,” including powdered carbon and plumbago, vary their resistance with pressure. All that Mr. Edison did discover was that certain substances, whose properties of being conductors of electricity had been known for years, conducted better when the contact between them was screwed up tightly than when loose. The experiments made to test this alleged “property” of carbon are absolutely conclusive. The author of this book has shown[39] that when a rod of dense artificial coke-carbon, such as is used in many forms of telephone transmitters, such as Crossley’s for example, is subjected to pressure varying from less than one dyne per square centimetre up to twenty-three million times that amount, the resistance of the rod did not decrease by so much as one per cent. of the whole. In this case any doubt that might have been introduced by variable contact was eliminated at the outset by taking the precaution of electro-plating the contacts.
In 1879, Professors Naccari and Pagliani, of the University of Turin, published an elaborate series of researches[40] on the conductivity of graphite and of several varieties of coke-carbon, and found, even with great changes of pressure, that the changes of electric resistance were practically too small to be capable of being measured, and that the only changes in resistance appreciable were due to changes of contact.
In January 1882, Mr. Herbert Tomlinson communicated to the Royal Society[41] the results of experiments on a number of electric conductors. The change of conductivity by the application of stress was found to be excessively small. For carbon it was less than one-thousandth part of one per cent. for an increase of fifteen lbs. on the square inch in the pressure. For iron it was slightly greater, and for lead nearly twice as great, but with all other metals less. If this alleged property were the one on which the action of telephone transmitters depended, then lead ought to be twice as good a substance as graphite; whereas it is not nearly so good.
Professor W. F. Barrett, in 1879,[42] made some experiments on the buttons of compressed lamp-black used in Edison’s transmitter, and found that when an intimate contact was satisfactorily secured at the beginning, “pressure makes no change in the resistance.”
In the face of all this precise evidence, it is impossible to maintain the theory that the electric resistance of plumbago or of any other such conductor varies under pressure. The only person who has seriously spoken in favour of the theory is Professor T. C. Mendenhall, but in his experiments he took no precautions against variability of contacts, so that his conclusions are invalid.
More recently still, Mr. O. Heaviside and Mr. Shelford Bidwell have experimented on the variations of resistance at points of contact.[43] Mr. Heaviside’s experiments were confined to contacts between pieces of carbon, and though extremely interesting as showing that the resistance of such contacts are not the same, even under constant pressure, when currents of different strength are flowing, do not throw much light on the general question, because they leave out the parallel case of the metals. Mr. Bidwell’s very careful researches were chiefly confined to carbon and bismuth. The choice is unfortunate, because bismuth the most fusible and worst conductor amongst metals (save only quicksilver) is the one metal least suited for use in a telephone transmitter. Mr. Bidwell’s conclusions, so far as they are comparative between carbon and “the metals,” are therefore necessarily incomplete.
Professor D. E. Hughes, whose beautiful invention, the Microphone, attracted so much attention in 1878, has lately thrown the weight of his opinion in favour of the view that with carbon contacts the effect is due chiefly to an electric discharge or arc between the loosely-contiguous parts. But Professor Hughes’s innumerable experiments entirely upset the false doctrine that a “semi-conductor” is necessarily required for the contact-parts. Speaking recently,[44] he has said: “I tried everything, and everything that was a conductor of electricity spoke.” In 1878, in a paper “On the Physical Action of the Microphone,” Professor Hughes stated:[45] “the best results as regards the human voice were obtained from two surfaces of solid gold.” Hughes also found carbon impregnated with quicksilver in its pores to increase its conducting power to work better than non-metallised carbon of inferior conductivity. Quite lately Mr. J. Munro has constructed successful transmitters of metal gauze, having many points of loose-contact between them.
It seems, therefore, much the most probable in the present state of investigations, that the electric resistance of a contact for telephonic purposes is determined solely by the number of molecules in contact at the surface, and by the specific conductivity of those molecules. The element of fusibility comes in to spoil the constancy of the surfaces in action; and hence the inadmissibility of general conclusions with respect to all metals drawn from the behaviour of the most fusible of them. At a mere point in contact physically with another point, there may be hundreds or even millions of molecules in contact with one another, all acting as so many paths for the flow of the electric current. An extremely small motion of approach or recession may suffice to alter very greatly the number of molecules in contact, and the higher the specific conductivity of the substance, and the denser its molecules, the shorter need be the actual range of motion to bring about a given variation in the resistance offered. Just as in a system of electric lamps in parallel arc, the resistance of the system of lamps increases when the number of lamps through which the current is flowing is diminished, and diminishes when the number of lamps connecting the parallel mains is increased; so it is with the molecules at the two surfaces of contact. Diminishing the number of molecules in contact increases the resistance, and vice versâ. Each molecule as it makes contact with a molecule of the opposite surface diminishes, by so much relatively to the number of molecules previously in contact, the resistance between the surfaces. Each molecule as it breaks from contact with its opposite neighbour adds to the resistance between the contact-surfaces. It may therefore be that the variations of resistance which are observed at contacts between all conductors, from the best to the worst, are all made up, though they appear to pass through gradual and continuous changes, of innumerable minute makes-and-breaks of molecular contact. The very minuteness of each molecular make-or-break, and the immense number that actually must occur at every physical “point” of contact, explain why the effect seems to us continuous. We owe, moreover, to Mr. Edison[46] the experimental proof that actual abrupt makes-and-breaks of contact can produce an undulating current when they recur very rapidly. Whether the heating action of the current itself may not also operate in changing the conductivity of the molecules which happen at the moment to be in contact is another matter. It may be so; but if this should hereafter be demonstrated, it will but confirm the contact-theory of these actions as a whole.
Assuming, then, broadly, that the observed resistance at a point of contact is due to the number of molecules in contact and to their individual resistances, it is evident that the property of varying resistance at contact ought to be most evident, ceteris paribus, in those substances which are the best conductors of electricity. Unfortunately, the cetera are not paria, for the question of fusibility comes in to spoil the comparison; and carbon, which has less fusibility than the metals, is commonly credited with giving a better result than any. This common opinion is, however, based on comparisons made without taking into consideration the question of range of motion between the parts in contact, and without taking into consideration the point that whilst some forms of carbon are excellent conductors, others do not conduct at all. In a telephonic transmitter so arranged that the actual range of motion shall be very small, the metals are just as good as carbon—some of them better. I have heard from a transmitter with contacts of pure bright silver better articulation than with any carbon transmitter. And this is exactly what theory would lead one to expect. As to the suggestion that plumbago makes a successful transmitter, because it is a “semi-conductor”—whatever that term may mean[47]—it is one of those suggestions which are peculiarly fitted to catch the unscientific mind as affording an easy explanation for an obscure fact; unfortunately, like a good many other similarly catching suggestions, it is not true. The very best conductor—silver—will serve to transmit articulate speech: and so will the one of the very worst conductors—lamp-black! So much for this fallacious doctrine of semi-conductors!
Reis used for his contact-points substances which, by reason of their non-liability to fuse or oxidize, were customary in electrical apparatus, and chiefly platinum. In his earliest transmitter (model ear), and in his last, platinum was used. In his lever-form of transmitter, so minutely described by von Legat, the material is not specified. The lever-shaped contact-piece was to be a conductor, and as light as possible, and since all metallic parts are particularly described as metallic, whilst this is not so described, the obvious inference is that this was non-metallic. The number of light, non-metallic conductors is so few that the description practically limits choice to some form of hard carbon. No other materials are named by Reis, but Pisko says (p. 103) that brass, steel, or iron might be used for contacts. Any one of these materials is quite competent, when made up into properly-adjusted contact-points, to vary the resistance of a circuit by opening and closing it in proportion to the vibrations imparted to the contact-points. That is what Reis’s transmitter was intended to do, and did. That is what all the modern transmitters—Blake’s, Berliner’s, Crossley’s, Gower-Bell’s, Theiler’s, Johnson’s, Hunning’s do, even including Edison’s now obsolete lamp-black button transmitter. Mr. Shelford Bidwell has very well summarized the action of the current-regulator in the following words: “The varying pressure produces alterations in the resistance at the points of contact in exact correspondence with the phases of the sound-waves, and the strength of a current passing through the system is thus regulated in such a manner as to fit it for reproducing the original sound in a telephone.”
Reis constructed an apparatus consisting of a tympanum in combination with a current-contact-regulator, or “interruptor,” which worked on this principle of variable contact, and he called it “The Telephone” (see pp. 57, 85). The very same apparatus we now-a-days call a “Telephone-transmitter,” or simply a “transmitter.” It is curious to note that Reis seems to have regarded his receiver or “reproducing-apparatus” as no new thing. He says explicitly (p. 56) that his receiver might be replaced by “any apparatus that produces the well-known galvanic tones.” “The Telephone” was with Reis emphatically the transmitter. Bell in 1876 invented an instrument which would act either as transmitter or receiver, and which, though never now used as transmitter, is still called “a Telephone.” Edison’s “sound-telegraph,” or “telegraphic apparatus operated by sound,” was patented in 1877. In his specification he never called his transmitter a “telephone;” that name he reserved exclusively for his receiver. He found it, however, convenient a year later to rechristen his transmitter as the “carbon telephone,” though throughout the whole of his specification neither “carbon” nor “telephone” are mentioned in connection with the transmitter! Within that year Hughes had brought out another instrument—“The Microphone”—which, like Reis’s instrument, embodied the principle of variable contact. Hughes’s instrument, usually constructed with contacts made of loose bits of coke-carbon, was simply a Reis’s Telephone minus the circular tympanum; and the really important new fact it revealed, was that very minute vibrations, such as those produced by the movements of an insect, when transmitted immediately through the wooden supports, sufficed to vary the resistance of a telephonic circuit, though far too slight in themselves to affect it if they had to be first communicated to the air and then collected by a tympanum. Put a specific tympanum to a Hughes’s microphone, and you get a Reis’s telephone. Take away the tympanum from a Reis’s telephone, and you get a Hughes’s microphone. Hughes is not limited to one material, nor is Reis. But the fundamental principle of the electrical part of each is identical. The Blake transmitter (Fig. 44), and the Berliner transmitter, and also Lüdtge’s microphone,[48] which was even earlier than that of Hughes, are all embodiments of the same fundamental principle of variable contact which Reis embodied in his “Telephone.”
The numerous experiments which Reis made, and the many forms of instruments which he devised, prove his conviction of the importance of his invention to have been very deeply rooted. He had indeed penetrated to the very soul of the matter. He did not confine himself to one kind of tympanum, he tried many, now of bladder, now of collodion, now of isinglass, and now of thin metal. He varied the forms of his instruments in many ways, introducing the element of elasticity by springs and adjusting-screws. Though he chiefly employed one metal for his contact-pieces, he did not limit himself to that one, but left us to infer that the principle of variable contact was applicable to any good conductor, metallic or non-metallic. He knew better, indeed, than to limit himself in any such fashion; better, indeed, than some of the eminent persons who are now so willing to ignore his claims. Modern practice has taught us to improve the tympanum part of Reis’s invention, and to obviate the inconveniences to which a membrane is liable: in that part we have gone beyond Reis. But in the question of contact-points for opening and closing the circuit in correspondence with the vibrations, we are only beginning to find how much Reis was a-head of us. We have been thrown off the track—blinded perhaps—by the false trail of the “semi-conductor” fallacy, or by the arbitrary and unnatural twist that has been given by telegraphists to Reis’s expression, “opening and closing the circuit,” forgetting that he practically told us that this operation was to be proportional to, “in correspondence with,” the undulations of the tympanum. When we succeed in freeing ourselves from the dominance of these later ideas, we shall see how much we still have to learn from Philipp Reis, and how fully and completely he had grasped the problem of the Telephone.