US2126464A - Electrical musical instrument - Google Patents

Description

9, 1938. L. HAMMOND 2,126,464

ELECTRICAL MUSICAL INSTRUMENT Filed April 2, 1958 10 Sheets-Sheet 1 [nvenfor Laure/7s Hammond A ffameys Aug. 9, 1938. HAMMOND 2,126,464

ELECTRICAL MUSICAL INSTRUMENT Zza 2 Laure/is Hammond y fl w, wwzm Affomeys 9, L. HAMMOND 2,126,464

ELECTRICAL MUS ICAL INS TRUMENT Filed April 2, 1938 10 Sheets-Sheet 3 J M A E-IO F/0 (THEORETICAL) '41 g" 50 h g VOLTAGE 0F POINT M (FIG. 1) AT START 60 OF NOTE DEFEND/N6 o/v KEY VELOCITY 0 6 CUTOFF VOLTAGE OF CONTROL TUBE/ VELOC/TY OF KEY 2 3g. 10 [nvemor Laurens Hammqnd y Mam, m 412; 5 m

A7 forneys Aug. 9, 1938.

L. HAMMOND ELECTRICAL MUSICAL INSTRUMENT l0 Sheets-Sheet 4 Laure/ 7s Hammond ZwMQM A ffo/"neys Filed April 2, 1958 vvvv 1 Q Q hmmuk wmmim ASE M v hmm domu ddom mom G82 WW I Filed April 2, 1938 I 10 Sheets-Sheet 5 ENE RGY ENERG Y [/7 venfor Laurens Hammond WW, $2 M 16M Affomeys Aug. 9, 1938. HAMMOND ELECTRICAL MUSICAL INSTRUMENT Filed April 2, 1938 10 Sheets-Sheet 6 N &$N Q m N www QON

B Laurens Hammond Aug. 9, 1938.

L. HAMMOND ELECTRICAL MUSICAL INSTRUMENT 10 Sheets-Sheet. 1

Filed April 2, 1938 WAVE ANALYZR Nofe No.1, .74

Note No.13

Nofe N0. 37 3 0 db/ 7 [nvenfor Laurens Hammond B MJWzMf-M l3 HAEMON/CS I Affomeys Aug. 9, 1938. L. HAMMOND ELECTRICAL MUSICAL INSTRUMENT- Filed April 2, 1958 10 Sheets-Sheet 8 V. 5 Z 2 ii 2 W m m M E J ddoodw w 52 5 ddoods A Jzy 20a HARMON/C 1 DEVELOPMENT OF NOTE N0. 25 AT VARIOUS VOLTAGES OF lnvenfo'r' Laure/7s Hammond Affomeys 5 m DA M Aug., 9, 13., L. HAMMOND 2,125,454

ELECTRICAL MUSICAL INSTRUMENT Filed April 2, 1958 1,0 Shets-Sheet 9 INTENS/ TV INTENS/TY V) 2 E z fnvemor Laure/7s Hammond MMMMA Affomeys Aug, 9 L. HA'MMQND 2,126,454

ELECTRICAL MUSICAL INSTRUMENT Filed April 2, 1958 10 Sheets-Sheet l0 [five/7m" Laure/vs Hammmd W W, MM am Patented Aug. 9, 1938 UNITED STATES PATENT. OFFICE 2,126,464 ELECTRICAL MUSICAL INSTRUMENT Laurens Hammond, Chicago, Ill.

Application April 2, 1938, Serial No. 199,613

49 Claims.

My invention relates generally to electrical musical instruments and more particularly. to instruments of this type in which the tones are produced by translating electrically originated and controlled current pulsations into sound.

The main object of my invention is to produce an electrical musical instrument playable from a keyboard on which it is possible to produce many novel and interesting effects, and at the same time make it possible for a pianist-adequately to interpret the great classical works of piano literature.

A further object of my invention is to produce an instrument upon which piano literature can be played, of a tone quality very reminiscent of piano, but with desirable characteristics not heretofore found in the tones of pianos, but which are found only in stringed instruments such as the violin. my invention there may be added to the piano tone a frequency shifting system resulting in a characteristic of the tone resembling the vibrato of a violin.

I am of course aware that electrical pianos have been built in which the hammer action and string have been retained, and by means of electrical translation devices the characteristic motions of the piano string have been translated into sound, through the operation of electrical pick-ups, ampliflers, and speakers. By this process it is possible to build electrical instriunents on which standard piano music may be adequately played, but in such systems it is not possible, for instance, periodically to shift the frequency of the various notes. In the instrument of my invention, Ii do not employ hammers or strings, but seek by purely electrical means to produce the same general acoustical efiect as that of the string under various intensities of hammer blow, with the collateral advantage of introducing other novel effects which are hereinafter described.

In order fully to understand the operation of the instrument of my invention, it is necessary to review, in a general way, some of the more obscure acoustic properties of the piano, so as to show what'eflects are necessary in the electrical system, in order that the music produced will be sufiiciently similar to that of the standard piano adequately to interpret the efiects intended by the composers of classical piano works.

In the first place, piano notes are, on the average, very complex. Thatis to say, a very large part of the energy of a single note, especially in the lower register, is contained in the energy of the harmonic frequencies as compared to that For example, in the instrument "oi.

of the fundamental. If all the notes of a concert type of grand piano, which is herein referred to as standard, are played with a blow of the same intensity, which I shall call medium intensity, it will be found that the brightness of the notes increases strikingly in playing down the scale. By brightness" is meant the amount of the total energy contained in the harmonics as compared to the energy in the fundamental. For example, the tones in a piano which I have measured, and the characteristics of which are graphed in Figure 12, the highest A or note No. 85(the notes being numbered consecutively from 1 to 88, the lowest note A, being note No. l)-has a total energy which exceeds the energy of the fundamental of that note by only 10%. That is to say, about of the energy is that of the fundamental tone, whereas the harmonic energy is only approximately 10%. In going down the scale to note No. 37, which is A below middle 0, it is found that thetotal energy on the harmonics has risen to a value as great as that of the foundamental alone. Harmonic energy then increases steadily and rapidly until for the lowest note of the piano, A, note No. 1, the total energy 'has risen to a value which is six and one half times as great as the energy of the fundamental frequency alone. So accustomed are we to this tonal arrangement in the piano that we naturally think of these tone qualities as being somewhat the same throughout the range of the instrument, whereas, by physical measurement they are entirely unlike in harmonic development.

If, instead of playing each note with the same intensity or" blow, a single note is selected and. played with intensities of blows ranging from the softest to the hardest, it will be found that the sound produced changes in two major respects. In the first place, it will be found that the total energy increases enormously with the g hard blow, the relationship being such that the total energy is approximately in direct proportion to the energy imparted to the key. The energy in an electric circuit is proportional to the square of the current or voltage. The energy necessary to move the key is proportional to-the square of the velocity imparted to the key, so it follows that in general, to simulate the touch response of a standard piano in an electrical instrument, it is necessary to produce an output signal of which the voltage is directly proportional to the velocity with which the key is struck. However, with key blowsof increasing intensity on a standard piano, the audibility of the note increases even more rapidly than its total energy increases,

because of the shift in the harmonic energy distribution. This is the second major effect. If a harmonic analysis is made of a soft note and then of a loud one, resulting from a change in the strength of blow delivered to the key, it will be found that the brightness of the note has increased along with an increase in intensity. The increase in brightness means that a larger proportion of energy is carried by the harmonics with the increasing strength of blow, which adds to the audibility, especially in the lower range, where the higher harmonics fall in the region of maximum sensitivity of the ear.

Thus, the skilled pianist has under his control, not only the loudness, but the harmonic content of the note, depending on the way in which he strikes the key. In the electrical instrument herein described, I have provided means whereby the key, which operates switches in timed succession, rather than actuating mechanical mechanisms such as hammers, is nevertheless so arranged as to be touch responsive in the same way as is the piano key. I mean by that, that the acoustic energy output will very closely approximate the square of the velocity with which the key is struck, and that the brightness of the note will increase with increasing key blow intensity, but in a manner which can be regulated in the electrical instrument, whereas it cannot be so regulated in the piano, except by taking out the hammers and changing the dimensions and properties of the felt covering the hammers, or changing the shape of the relatively harder cores of the hammers themselves.

Whereas, as stated above, the acoustic energy output of the piano is approximately proportional 'to the energy of the key blow, this relationship does not hold true for very light blows. In the standard piano there is a threshold value of key velocity below which there is no energy output. This is due to the fact that the velocity with which the key is struck must be sufllcient to throw the hammer far enough to reach the string. This limits the pianist in very soft passages because if he attempts to play more and more softly he reaches a point where his technique no longer permits him to play each note with equal intensity, or at least with the intensity which he desires and expects. This is one of the features of the standard piano which I believe can be improved upon, and it is therefore a further object of the invention to make the key operate in such manner that there is no threshold of key velocity below which the acoustic energy falls to zero. The operation of the key has been made such that the lowest possible velocity produces a threshold value of acoustic output. It therefore follows that there is a small range of intensities of key blows, all of them exceedingly soft, which gives rise to the same energy output and therefore greatly facilitates the pianist's technique in soft passages.

A further important characteristic of the standard piano is the way in which the sound decays after the key is struck and held down. Once the string is set in motion, its energy is gradually transferred to the sounding board and the acoustic output decreases in a logarithmic fashion to approach the value of zero asymptotically. That is, it would theoretically require an infinite length of time for the string to come to rest. The rate of decay of the sound is strikingly different for the highest notes and the lowest ones--the lower notes sounding much longer. The exact rate of decay plays a very important part in determining the characteristics of the piano tone. In the instrument of my invention, the rate of decay may be controlled, and the constants of the circuit are so arranged as to give, with at least one particular setting, the same rate of decay as that of the standard piano. If these values are departed from, the instrument may produce most interesting and novel effects, but the similarity to the piano will immediately be lost and many works of piano literature cannot adequately be played if these constants be appreciably changed.

We have seen therefore that a piano note consists of a complex tone quality which starts suddenly and thereafter decays logarithmically to zero. This is not the whole story. This may be illustrated by the use of. an electric organ of the type diclosed in my U. 8. Letters Patent No. 1,956,350 granted April 24, 1934. On this instrument it is possible to set up a tone quality which for any particular note in the middle range may be quite similar to that of the standard piano. By pressing the key and closing the swell pedal while so doing, it is possible to make the note come on instantly and decay logarithmically to a lower level. In this case, however the quality of the note does not change as it decays. While the note sounds generally like a piano note, there is a certain definite difference noticeable to the ear. To simulate the piano, it is necessary that the electrical apparatus be such that each note have a different harmonic content depending on the intensity with which it is sounding. This is similar to the effect discussed previously in connection with soft and hard key blows. Owing to the fact that the note, when loud, is brighter than when it is soft, the audibility of the note as it decays, decreases more rapidly at the start than the straight logarithmic decay of a note of constant harmonic content. As far as I am aware, I am the first to propose an instrument without hammer or strings in which this characteristic of the sound of a string struck by a hammer is obtained, and it is therefore a further object of the invention to reproduce this effect.

A further characteristic of the standard piano lies in the fact that the tone of a piano note, especially in the lower register, is made up of a fundamental and a long series of partials, which partials are not true harmonics of the fundamental-in the sense that they are of frequencies which are not exact integral multiples of. the fundamental. This is due to the fact that whereas the string continues to generate sound over a considerable period of time, it is not continuously excited, but receives all of its energy during the blow of the hammer. If the string were an ideal one--that is to say, having no stiffness of its own, a uniformly distributed mass in a diameter which could be completely neglected by comparison with its length-then it could be shown that such a string would give rise to partials of frequencies which were exact harmonics of the fundamental. This is not the case in the standard plane as may quickly be shown by watching the pattern shown on a cathode ray oscilloscope operated from a microphone picking up the sound of a piano note. If all the partials of the tone were exact harmonics, the shape of the pattern would not change appreciably during the decay of the note. Actual experiment will immediately show small amplitude high frequency waves which travel rapidly along on the wave produced by the lower frequencies.

Pianos are usually tuned by a beat system, and

-' Let us assume therefore, that a piano has just been tuned in accordance with the best, practice, and assume further, that double octaves are struck, for instance, As' Nos. 13, 25, 37 and 49. The second harmonic of the lowest M13) is very close in frequency to the fundamental of the M) an octave higher. Similarly, the fourth harmonic of the A03) is of a frequency in the vicinity of the fundamental of the M37) two octaves higher, and in the vicinity of the second harmonic of the M25) one octave higher. The eighth harmonic of the-lowest M13) is close in frequency to the fundamental of the M49) three octaves up and close to the fourth harmonic of the M25) one octave higher, and close to the second harmonic of the A(37) two octaves higher, etc. Now it is not possible to tune these four A's so that there will be no beats between them. Not only is it not possible to do so in a practical sense, but it is theoretically impossible as well. Without therefore considering the effects of the tam pered musical scale which introduces beats in chords where notes of other than unison pitches are involved, it will be seen that beats will occur between frequencies which are close to one another and which are derived from the different notes of chords.

Further, beats are introduced for another reason. Except for the very lowest notes of the standard piano, either two or three strings are used for each note. The multiple strings cannot be tuned exactly alike as a practical matter. or even if they were, they would not remain so for very long after tuning. Chords played on a freshly tuned piano give rise to groups of frequencies which are very close to one another but are not the same. The amounts of their differences are not of an order to suggest to the listener that the piano is out of tune, but these frequency differences play a very important part in characterizing the general effect and in giving life" to the tone. This is due to thefact that when two frequencies which are close to one another are temporarily in phase with one another they reinforce each other, whereas when they are temporarily out of phase they tend to cancel one another. Thus, the acoustic spectrum is continuously changing-certain frequencies coming on while others are going off, etc. These effects are generally similar to the effect of a vibrato on a stringed instrument such as a violin.

In listening to the piano, one is ordinarily not aware of the vibrato effect because of the fact that a vibrato is associated in our minds with a definite simple rhythmical change such as is heard when a violinist rocks his finger on the string. In that case, he changes the effective length of the string in such a way as periodically to raise and lower the pitch. This is a change in ,frequency which, together with the pattern of the enclosure in which he is playing, produces at the listener's ears a change both in frequency and in intensity.

A simple vibrato on a melody instrument such as the violin occurs at a periodicity which may be counted by a listener. Depending on room conditions and other factors, this may sound very much like a tremulant to the listener, and there is no very clear difference, under some circumstances, between a vibrato, which is customarily thought of as a periodic change in frequency, and

ic change in intensity.

with the solo violin note.

a tremulant, which is usually defined as a period- The eflect of the vibrato is enhanced by applying it at different frequencies to different notes, such that when a chord is played no definite rhythmic beating is perceptible. It may be stated in a general way, that when the listener becomes conscious of such rhythm to the extent that he is aware of the beat itself rather than of the effect' of warmth which it gives to the tone, the effect is then unsatisfactory to many listeners. He is never able to count the vibrato effects of the piano and has therefore come to think of the piano as an instrument without vibrato, which is not the fact in a technical sense. Very strong vibrato effects are present in the standard piano, but they are of a very complex nature. The beats between the different frequencies do not last long because the notes decay, and there are a great many different rates of beat going on at the same time. To count them it is necessary to focus the attention on one particular beat and to start to count. Time does not permit muchin this process, because before the rate is established clearly in the mind, the effect is gone.

In the instrument of my invention, I have a system for continuously shifting the frequencies of each note in a rhythmical manner. Thus, if a single note is played, the effect may be counted, if the attention is attracted to it, as is the case I provide a method, however, for putting a different vibrato rate on the different notes so that when chords are played there are many vibratos of different rates going on at the same time. This produces an effect-too complex for the listener to analyze menta1lyin the sense that he is unable to pick up these complex rhythms before they are gone. The shifting of the frequencies of different notes in a manner in which one is independent of another produces a musical effect which is highly desirable. The complex tone of a chord, as in the standard piano, is then made up of many groups of frequencies which are shifting in audibility, so that some are growing more prominent while others are becoming less so. The specific means for accomplishing this result is not shown herein but is disclosed in detail and claimed in my copending application Serial No. 199,612, filed April 2, 1938.

Other and more specific objects of my invention will be apparent from the following description, reference being had to the accompanying drawings in which: a

Figures 1 and '2 together constitute a wiring diagram of a representative portion of the instrument.

Figures 3 to 9a inclusive show wave forms, illustrating the operation of the distortion and control tube.

Figure lii is a graph showing the effect of the velocity of depression of the keys of the instrument.

Figure 11 is a wiring diagram showing the adjustable potential supply means for the circuits shown in Figures 1 and 2.

Figure 12 is a graph showing the relative energy in the fundamental and the harmonics of a piano tone.

Figure 13 is a transverse sectional view of the keyboard and associated switches.

Figures 14 to 18 inclusive are sectional views of the keyboard taken on the lines H-H, iii-l5, i8-l8, l'I--Ii and l8--i8 respectively, of Figure 13.

Figure 19 is a chart showing the harmonic development of a number of notes.

Figure 19a is a wiring diagram showing the setup used to obtain thereadings charted in Figure 19.

Figure 20 is a chart showing the effect of increased key velocity upon the harmonic development of the note.

Figure 20a is a diagram showing the setup utilized in obtaining the readings charted in Figure 20.

Figures 21 to 27 inclusive are various intensity envelope diagrams showing a number of different ways in which the intensity of the note sounded may be varied.

Figure 28 is a diagrammatic representation of a modified form of note control switch mechanism.

Figure 29 is a diagrammatic front elevational view of the unitary bus contact bar; and

Figure 30 is a view similar to Fig. 29 showing a sectionalized bus contact bar.

General plan of instrument I believe that a better understanding of the invention will be obtained if the general operation of the instrument as a whole is first outlined. As a complete instrument it includes thousands of parts, including at least 180 vacuum tubes with their circuits, power supply systems, amplifiers, etc. The whole instrument may be divided into five general systems as follows: (1) Generator system; (2) Control tube system; (3) Touch responsive keyboard; (4) Touch control system; and, (5) Output system.

Generator system The generator system includes 88 alternating .current generators which are continuously operating to produce frequencies of a tempered musical scale corresponding to the 88 notes of the standard piano. The generator system per se forms no part of the invention hereinafter described, being the subject of my aforesaid copending application, Serial No. 199,612, filed April 2, 1938.

While any one of a number of suit-able high audio-frequency generators known in the art could be used, but for a numberof considerations which will appear later, I prefer to employ 88 oscillating vacuum tubes. In one of the arrangements which I prefer, there are 12 vacuum tube oscillators which I refer to as master oscillators, and which oscillate at the frequencies of the 12 highest notes of the piano. In addition to this, there are '76 tubes, only one of which is illustrated in Figure l as tube 280. These tubes are special 3 element gas filled tubes and they operate as relaxation oscillators. The grid of each such tube receives a signal from a secondary winding 2", of a transformer, the primary of which is in the plate circuit of a similar tube one octave up. Thus, the frequency at which each gas tube "relaxes" is controlled, in cascade fashion, from a frequency twice as great, produced by another similar relaxation oscillator which is supplying the signal for a note one octave higher. By means which it is unnecessary here to describe, the frequency of every gas tube oscillator for similar notes in each octave, such as the A, for instance, is thus derived from the frequency of the master oscillator for the note A of the highest octave of the instrument. If this last frequency be shifted up and down through a smallrange, or even several notes, the pitch of every other A for the entire instrument will shift accordingly. The generator system also includes (but not shown in this application) a selectively operable system for periodically shifting the frequencies oi all the master oscillators, at different rates, so that the control tubes 2 hereinafter described do not necessarily function at steady frequencies.

Control tube system The control tube system includes 88 vacuum tubes, one for each note, for which purpose I prefer to use pentodes with sharp cutoff characteristics. Only one of these tubes is illustrated as tube 294 in Figure 1, and its operation is described in detail elsewhere. Briefly, this tube is continuously receiving a signal from the corresponding relaxation oscillator for that note, this signal being applied to the control grid "2. All of these control tubes are connected to transmit the signal from their plates P-IB, P--I0, etc., into a common output circuit described hereinafter. They do not normally supply any signal, however, because of the fact that the cathode of each tube is connected to the common ground oi the system by means of the condenser C4, and unless the cathode receives direct current through a resistance R1, it is of necessity cut off, and no plate current is available to carry a signal to the output circuit. This means that if any direct current flows through the resistance R1, such current will be immediately "chopped up" into impulses which pass into the output system. as a result of the continuously varying potential of the grid 292. The amplitude of these impulses will depend wholly on the amount of direct current supplied to the cathode from the touch responsive keyboard.

Touch responsive keyboard contacts occurring in regular succession in accordance with the motion of the key, so that the electrical events occurring in the circuits associated with these switches take place in succession with a rapidity which is wholly determined by the velooity of the key on its downward stroke. One of the switches shown as switch 262, 286 of Figure l, is adapted to be operated upon by a spring catch 2' which is controlled by a foot pedal 210 corresponding to the damper or sustaining pedal of the standard piano. The general purpose of the switches, operated by the keys and sustaining pedal, together with their associated circuits, is to supply to the control tube system impulses of direct current which are to be convertedinto signals by the operation of the control tubes 294 Just described.

Thus, the direct current supplied from the keyboard switches to the control tubes 284 may be thought of as determining the envelopes of the waves transmitted by the control tubes to the output systems Thus, if the direct current supplied by the keyboard system be one which starts suddenly, continues at the same rate for a period of time and then suddenly stops, the signal transmitted from the control tube to the output system will resemble a sustained note, such as an organ note, for example. If, on the other hand,

the direct current supplied by the keyboard system starts suddenly and then immediately begins decreasing in a logarithmic fashion, the note produced will resemble that of a percussion instrument, such as a piano. Thus, the switches operated by the keys merely control direct currents and it would therefore be possible, if it were desired, to place the keyboard system at a distance from the rest of the apparatus, in the same manner that organ consoles may be placed remotely from the body of the instrument. Except for a common ground return, only one wire per note would be necessary for this purpose.

Touch control system Unlike any other musical instrument with which I am familiar, the instrument of my invention contains means whereby the player may instantly control the touch responsive behavior of the keyboard upon which he is playing. This touch control system includes a number of separate controls described in detail elsewhere, whereby the velocities of key motion may produce different effects not only in the control of the envelope of intensity, but also in the quality of each note. This is accomplished in a general way by shifting the direct current potential of different points in circuits associated with the key and sustaining pedal operated switches.

Output system This system includes circuits whereby the output of all the control tubes can be combined, modified, controlled as to intensity, amplified, and transmitted to an electrical output system for conversion into sound waves in a manner generally similar to that employed in electric organs, but with such modifications as are described in detail hereinafter.

Key action and switches The keyboard employed in the instrument of my invention will, in external appearances, resemble that of the standard piano and extends throughout the tonal range customary in standard pianos, namely, throughout 88 notes commencing with Ar and extending to C For convenience, the notes of the piano are herein designated by their respective numbers, 1 to 88.

While the key action per se does not form a part of the subject matter claimed herein, a brief description thereof is believed to be necessary for an understanding of the invention as a whole, and I have therefore illustrated in Figures 13 to 18 inclusive, 29 and 30, the more important parts of the key action, especially as it is related to the subject matter of the invention claimed herein.

For each .of thje notes there is a key 200, preferably of a molded plastic composition of the general character disclosed and claimed in my co-pending application, Serial No. 91,283 filed July 18, 1936, which has matured into Patent No. 2,117,002, granted May 10, 1938. The keys 200 are suitably secured to key bars 202. which are pivoted upon a rod 204, the latter being suitably secured in a keyboard assembly and being supported adjacent each group of keys by the upwardly projecting portions 206 of key frame plates 208. There is riveted to each of the key bars 202, adjacent its rearward end, an off-set downwardly extending arm 2), to the lower end of which one end of a tension spring 2|2 is secured. The other end of the tension spring 2I2 is anchored to bar 2 extending lengthwise of the keyboard and provided with adjusting screws 2 l6 by means of which the bar 2 ll may be moved forwardly or rearwardly in guide slots 2 ll formed in frame plates 208 'to vary the tension of the springs 2|2 and thus adjust the keyboard as a whole, or any portion thereof, for any desired key tension or "touch.

Upward movement of the key is limited by a thin flexible metallic strip 2l8, the lower end of which is suitably anchored by means of clamping strips 220, and the upper end of which is secured to the forward end of the key bar 202. Upon depression of the key, the strip 2|8 bows forwardly as indicated by the dotted lines in Figure 13.

Suitable means is provided to prevent rebound of the key as it returns to its uppermost position. This means comprises a snubber consisting of a fabric ribbon 222, one end of which is clamped to the center of the strip 218, and the other end of which is shown as secured to the upper free end of. a leaf spring 224. In actual practice the latter may be secured to a suitable adjustable clamping means 225 whereby its effective length may be adjusted to maintain a slight initial bow in the strip 2 IS. The ribbon 222 passes partially around snubber rods 228, 228. As the key is depressed to its lowermost position, and the strip 2l8 bows forwardly to its dotted line position, the spring 224 will take up the slack produced in the ribbon 222 so that upon the return stroke of the key, the ribbon will be under tension applied by the spring 224, and due to its frictional contact with the snubber rods 226, 228 will dissipate energy at a rate increasing as the key approaches the upper limit of its stroke and as the strip 2l8 resumes its normal position as shown in full lines in Figure 13. As a result, the key will return to its normal position without any rebound whatsoever despite the fact that the tension of the spring 2&2 is sufficiently great relative to the rotary moment of inertia of the key bar and parts carried thereby, that the return stroke of the key is very rapid.

Downward movement of the key is limited by a felt pad 230, which is suitably mounted upon a cross-bar 232 extending longitudinally of the keyboard, which has notches 233 receiving portions of the frame plates 20%, being held in engagement with the latter by screws 234. Suitable means for guiding the forward ends of the key bars 202, and for holding the plates 208 in properly spaced relation, are provided, but it is believed to be unnecessary to describe these parts in detail in this application.

Each of thekeys is provided with a suitable bearing bracket' 236 which is secured to the key bar 202 and at its lower end is provided with a pivoted touch-response switch arm 238. arm pivots freely but is forced to move clockwise (Fig. 13) by a tension spring 240, one end of which is suitably anchored to a terminal part 24! fixed to the key bar 202 (but insulated therefrom), and the other end of which is secured to the tail portion 242 of the switch arm 238. When the key is in its normal upper position, the switch arm 23!! makes contact with a bus contact 244, which may extend the full length of the keyboard as shown in Fig. 29 or may, if desired, be made in separate sections such as the sections 244a, 244b, and 244c shown in Fig. 30, each section for one or more keys. Adjacent the bus contact 2 is a contact 246 individual to each key, while forwardly of the. latter there is an insulating strip 248, and forwardly of the insulating strip is a thirdcontact 250 individual to each This key. The upper edges of the contacts 244, 266, 266 and insulating strip 246, are at different elevations, so that upon depression of the key 266, the contact arm 266 will successively make contact with the contact 246; break contact with the contact 244; break contact with the contact 246; and make contact with the contact 266. Upon the return stroke of the key, these contacts are made and broken by the contact arm 266 in the reverse order.

The "repeat" switch comprises a pivoted arm 262 carried by a bracket 264 secured to the key bar 262. The repeat arm 262 does not pivot freely, but a certain amount of frictionis provided so that the arm will not be moved by the forces of gravity and momentum to which it may be subiected. The am 262 projects through an aperture 266 formed in an insulating strip 266, and is adapted shortly prior to the completion of the downward stroke of the key to make contact with a contact 266, the contact with this contact 266 being broken Just prior to the completion of the down stroke of the key and just prior to the time rocker switch arm 236 makes contact with contact 266. Upon the return stroke of the key the switch arm 262 does not make contact with the bus 266.

A sustaining switch arm 262 is suitably clamped at 264' and is in the form of a leaf spring having a contact making element 266 at its rear free end which is adapted to close a circuit by contacting with a contact member 266. In normal position, as shown in Figure 13, the circuit is completed between the switch contact arm 262 and contact member 266. Upon depression of the key, an actuator 266 which is carried by the key bar 262, is adapted to engage the contact am 262 and move the rearward free end of the latter away from the contact 266, thereby opening the circuit.

A shaft 266 is pivotally mounted in the frame plates 266 (other suitable bearings may be providcd) and may extend the full length of the keyboard, or may be made in two or more sections. The shaft 266 is adapted to be moved clockwise (Fig. 13) by means of a sustaining pedal 216 (illustrated schematically in Fig. 1), the sustaining pedal being connected by a suitable tension member 21l with an arm 212 secured to the shaft 269. If the shaft is made in two or more sections, each section may be provided with a separate sustaining pedal and linkage for actuating that section of the shaft, duplicating the parts 269, 216, 2H and 212 shown in Fig. 1. For example, the instrument may be provided with a shaft section 269 which extends throughout the treble register of the keyboard and a similar shaft section for the remaining keys of the keyboard, each section being controlled by a separate sustaining pedal. The sustaining pedals, however, are preferably so positioned that the player may easily with one foot selectively depress either of the sustaining pedals, or both of them.

The shaft, or shaft sections, 269 have secured thereto resilient latches 214 which, when the shaft 269 is swung clockwise by the depression of the sustaining pedal, will press against the ends of the sustaining switch arms 262. If the latter are in their uppermost position, the pressure of the resilient latches 214 against the ends thereof will not have any effect. If, however, a key is being held down while the sustaining pedal is depressed, the resilient latch 214 associated therewith will swing above the free end of the sustaining switch arm 262 and ho d t i latter away from its complementary contact 266, when later the key is released, as long as the sustaining pedal is held depressed. Similarly, if keys are operated while the sustaining pedal is being held depressed, the sustaining switch arms 262 of such keys will be engaged by their respective resilient latches 214, and these switch arms 262 will thereby be held away from their associated contacts 266 until the sustaining pedal is released and the resilient latches 214 thus swung counterclockwise (Fig. 13) out of the path of movement of the free ends of the sustaining switch arms 262, whereupon these arms will resume their normal positions in contact with contacts 266.

From the above description it will be seen that the depression of a key will result in switching operations in the following sequence:

(0) Making contact between switch arm 266 and contact 246;

(b) Breaking contact between switch arm 236 and bus contact 244;

(c) Breaking contact between switch arm 266 and contact 246.

(11) Making contact between repeat switch arm 262 and contact 266; a

(e) Breaking contact between switch arm 262 and contact 266;

(f) Breaking contact between sustaining switch arm 262 and its complementary contact 266;

(9) Making contact between switch arm 266 and contact 256; and upon the return stroke:

(h) Breaking contact between switch arm 226 and contact 266; I

(i) Closing contact between sustaining switch arm 262 and its complementary contact 266 (if the sustaining pedal is not depressed at the time the key is released) (1) Making contact between switch arm 226 and contact 246;

(Is) Making contact between switch arm 266 and bus contact 244;

(1) Breaking contact between switch arm 226 and contact 246.

The functions and effects obtained by the sequence of switching operations above set forth, will be apparent from the description of the circuits and their operation which appears hereinafter. The key action and associated switches, as distinguished from the combinations in which these elements are used, are disclosed in greater detail and are claimed in my copending application Serial No. 216,336, filed June 28, 1938.

Oscillator generators As previously stated, the oscillators for the various note frequencies may be of any of the well-known types, but I prefer to use vacuum tube oscillators for the twelve highest frequencies and may use gas filled tube relaxation oscillators for the remaining '16 notes. The gas filled tube oscillators have their frequencies controlled by means of the master oscillators, each master oscillator controlling the frequency of the gas tube oscillators which are of the same note in the octave, i. e., whose frequencies bear a submultiple octave relationship to the frequency generated by the master oscillator. The gas tube oscillators of aseries are controlled in cascade in a manner such that the master oscillator controls the frequency of the gas tube oscillator which has a frequency one-half that of the master oscillator, and this gas tube oscillator controls the frequency of the next gas tube oscillator of the series which has a frequency oneiourth of that of the master oscillator, and onehalf the frequency of the preceding gas tube oscillator, etc. This system or method of generating oscillations of the frequencies of the tones of the musical scale is disclosed in greater detail, and is claimed in my copending application Serial No. 199,612, filed April 2, 1938. Another simpler and less costly arrangement is to have the oscillations of the master oscillators successively divided by two by means of the frequency divider circuit disclosed and claimed in my co-pending application Serial No. 199,614,

filed April 2, 1938.

In Fig: 1, I have illustrated one such gas filled tube relaxation oscillator as comprising a gas filled triode 280. The plate circuit of each oscillator includes a resistance 282 and an inductance 284 (which constitutes the primary of a transformer, the secondary of which is connected in the grid circuit of the gas tube oscillator generating a frequency which is an octave lower than that of the oscillator illustrated). a source of adjustable direct current potential 286, a resistance R11 and a resistance R16. A condenser C15 is connected between the source 266 and the cathode of the tube 286. The grid of the tube 280 is connected to ground through a resistance 28! and an inductance 288, the latter constituting the secondary of the transformer by which the stabilizing signal from the oscillator of a frequency twice that of the oscillator illustrated is impressed upon the grid of the tube 280. The resistance R17 is shunted by a condenser i314 and may be shunted by an additional condenser C10 upon closure of a switch 2%. The signal from this relaxation oscillator is taken from the point P between the resistances Rm and, Ru and is impressed upon the control grid 292 of the control tube 294. The heaters for the cathodes of tubes 2% and 2% are omitted from Fig. l for the sake of clarity. A description of the operation of this oscillator will appear hereinafter.

Control tube and circuits The functions of the control tube are fourfold: (a) It serves as a distorter tube to change the harmonic content of the note fed to the output circuit from that oi the signal which is applied to its grid; (b) it serves to control the envelope of the energy output, beginning-at cutoff, when the note is not sounding, to maximum intensity. and progressively to lower. intensi ies. finally to cutoff again; it serves as a 1 means for changing the harmonic content oi the note in a manner which results in increasing the brightness with increasing output inten ity: (d) lastly. the tube operates in such a way that it makes possible a wide variety oi tone qualities supplied to the output circuit by a change in the value of a condenser connecting its control grid to ground. A change in the value of this condenser changes the sharpness of the positive peak of the signal impressed on the grid. and this change produces a change in the harmonic content of the output signal in a manner which is entirely different from the effect a condenser would have in a linear circuit.

The cathode of the control tube 294, of which there is one for each note of the instrument. is.

connected to a terminal N which is connected to the ground through a blocking condenser C A high resistance R1 is connected between a junction M and the terminal N. The junction M is connected to ground through a blocking condenser C1. The junction M may be connected to a direct current source of adjustable potential through a resistance R14 upon closure of a switch 296 which isconnected to terminal G. The junction M is connected to a terminal D through a resistance Re, the terminal D being connected to a direct current source of adjustable potential.

The junction M is connected through a resistance Ra with a terminal L, which latter is connected to the fixed repeat switch contact 260. The repeat switcharm 252 is connected by a conductor 298 with a terminal F, which in turn is connected to a direct current source oi adjustable potential. A resistance Rm is connected between terminal L and a terminal K, which is the terminal on the sustaining switch contact 266. The resilient sustaining switcharm 262 is connected to a terminal E, which in turn is connected to adirect current source of adjustable potential. A resistance R0 is connected between the junction M and the terminal J on the contact 250. i

The contact 246 is connected through a resistance R13 to terminal B, which in turn is connected to a direct current source of adjustable potential. The bus contact 244 is connected to a terminal A, which is connected to a direct current source of adjustable potential. The touch response switch arm 236 is connected to a terminal H, the latter being connected through a resistance R12 to a terminal 300, and the latter terminal is connected to ground through a condenser Cs. and connected to terminal C through a high resistance Ru, the terminal 0 being connected to a direct current source of adjustable potential.

The screen grid 3% of the control tube 294 is connected to a source of direct current of adjustable potential, while the suppressor grid QM may be grounded. The plate 79-49 of the tube 94 is preferably connected to a wire 306 in parallel with the plates E -d, P-Ell and P--2 which are the plates of the control tubes 294 for the adjacent notes Nos. 50, 51 and 52.

Operation of control tube and circuits I will first describe the operation or the tube 2% with r'espect to its control of the energy output. When playing the instrument as a piano this energy output is of course changing all the time from the beginning of the note when it rises suddenly, until the end, when the tube is again cut oil by release of the key or release of the sustaining pedal, in the event the latter has been used. The tube I prefer is a pentode having sharp grid cutoff characteristics. A tube oi. the type 666, for instance, is satisfactory. The control grid 293 receives a continuous signal from a generator, which in this case is illustrated as a continuously oscillating vacuum tube. One of the systems which I prefer, and which is herein partly shown and described, is the three-electrode gas filled tube 280 previously mentioned, but which'might be some other form of oscillator or generator.

The screen 302 of the control tube 294, like the screen of the control tube for every other note, is supplied with a steady potential of 25 volts. The suppressor 304 is connected to ground and the plate is continuously supplied with voltage, (at point 3i2, Fig. 2) in this case some 200 volts, from an output circuit which receives the signal from all the control tubes. Since the cathode is connected by means of the condenser C4 to ground and by means of the high resistance R1 to ,the touch responsive key circuit, no direct current can flow to the cathode of the control tube unless it does so by way of the high resistance R1. Thus, if the control tube 294 is not supplied with direct current from the key circuit, it cannot continue to function to transmit the signal (which is continuously put on its grid 292) to the output circuit. when direct current stops flowing in R1 the potential of the cathode will shift in the positive direction such that the control grid becomes more negative with respect thereto, until the tube is cut off. In this condition there is no output signal despite the fact that the grid continues to receive an input signal, and the plate P-49 is shielded from the capacltative effect of the grid by the screen I02 and suppressor I04. The key circuit, which is described elsewhere herein, controlslthe direct current component of the pulsating current in the cathode circuit. The impedance of the tube 204 from cathode to ground for direct current is extremely small by comparison with the impedance of the resistance R1. This is due to the fact that if the potential of the point N with respect to ground should change by a very small amount in the negative direction, an enormous plate current would immediately now through the tube.

The control tube may be thought of as normally cut off. As soon as cathode current is supplied from the key circuit, the direct current po-- tential of the cathode will shift slightly such that the next most positive excursion of grid potential will be just sufllcient to allow an impulse of current to pass through the tube during a fraction of the cycle of grid voltage. During the remalnder of the cycle the tube will be cut off. and the potential of the cathode will again be sinking in the negative direction for another impulse of current on the succeeding positive grid swing.

To understand the operation, it may be convenient to think of the tube as a device whereby whatever direct current is supplied by the resistance R1 to the cathode, will be immediately "chopped off" by the tube and delivered in impulses to the output circuit, the frequency of these impulses being of course determined by the frequency of the signal which is continuously supplied to its grid. If the current supplied by the resistance R1 be double, for instance, then the cathode current will be doubled and so will the average value of plate current. The control tube operates to amplify the signal supplied to its grid and to transmit a signal to the output circuit, but it does not act as a linear amplifier. The harmonic content of the output differs from that of the input, and the tube may therefore be thought of as .a distortion device. It is the particular type of distortion which is here of interest and which may be explained as follows:

In the circuit illustrated in Figure l the signal applied to the grid 202 of the control tube 294 is derived from a generator having an output which is not a sine wave. I may use a sine wave, however, and in the preferred form of construction do use a sine wave for the highest octave of notes. If a sine wave is supplied to the grid, the tube functions for only a portion of the cycle during the period of most positive grid excursion, so that the quality of the output is dependent on the shape of only that portion of the input signal wave which is sufliciently positive to operate the tube.

To illustrate this point, there is shown in Figures 3 to 6 inclusive a number of different input waves which would result in the same output signal, illustrated in Figure 9a. The harmonic content of these various input signals would be entirely different, and yet the output would be the same because the shapes of the upper parts of the waves above the dot-dash cut-off lines, are similar in each case. The distortion introduced by the tube is therefore of a peculiar nature. A sine wave input will produce an output wave which may be rich in harmonics.

If the output wave, such as that shown in Figure 9a, is connected to a harmonic analyzer (using a circuit such as shown in Figure 19a) and the amount of energy of each harmonic is measured separately, it will be found that the amplitude of each harmonic bears a simple relationship to the amplitude of all other harmonics. This relationship is a simple geometric one in which if the amplitude of the fundamental is unity and the amplitude of the second harmonic is, for example, .8, then the amplitude of the third harmonic will be .8 .8 or .64, and the amplitude of the fourth harmonic will be .8 .64 or .512, etc. If the amplitude of each harmonic be plotted on semi-logarithmic paper, a series of points such as that shown in Figure 19 will be obtained, and it will be found that these points lie on a straight line with a high degree of accuracy. If the amplitude of the input wave be changed, a different series of harmonics will be obtained, but in this case the points will again fall in a straight line with a difference in the slope of the line.

The harmonics present from what appears to be an infinite series extending upward in frequency in accordance with this formula as far as the measuring apparatus will permit. In measurements which I have made, the ratio of energy between the 42nd and 43rd harmonics, for example, will be found to be just the same as that between the second and the third harmonic. The complete harmonic analysis of tone is therefore determined by simply specifying the difference in intensity measured in decibels between any harmonic and the next succeeding one. This is of course the same as specifying the slopes of the lines in Figure 19, and I therefore refer to this slope in terms of decibel decrease per harmonic number as, for instance, 2.05 db/H for note No. 25. The curves of Figure 12 are plotted with a value oi db/H as abscissas and various energies as ordinates. For a wave having a fundamental of a given ampltiude it will be found that the total energy increases very rapidly as the harmonic content increases, corresponding to arm-- lyses in which the db/H factor decreases. This factor is therefore a convenient measure of the brightness of the tone.

Except for the highest octave of notes, each of the control tubes 294 obtains its signal from a relaxation oscillator shown as comprising the gas filled triode 280 in Figure i. In the cathode circult of this tube are two resistors R16 of 300,000 ohms, for example, and the resistance R11 of 10,000 ohms. A pulsating direct current flows through these resistors and the wave of this current is of a saw-tooth nature illustrated as the curve shown in Figure '7. The voltage developed across the resistance Rn would be of the same shape were it not for the eiTect of the condenser On, which shuts it. If this condenser be of proper value, the voltage across the mesh of the resistance R17 and the condenser C will be of wave form illustrated in the curve of Figure 8.

It will be noted that the curve of Figure 7 is of saw-tooth form, one leg of which is almost vertical, and there is a sharp corner to the wave on the most positive and negative points. The excursion of voltage in the positive direction does not produce a vertical line because of the impedance in the plate circuit of the tube which does not permit of an instantaneous change in charge of the condenser C15. The slope of this portion of the curve may be controlled by changing the constants in the plate circuit, The curve of Figure 8 shows the effect of the condenser C14 which has the effect of rounding off the most positive portion of the wave. The shape of the most negative part of the wave is substantially unaffected. Comparison of this curve with the curves of Figures 3, 4, 5 and 6 will show that only the most positive part of the wave is of importance in determining the signal put out by the control tube. Thus, a saw-tooth wave may be employed in the same manner as a sine wave, and the harmonic content of the output of the control tube will have the same general characteristics as those described for the sine wave input.

An additional condenser Cm may be shunted across the condenser Cm by means of the switch 290. In a complete instrument this switch is a gang switch, operating to ground one end of a similar condenser provided for each note for which a relaxation oscillator is used as the generator. With the added capacity, Cm, the curve representing the voltage at the point P (with respect to ground) now changes to a shape more like that of curve of Figure 9. In this wave the positive loop is rounded off even more.

Figure 20 shows the result of the analysis of the harmonic content of the signal put out by the control tube of note No. 25 of a complete instrument. The circuit. used and the constants used are shown in the diagram, Figure 20a. Five different series of points were obtained which may be joined by substantially straight lines, representing the voltages of each separate harmonic entering into the composition of the wave. The five diiferentlines arise from five different voltages of the battery B1. These curves show how an increase in this voltage produces an increase in the energy of the output signal, and more particularly, how the harmonic content of the resulting note changes with an increase in intensity.

The slope of each line is denoted by the brightness factor db/H previously described, and it will be noted that the decibel decrease per harmonic number" changes from a value 2.75 for the softest note to 1.41 for the loudest, and that this change is systematic. Further experiments have shown that'the harmonic content of the output may be controlled in two major respects by changing the values- OfQm and C4 respectively. For any given battery voltage the slope of the line, or db/I-I factor may be changed by changing the value of the condenser C14. Once this slope is established for a certain battery voltage, the rate of change of slope with increasing and decreasing battery voltage may be largely controlled by the proper selection of the value of the condenser C4. The desirability of the brightness change to simulate piano-like effects was described previously. I I

In Figure 19 are plotted the harmonic analyses for a plurality of different A-s of the complete instrument. The same battery voltage was used in each case, and was of a potential corresponding to the starting potential of a blow considered to be of medium strength, In this case,

'H and A.

the values of the condensers C14 and C4 for each note are arrived at by voicing" the instrument in such a manner as most closely to approximate the tone quality for a similar blow on a certain concert grand piano of well known make. The curves show the enormous change in the harmonic content for the same note in different octaves. Similar curves plotted for different intensities of blows show the same general arrangement, except, of course, that all notes are more harmonically developed with increasing strength or blow.

Operation of touch responsive key circuits As previously described the touch responsive key circuit is connected as shown in Figure 1, in which, when the key is in its upper position of rest, the rocker arm 238 electrically connects the point H with the terminal A. As the .key goes down, the rocker arm connects H to B through R1: and opens the connection between After a period of time, depending upon the velocity with which the key is moving, the contact from H to B (through resistance R13) is in turn broken and when the key is all the way down, contact is made between H and J.

When the key has been depressed a certain distance the damper switch 262-266 connecting E i to K is likewise opened. When the key is allowed to return upward this switch will again be closed unless it is held open by the action of the sustaining pedal 210 which operates to turn the shaft 269 through a small angle in a clockwise direction. In that event, the damper switch will remain open regardless of whether the key is allowed to rise or not.

The repeat arm 252 whose mechanical operation has been described, connects the terminal L with the terminal F for a brief period during only the down stroke of the key. The circuit from L to-F is open both at the top and at the bottom of the key stroke and the parts are so dimensioned that contact is made on the down stroke at about the same time that the rocker arm 238 is passing over the insulating strip 248, so that at that time, point H is not connected to either A or J.

The terminals A, B, C, D, E, F, and G are, for purposes of illustration, each connected by means of a battery to ground. The switch 296 is herein assumed to be open. These batteries are merely to indicate that each of these points is to be held at a fixed potential relative to ground. The voltshunting this supply with a tapped resistor in a manner well known in the design of power supply systems for radio sets and the like and as described hereinafter with respect to Figure 11. By

means of various multiple switches, hereinafter described in detail, the performer may change the potential -of each one 01' these points independently in order to obtain a different mode of operation for the circuit.

In order toexplain the operation of the key circuits we will assume that the performer wants to obtain a normal touch responsive action which will closely approximate that of the standard piano. In that case, he will make such adjustments that the potential at the various points with respect to ground is as follows: A at minus 450 volts, B, C and F at ground, D and E at plus 6 volts. The points A, B, C, D, E, F, and G are common for all notes of the keyboard, (or, if desired, for only certain registers thereof) but the values of the resistors or condensers, or both, may change' from note to note. Assuming that the diagram applies to note No. 25, for instance, then good values for the various elements are as follows: C1 and Ca, .3 mfd; R1 and R11, 4 megohms; R12 and R13, 6000 ohms; R0 for this note may have a value of zero and consequently be omitted, (but for higher notes Re may be considerable); Ra. 10,000 ohms; Rm, 600,000 ohms; and Re 20 megohms.

I prefer to make the condensers C1 and C1 of the same capacity, and it will be noted that one end of each is connected to ground. The charges on these condensers vary during the playing of the note, the general plan being that a charge on C1; is carried and delivered to C1. A simple mechanical analogy may beused to illustrate the operation as follows: A runner has a bucket withwhich he is carrying water from a lake to flli a tank. The bucket has a hole in it. .On his way between the lake and the tank, the water is continually running out of the bucket. He always starts with a full bucket but when he reaches the tank, he may have much or little, or no water at all, depending on how fast he ran. If he goes slowly enough he will obviously arrive with no, water. In this analogy, the condenser Cu is the bucket. Condenser C1 is the tank. The hole in the bucket is the resistance path from B through the resistance Ru, the rocker arm to H, the resistance R11 to the bucket condenser Ca. The lake is the source A, the potential of which is 450 volts more negative than ground.

With the key up before the down-stroke of the key, the potential across Cu is that of the source A, and the condenser therefore has a high negative charge. The potential across C1 is the same as that of the points D and E which were both assumed to be at the same voltage, namely, plus 6, with respect to ground. No current is flowing in the resistance R1 because the control tube 284 is cut off.

Assume that the key is struck very violently. In that case, connection through the rocker arm 238 from H to B will not last long, and despite the fact that B has been assumed to be at ground potential, time will not permit a great discharge of the "bucket" condenser C; because of the resistances R1: and R1: in the circuit. As a result, when the connection H to J is established through the rocker arm 238, the charge on the condenser Ca will be something less than 450 volts, say for example 250 volts, depending of course on how hard the key was struck. As soon as the rocker arm 238 establishes contact between H and J, the two charges on C1 and Ca will begin to equalize. This takes only a very short time because of the low resistance of Ru and Re, and as a result, the potential at the point M will suddenly shift from its previous value of a low potential with respect to ground to a potential about halfway towards lninus 250 volts, or say minus volts.

Had the key been pressed very slowly the potential delivered by Ca would be substantially nothing. That is to say, its charge would be substantially lost and when the connection H to J was established the condensers C: and C1 would neither of them have any substantial charge. In that event, the potential at the junction point M would be left substantially unchanged.

In Figure 10, a curve has been plotted in which the abscissa is the velocity of key motion. -The ordinates are the voltages of the point M at the start of the note. These voltages are not obtained, of course, until, after a very short period of time when the two condensers Ca and C1 have had their charges equalized 'through the path from M through Re to J to H and through Rm.

It will be noticed that if the velocity of the key is near zero, the voltage of point M will likewise be close to zero. As the velocity increases to 1, voltage begins to rise slowly, the curve being shown as the curve A-IO.

The characteristics of the control tube and the constants of the circuit are such that the tube is cut oi! when the cathode is at a voltage of plus 6 volts with respect to ground. Accordingly, if the cathode is more positive than plus 6 volts, there is no direct current path from the point N through the resistance R1 to ground. Likewise, the point N always remains at a potential which is exceedingly close to plus 6 volts with respect to ground, because as soon as any current flows through R1 tending to make the point N grow more negative, the current through the tube will increase very sharply. For operation of the key circuit it may therefore be assumed that the point N is always at a fixed potential of plus 6 volts with ,the cathode current of the tube and there is,

therefore, a signal,.albeit a small one, with M at the potential of zero.

From the curve in Figure 10, it may therefore be seen that if the velocity of the key is anything less than 1, the intensity of the signal will be substantially constant and may be represented by the voltage between the cutoff voltage of the tube and the voltage of the point M. A straight line F-Jfl drawn from plus 6 volts and zero velocity toward the point B-l0 will therefore represent a theoretical value which the voltage of point M should have if the intensity of the signal, expressed in volts, was exactly proportional to the key velocity. This relationship will be exactly true at three points where the straight line intersects the curve 11-", namely, at C, D and E. The agreement, generally speaking, is very close except for very low velocities, which gives this system a great advantage over the touch responsiveness of the standard piano, for reasons previously described.

The constants of the circuit, especially the values of the resistors R1: and R12 must be my chosen that the hardest blow which a good pianist can deliver to the key will not carry the voltage of the point M a great deal beyond the point and the point M has reached a negative potential with respect to ground, of 125 volts There are now three paths whereby the charge on the condensers C1 and Ca may leak off. One path is through theresistance R1 to the point N, the potential of which is plus 6 volts. Another path is through the resistance R11 to the point C, the

potential of which is zero, and the third is through

Images