US3117304A - Matrix with reflected wave energy crosspoints - Google Patents

Description

Jan. 7, 1964 l. G. AKMENKALNS 3, 7,

MATRIX WITH REFLECTED WAVE ENERGY CROSSPOINTS Filed Nov. 2. 1960 s Sheets-Sheet 1 9 7K"'" '"""""""E j7 j 9 x L 5 OPEN CIRCUIT ELEMENT TO BE SELECTED TERMINATION {J3EE PFL'SE V F E- A TIME l REFLECTED I 151 PULSE OVERLAPS V ZndPULSE 1 i TIME 7 L LO Zo 3 3 9 8 OPEN CIRCUIT l %Zo=RL=ELEMENT TO BE TERMINATION SELECTED FIG. 10

INVENTOR IVARS G. AKMENKALNS ATTORNEY MATRIX WITH REFLECTED WAVE ENERGY CROSSPOINTS Filed Nov. 2, 1960 1964 l. G. AKMENKALNS 5 Sheets-Sheet 2 :ELEMENT TO BE SELECTED T FIG. 2

V? Zo RFELEMENT vTo BE SELECTED FIG. 20

l DRIVER Row 1 Jan. 7, 1964 l. G. AKMENKALNS MATRIX WITH REFLECTED WAVE ENERGYCROSSPOINTS Filed Nov. 2, 1960 "5 Sheets-Sheet 3 COLUMN 0 DRIVER J W3 1 a DRIVER *4 PLANE United States Patent M 3,117,304 MATRIX WITH REFLECTED WAVE ENERGY CROSSPOINTS Ivars G. Akmenkalns, Endicott, N.Y., assignor to International Business Machines Corporation, New York,

N.Y., a corporation of New York Filed Nov. 2, 1%0, Ser. No. 66,819 8 Claims. (til. 340-166) This invention relates to matrices and more particularly to an improved means of selecting a particular element in a matrix.

In the prior art, two coordinate matrices are commonly employed; such matrices also generally comprise elements having at least two stable states of energization. The elements may be considered to be arranged in rows and columns and a plurality of drive lines pass through the rows and columns of elements, that is, the drive lines in the rows cross the drive lines in the columns and a binary element is located at each intersection. To energize any one of the elements in a matrix, it is necessary to coincidentally energize a drive line in one of the rows and a drive line in one of the columns. The magnitude of the current at the intersection of the drive lines which are energized is suflicient to drive the element located at said intersection to one or the others of its stable states.

It is a principal object of this invention to provide an improved means for selecting a particular element in a matrix.

It is another object of this invention to provide a means of selecting an element other than by utilizing coincidence or anticoincidence current flowing in associated drive lines.

It is another object of this invention to provide a matrix in which the dimension of time is utilized to select a particular binary element.

It is another object of this invention to provide a matrix in which a single set of lines is utilized to drive the elements.

It is still another object of this invention to provide a matrix in which a minimum of current drivers are utilized.

It is yet another object of this invention to provide a compact extremely fast acting scheme for selecting an element of a matrix.

In one preferred embodiment, the invention provides a matrix comprising a plurality of binary elements arranged in a plurality of groups; for simplicity in description, the elements are considered to be arranged in rows. Each of the elements is connected to a distinct position in each of the drive lines. One end of each of the drive lines is connected to a pulse generating means and the other end of each of the drive lines is terminated such as to reflect pulses sent down a line by the generator. An element is selected by actuating a pulse generating means to send a pulse down the respective drive line and then actuating the pulse generating means to send a second pulse down the drive line to meet the reflected first pulse at a desired point. The reflected first pulse will reinforce the second pulse and total energy will be of suflicient amplitude to energize the binary element positioned at the selected point.

The foregoing and other objects, features and advantages of the invention will be apparent from the following more particular description of preferred embodiments of the invention, as illustrated in the accompanying drawings.

In the drawings:

FIGS. 1 and 1a are schematic drawings of an open circuited transmission line useful in describing the concept of the invention;

FIGS. 2 and 2a are schematic drawings of a short cir- 3,1113634 Patented Jan. 7, 1964 cuited transmission line also useful in describing the concept of the invention;

FIG. 3 is a schematic diagram of a drive line comprising a microwave strip line to which are connected a plurality of spaced elements in accordance with the invention;

FIG. 4 is a graph showing the operation of the binary elements comprising Esaki diodes of the circuit of FIG. 1;

FIG. 5 shows a plurality of microwave strip lines as shown in FIG. 4 arranged in a matrix.

FIG. 6 shows a schematic diagram of a word memory array in accordance with the invention.

Before describing the structure and operation of circuits constructed in accordance with the invention, an over-all principle of the properties of transmission lines will be reviewed.

Referring to FIG. 1, if an ideal lossless transmission line 7 is properly terminated at the driver 8 end, and if the other or receiving end 9 of the line is open circuited, no energy can be dissipated at the receiving end. Therefore, a pulse generated at the driver end will propagate down the line and upon reaching the receiving end will be reflected back toward the driver end. Upon reaching the driver end of the transmission line, the energy will be dissipated in the characteristic impedance Z0 and driver internal resistance. If the transmission line is short circuited at the receiving end, again no energy can be dissipated in the termination and the energy will be reflected and finally absorbed in the generator resistance.

Assume that two energy pulses with some time interval therebetween are generated by the driver and propagated down the line. The first pulse, after being reflected from the termination, will meet the second pulse at some point on the transmission line. If the two pulses have finite pulse widths, they will overlap over some finite length or region of the transmission line. Over this finite length of line, the energy per unit length of line will be doubled due to coincidence of the two energy pulses.

Consider now the voltage and current distributions on the line when operated in the foregoing manner. An open ended termination, as shown in FIG. 1, imposes the requirement on the energy pulse that the current at the termination has to be zero and for this reason the reflected current will have an opposite sign from the incident current. As the two pulses overlap over some part of the line, the two current components will cancel and all the energy will be stored in the electric field. At this location, the voltage on the line will double. If a relatively large resistance RL is connected across the line at some point, point X in FIG. 1, in the region of pulse overlap, a voltage or a current proportional to the line voltage may be obtained for control of an element 5 (device) having at least two stable states. Since a relatively large resistance is used, a minimum of disturbance is presented by the load on the line. Also, a properly designed matching network, as known in the art and as shown in FIG. 1a, may be used to couple the load RL to the line.

The region of the transmission line over which the pulse overlap occurs is controlled by the pulse width and spacing of the pulses, as will be described hereinbelow.

In the case of a shorted termination. a voltage of opposite polarity to the incident voltage will be reflected and cancellation of voltages will take place somewhere on the transmission line. In this region, the entire energy will be stored in the magnetic field and the current magnitude will be doubled. A magnetic field activated device may be controlled by the field if placed in series with the transmission line, as shown in FIG. 2 at the region where the current magnitude is doubled. Again, the circuit of the element 6 (device) should present a minimum series impedance; a properly designed matching network, as known in the art and as shown in FIG. 2a may be used to minimize any disturbance presented by the load to the line.

Utilizing the foregoing principles, circuits according to the invention have been constructed to selectively control elements, for example, binary elements or devices, by using only one set of drive lines, as will hereinafter be described in some detail.

Referring to FIG. 3, a group of similar binary elements or devices comprising Esaki diodes 11, 12, 13 and 14 are connected through similar respective resistors, indicated generally as R, to spaced points on a drive line comprising a microwave strip line 19 commonly referred to as a microstrip line. A voltage pulse generating means or driver 24 is connected through a suitable line terminating (matching) resistor 25 across one end of the micro strip line 19. The other end of the microstrip line 19 is open circuited.

An article in the Physical Review of January, 1957, on pp. 603-605 entitled New Phenomenon in Narrow Germanium P-N Junctions by Leo Esaki, describes a semi conductor structure now known as the Esaki diode or tunnel diode. The Esaki diode may be said to be a P-N junction diode wherein both the P region and the N region contain a very high concentration of the respective impurities resulting in a current versus voltage characteristics which exhibit a short circuit stable negative resistance region. As seen from FIG. 4 which shows the current versus voltage characteristics, the Esaki diode may also be defined as exhibiting a first region of positive resistance over a low range of potentials and adjoining at a peak current value, a second region of negative resistance, and then a third region of positive resistance. Thus, by properly providing potentials to the Esaki diode, a bistable element may be obtained. The Esaki diode is particularly suitable for circuits formed in accordance with the invention since its response time is in the order of millimicroseconds and its operating potentials are relatively low in the order of 0.05 volt at the beginning of the negative resistance region to 0.4l.2 volts, depending on the diode material used at the end of the negative resistance region.

In FIG. 3, the Esaki diodes 11-14 are each biased for bistable operation by a voltage +E coupled thereto through a resistor r. Diodes 11-14 are coupled to the microstrip line 19 through the resistor R having a resistance which is large compared to the characteristic impedance of the line. A capacitor and resistor series network instead of resistor R can be used for coupling, if the direct current fed back by the diode circuit creates problems. A conventional three-way matching network may also be used to couple the diode circuit to the line. Driver 24 is a bipolar voltage generator to provide opposite polarity pulse to thus effect set and reset states to the bistable Esalti diodes. The outputs of the Esaki diodes may be connected to an OR circuit of any suitable known type such as backward diode and capacitance type for output sensing.

The microwave strip line 19 comprises a solid dielectric microwave strip conductor 21, a dielectric 22, and a ground plane 23. The characteristics of solid-dielectric strip lines are known in the art and have been described in numerous publications, for example, in one place in the March 1959 issue of The Proceedings of the Institution of Electrical Engineers in an article entitled Parallel-Plate Transmission System for Microwave Frequencies by A. F. Harvey. It has been found that a pulse inserted at one end of a microwave strip line is propagated down the line and the velocity of propagation is determined by the properties of the dielectric material which separates the strip conductor 21 and the ground plane 23. However, as the velocity of propagation (V) is decreased, the attenuation which is due to the conductor, dielectric and radiation losses is increased.

The relationship between the line constants, signal pulse width, pulse spacing and spacing between elements on the line is derived hereinbelow. For simplicity, assume rectangular pulses are employed.

A pulse of pulse width w, when propagating through a transmission line with a velocity of propagation v, will have its energy contained in a region of d units length along the transmission line.

The following relation is obtained:

d v.w

In coaxial transmission lines:

c=velocity of light e relative dielectric constant where Therefore, the region at that the pulse will occupy on the line at any instant may be expressed as:

The region d corresponding to the maximum overlap of the pulses will be equal to the length of region d of the shorter of the two pulses. This then establishes the minimum distance s between any two devices connected to the transmission line. Thus:

In terms of the line constants, this relationship can be expressed as:

the wider pulse may have a maximum width of For a given spacing s between devices, the minimum pulse width is determined by the minimum time of pulse overlap required in order to control the binary device.

The selection of a particular element on the line is determined by the time interval between the leading edges of the two pulses at the driver end of the line. Let us consider element X located a units of distance from the driver end of the line and b units of distance from the receiving end of the line. The leading edge of the first pulse will be sent out at time t=0 and upon reflection will reach the point at which element X is located, T seconds later. The following relation applies:

T= a+2b a+2b fi In order for the leading edge of the second pulse to reach X at time T, the leading edge of the second pulse must leave the generator at time:

and therefore,

1,: a+2b 1 E-a-lE=2b-l E c c c In other words, the time interval between the two leading edges of the energy pulses has to be equal to the time required for a pulse to travel twice the distance from the desired selection point to the receiving end of the line. This relationship then determines the selection of the particular region on the line where the pulse overlap will occur. Note that this discussion applies to voltage overlap in case of an open ended line and current overlap in case of a shorted line.

Consideration must also be given to the distances between the first element and the driver and the last element and the receiving end of the line. It is obvious that the distance from the driver to the first element requires no special considerations. The distance from the last unit to the receiving end, however, has to be greater than onehalf of the distance of the line occupied by the wider pulse, otherwise, overlap of each pulse with itself as it is reflected may cause an error in the control function.

The time interval between two selection and control operations depends on the time it takes to reflect the sec ond pulse and absorb said pulse .at the receiving end. The longest time interval will be required for the selection of the first unit.

An appreciation of the magnitudes involved may be obtained from the following example. Assume that a 5 bit storage register using high speed Esaki diodes is re quired. Also, assume that a glass dielectric line (e=9) is used that a l millimicrosecond voltage signal will provide the required control.

The minimum required overlap is therefore 1 millimicrosecond. Assuming that the pulse width w+10% tolerance may be maintained on the drive pulse, the maximum pulse Width of the shortest pulse will be 1.2 millimicroseconds.

The minimum distance between two elements is therefore,

Assume there is a 6-inch spacing between devices. Arbitrarily selecting a space of 6 inches between the driver and the first element, and a three inch spacing between the last element and the end of the line, length l of the register is obtained as: (note that the spacing between the last element and the end of the line is only 0.5 of the spacing between the elements themselves) l=5.5s:5.5 6=33 inches In order to select the fourth device in the register, the two pulse leading edges have to be spaced in time by an interval 1.2X 10 392 ft.=4.7 inches Now, assume that the second pulse has its maximum pulse width of 2 millimicroseconds and the first element is being selected. The reflected trailing edge of the second pulse will arrive at the driving end after T seconds 1 2b 2X .75 =46 millimicroseconds 30.6 millimicroseeonds Consequently, a minimum of 30.6 millimicroseconds is required between two successive control operations.

Thus, in the foregoing examples, it the drive pulse width is 1 millimicrosecond and the total length of the line is 22 inches, the first and third pulses must be separated in time by at least 20.4 millimicroseconds in order to selectively energize the diodes. The first and second pulses are considered as one control operation, as are each subsequent pair of pulses.

As shown in FIG. 1, it will be appreciated that if the two pulses are wider than is necessary for the energization of an element, the reflected pulse and the incident pulse need not exactly coincide, -i.e., need not completely overlap, at the region across which the elements are connected; only portions of the pulses need to overlap at the region across which the element is connected.

FIG. 5 shows a relatively simple form of a twocoordinate matrix 31 formed in accordance with the invent-ion. Matrix 31 comprises a group of similar microstrip lines and associated diodes, as shown in FIG. 1. A total of sixteen Esaki diodes are thus arranged in the two-coordinates system and, more specifically, in four rows and four columns. The columns of the Esaki diodes are designated alphabetically as A, B, C and D; the rows are designated as l, 2, 3 and 4. Each diode may be designated by its coordinate; for example, diode 2B is positioned at the intersection of row 2 and column B. To select a particular Esaki diode, a driver connected to the associated microstrip line is activated to send a first pulse of a given polarity down the line at a controlled later time and a second pulse of said given polarity the line to meet the reflected first pulse at the point on the microstrip line across which the selected diode is connected. Assume, for example, diode 2B is to be selected, driver 24!) would be activated to send pulses time spaced, as discussed hereinabove, along line 19b to energize diode 23. By activating a particular driver and controlling the timing of the pairs of pulses sent down the line, any one of the elements of the matrix may be selected.

The drivers are bipolar devices. During the reading operation, pulses of opposite polarity to the first mentioned pulses are propagated down the line to shift the states of the selected binary elements to their initial state.

The practical size of the matrix and the number of elements used will be limited by dissipation losses due to the lines and also by the speed at which it is desired to operate the matrix. Obviously, as the length of the line is increased, the time required for a pulse to propagate through the line and be reflected back is increased.

Since the control cycle time depends on the time required for the resistor connected to the driver to absorb the reflected second pulse, some cycle time reduction is possible by placing the driver at some other point on the line, such as the center of the line, and terminating both ends of the line either by an open or short circuit. In the foregoing manner, one or two elements connected to the same line can be selected. For example, two elements may be selected by spacing the elements on either side of the line the same distance from the driver while different elements may be selected by spacing the elements on one side of the driver differently from the elements on the other side of the driver and selectively controlling the timing of the driver pulses.

Referring to FIG. 6, a word memory array may be constructed using conventional ferrite cores. The memory is arranged such that in each single plane, generally indicated as 42, only the cores 41 on the plane that are to be addressed simultaneously are threaded by an associated drive line 43. Due to the relatively long control signal required by ferrite cores, the width of pulses necessary to energize the cores would overlap over a considerable, not along the entire length of the drive line.

In accordance with the invention, the groups of cores representing groups of digits are separated electrically by using delay lines 44 between planes, as shown in FIG. 6. The delay line 4-4 at the end of the drive line 43 will be terminated, as discussed hereinabove, to reflect the pulse energy. In this embodiment, the last delay line 44 is short circuited to provide an increase of current at the desired location.

The number of individual planes and the number of cores in each plane are practically limited by losses in the 7 lines and the required speed of operation, as noted above with respect to FIG. 5

While the invention has been particularly shown and described with reference to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.

What is claimed is:

1. In combination: a transmission line, means for coupling a plurality of energizable elements to spaced points on said line, said line being terminated to reflect energy propagated through said line, and pulsed energy generating means for providing time spaced pulses to said line, succeeding ones of said pulses supplementing the reflected preceding ones of said pulses at selected regions along said line to energize predetermined ones of said elements.

2. In combination: a transmission line, means for coupling a plurality of energizable elements to spaced points on said line, pulsed energy generating means controlled to provide selectively spaced pulsed energy to said line, said line being terminated to reflect energy propagated through said line, and said generating means during a control opration providing an initial pulse and a time spaced later pulse which supplements the energy of the reflected initial pulse at a predetermined region along said line to energize a selected one of said elements.

3. A plurality of energizable elements each having at least two stable states, a transmission line, said elements being coupled in parallel to spaced points on said line, pulsed energy generating means being controlled to provide selectively spaced pulses to one end of said line, the other end of said line being open circuited to reflect pulse energy propagated through said line, and during each control operation said generating means providing a first voltage pulse and a time spaced later voltage pulse which supplements the energy of the reflected first pulse at a determined region along said line to shift a selected one of said elements to one of its stable states.

4. A circuit in accordaance with claim 3 in which line matching networks couple said elements to said line.

5. A plurality of energizable elements each having at least two stable states, a transmission line, said elements being coupled in series to spaced points on said line, pulsed energy generating means being controlled to provide selectively spaced pulses to one end of said line, the other end of said line being short circuited to reflect pulse energy propagated through said line, and during each control said generating means providing a first current pulse and a time spaced later current pulse which supplements the energy of the reflected first pulse at a determined region along said line to shift a selected one of said elements to one of its stable states.

6. A circuit in accordance with claim 5 in which line matching networks couple said elements to said line.

7. A matrix comprising a plurality of energizable elements each having at least two stable states, a plurality of drive lines each arranged to reduce the velocity of propagation of energy therethrough, means for coupling said elements to spaced points on respective ones of said drive lines, pulse energy generating means being controlled to provide spaced pulses to respective ones of said lines, each of said lines being terminated to reflect pulse energy propagated through said line, and during each control operation one of said generating means selectively providing a first pulse and a time spaced second pulse which overlaps at least a portion of the reflected first pulse at a determined region along the associated line to energize a particular one of said elements to one of its stable states.

8. A matrix comprising a plurality of bistable cores arranged as groups in distinct planes, drive lines, one such drive line passing through corresponding groups of cores in each of said planes, delay networks comprising a portion of said line extending between planes, pulse energy generating means connected to one end of respective ones of said lines and controlled to provide selectively spaced current pulses thereto, the other end of each of said lines being short circuited to reflect pulses propagated through said lines, during each control operation a selected generating means providing a first pulse and a time spaced second pulse which overlaps at least a portion of the reflected first pulse at a determined region along the associated line to shift a group of cores in one plane to one of their stable states.

References Cited in the file of this patent UNITED STATES PATENTS 2,795,775 Faymoreau et al June 11, 1957 2,907,000 Lawrence Sept. 29, 1959 2,989,732 Young June 20, 1961 2,991,394 Archer et al July 4, 1961

Claims (1)

1. 7. A MATRIX COMPRISING A PLURALITY OF ENERGIZABLE ELEMENTS EACH HAVING AT LEAST TWO STABLE STATES, A PLURALITY OF DRIVE LINES EACH ARRANGED TO REDUCE THE VELOCITY OF PROPAGATION OF ENERGY THERETHROUGH, MEANS FOR COUPLING SAID ELEMENTS TO SPACED POINTS ON RESPECTIVE ONES OF SAID DRIVE LINES, PULSE ENERGY GENERATING MEANS BEING CONTROLLED TO PROVIDE SPACED PULSES TO RESPECTIVE ONES OF SAID LINES, EACH OF SAID LINES BEING TERMINATED TO REFLECT PULSE ENERGY PROPAGATED THROUGH SAID LINE, AND DURING EACH CONTROL OPERATION ONE OF SAID GENERATING MEANS SELECTIVELY PROVIDING A FIRST PULSE AND A TIME SPACED SECOND PULSE WHICH OVERLAPS AT LEAST A PORTION OF THE REFLECTED FIRST PULSE AT A DETERMINED REGION ALONG THE ASSOCIATED LINE TO ENERGIZE A PARTICULAR ONE OF SAID ELEMENTS TO ONE OF ITS STABLE STATES.

Images