US3002183A - Digital computing - Google Patents

Description

P 1961 s. S. GUTERMAN 3,002,183

DIGITAL COMPUTING Filed Dec. 15, 1954 INPU T 0 aurpu-r OUTPUT aurpur au-rpur INVENTOR SA D/A SYDNEY GuTERMA/v ATTORNEY nite States This invention relates to computing, and particularly to the use of magnetic or electric field-sustaining elements in the handling of computations or analogous information in code pattern as, for example, a binary code utilizing two contrasting digits adapted for representation in the form of two contrasting magnetic or electric field conditions brought about alternately in each of said elements by application of code-controlled pulse energy thereto.

The invention is characterized by the application, to one or more field-sustaining elements of the character indicated, of circuitry operable to cause a single one of such elements to deliver signal energy in selective manner to a designated one of two outlet points so that said single field-sustaining element becomes, in effect, a nuclear point from which signal energy may flow selectively to two ultization points, each of equal and coordinate accessibility and therefore individually available for receipt of a given transmitted signal to the exclusion of the other utilization point, according to the selection impressed upon the transmitting element constituting said nuclear point.

The invention is herein illustrated and described in several forms it may assume, each of which forms incorporates a plurality of toroidal cores, each being of material having high magnetic retentivity, and each having the potentiality for relatively slow variation of magnetic flux intensity from a maximum in one polar direction to a maximum in the opposite polar direction, in response to application of correspondingly low flux-shifting energy to conductors wound upon said cores. It is to be understood, however, that the invention may be embodied in other forms differing from those herein disclosed, but based upon the same controlling principles.

In the drawing:

FIG. 1 is a diagram of components and electrical connections constituting an embodiment of the invention;

FIG. 2 shows a second embodiment; and

FIG. 3 shows a flow pattern diflfering from those of FIGS. 1 and 2.

Referring first to FIG. 1, the embodiment, there illustrated, includes a rectangular two-dimension pattern of magnetic cores A to J of toroidal form and of ferromagnetic composition providing a rectangular or nearrectangular hysteresis loop characteristic facilitating the driving of each core to a condition of saturation in one polar direction or the other, as driving shif current is applied to one or the other of the x and y shift windings, which windings are connected in separate circuits (not shown) including separate driver elements normally in non-conducting states, but individually tiriggerable to cause delivery of an x or y shift pulse, according to whether it is desired to advance the coded information along vertical rows (such as row ACF) of the twodimension register, or along horizontal rows of the register, such as row A-B-D.

In addition to the shift windings x and y, each core has input windings i and i and a single output winding 2. The windings i and i on the respective cores serve as input whereby signals may be applied to the cores by means of signal energy delivered to a selected one of said windings, i or i as the case may be. Delivery of such signal energy to a given core, by way of either of said input windings, will operate to reverse the flux-saturation condition preexisting in said core, thus writing in Patented Sept. 26, 1961 to said core a new digital value. out of such value will be by way of the output winding z, now to be described.

The output wind-ing z of any given core has connection with the input winding i of the horizontally adjacent core, and also with the input winding i of the vertically adjacent core, but in said connections there are interposed delay networks N and N respectively, preceded by diodes D and D respectively, the former diode being receptive to E.M.F. of positive polarity only, whereas the latter is relatively impervious to positive but receptive to negative Both the D and D sets of diodes are constructed and composed in such a manner as to have maximum resistivity when the application is of low magnitude, and to carry out the purposes of the invention the delay networks N and N (as well as associated components of the register) are so designed that the energy transfer from core to core occurs at relatively slow speed. Because of this relatively slow rate of signal energy transfer, the core flux-reversing process is correspondingly protracted; hence development in the intercore circuitry is of a low order, so that current values through the diodes remain at a minimum level that is too insignificant to be a signal transferring factor or a circuit loading factor.

Let it now be supposed that core A is flux-saturated in the polar direction corresponding to the code value of l, and let it further be supposed that a shift pulse is applied to the x winding of core A. The positive E.M.F. developed on output winding z of core A will be high, as the shift pulse is relatively strong. The output current will flow thrugh diode D and network N arriving (after the prescribed delay period inherent in the design of N at the input winding i of core C, where it becomes eflfective (over a protracted time interval, as above noted) to reverse the flux-saturation polarity of core C from the 0 state to the 1 state. Thus the code value 1 is shifted horizontally from core A to core C of the register under the control of the horizontally acting shift pulse P The described flux reversal in core C operates to develop a negative in output winding z of core C. Because of this a current can pass through diode D linking cores C and E. However, the accompanying E.M.F .is very small; hence such current fiow is too small in magnitude to present any appreciable loading factor to interfere with the proper functioning of core C, nor will it be in any respect capable of altering the flux status of the core E. In other words, the code signal transfer will be from core A to core C only (under the hypothesis assumed) and neither core E nor any other core of the two-dimensional register will be aifected. In analogous fashion the coded signal could have been shifted vertically from core A to core B if shift pulse P rather than pulse P had been applied.

It is to be understood that the two dimensions of the register may be increased to include more stages, as desired, both vertically and horizontally. Also, all shift pulses P to the horizontal shift-controlling windings x are preferably applied simultaneously by wiring all of the windings x in series relation to a suitable potential source at one terminal of the x shift line, whose other terminal would be constituted by a suitable driver element, such as a pentode tube, facilitating control of the timing and duration of each shift pulse, in conventional manner. The shift line may likewise include all windings in series relation, and a similar driver permitting P pulse application at time period-s other than those marked by P pulse application, and in accordance with whether vertical propagation of the coded data, rather than horizontal propagation, is desired.

The FIG. 2 embodiment is the same as that of FIG. 1 except that the i and i windings of each core are con- The subsequent readsolidated into a single i winding for each core, which single winding may receive input signals of opposite polarities on diiferent occasions, to produce opposite flux saturation conditions for flux reversal in the associated cores, as requirements dictate. In allother respects the mode of operation is the same, and'the corresponding parts are correspondingly designated in' both views.

In refering to vertical and horizontal operation, it is intended to embrace any combination of diverging data transfer paths by means of which information may be moved along the register to two selected output points in predetermined fashion, and irrespective of the physical degree of divergence of one path in relation to the other. FIG. 3 illustrates a number of paths diverging at various physical angles.

This invention is not limited to the particular details of construction, materials and processes described, as many equivalents will suggest themselves to those skilled in the art. It is accordingly desired that the appended claims be given a broad interpretation commensurate with the scope of the invention within the art.

What is claimed is:

1. A system comprising a plurality of magnetic flux sustaining elements, first and second windings on each of said elements, diverging means including a delay network in circuit with said second winding for feeding signals from the second winding of any one of said elements to the first winding of each of two other elements, and means including third and fourth oppositely poled windings on said elements for activating a selected one of said two diverging means, said diverging means including unidirectional impedance means of opposite polarities in circuit with said delay network.

2. A system comprising a plurality of magnetic fluxsustaining elements, first and second windings on each of said elements, diverging means including a delay network in circuit with said second winding for feeding signals from the second winding of any one of said elements to the first winding of each of two other elements to reverse the polarity of flux saturation thereof, means for activating a selected one of said two diverging means, said activating means comprising additional oppositely poled windings on said elements, and means for energizing said additional windings in a selected polar direction.

3. A system comprising a plurality of magnetic fluxsustaining elements, first and second windings on each of said elements, diverging means for feeding signals from the second winding of any one of said elements to the first winding of each of two other elements, said diverging means including a delay network in circuit with said'secnd winding, means for activating a selected one of said two diverging means, said activating means comprising additional oppositely poled windings on said elements, and means for sending pulse energy of a selected polarity into said additional windings.

4. A system comprising a plurality of magnetic fluxsustaining elements, first and second windings on each ofsaid elements, diverging means for feeding signals from the second winding of any one of said elements to the first winding of each of'two other elements, said diverging means including a delay network in circuit with said second winding, means for activating a selected one of said two diverging means, said activating means comprising third and fourth oppositely .pdle'd windings on said elements, means -for sending pulse energy "for-one polarity into said third winding, and means for sending pulse energy of the opposite polarity intosaid fourth winding.

5. A system comprising a plurality of magnetic fluxsustaining elements, diverging means including a delay network and unidirectional current means of opposite polarity for feeding signals from any one of said elements to two other of said elements, means for activating a selected one of said diverging means, said activating means comprising electric pulse enengy-receiving means including oppositely poled windings on said elements for driving the field of said elements to saturation in a selected polar direction, and thereby generating an'E.M.F. for operation of the selected signal-feeding means.

6. A system as defined in claim 1, wherein said unidirectional impedance'rneans has maximum resistivity when the applied thereto is of low magnitude, said delay network facilitating selective maintenance of said at said low magnitude.

7. 'A system comprising a plurality of magnetic fiuxsustaining elements, input and output windings on each of said elements, diverging means for feeding signals from the output Winding of any one of said elements to the input winding of two other of said elements, said diverging means including a delay network in circuit with said output winding, and means including third and fourth oppositely poledwindings on each of said elements for activating a selected one of said two diverging means, said diverging meansincluding unidirectional impedance means of opposite polarities'in circuit with said delay network.

References Cited in the'file of this patent UNITED STATES PATENTS 2,652,501 Wilson 'Sept. 15, 1953 2,673,337 Avery 'Mar. 23, 1954 2,708,722 An Wang 'May 17, 1955 2,709,798 Steagail May '31, 1955 2,720,597 'Stuart Williams Oct. 11, 1955 2,886,799 Crooks May 12, 1959 OTHER REFERENCES Thesis by M. K. Haynes, pp. 46-50, 57-58, Dec. 28, 1950.

Thesis by R. C. Minnick on The Use of Magnetic Cores as Switching Devices, Harvard Progress Report No. BL-3, September 1953, pp. 5-14 through 5-17.

NBS Report No. 2940, The Diode-Capacitor Memory, by A. W. Holt, November 1953,,pp. 1-6 and three additional sheets containing FIGS. 1-3.

Images