US2901731A - Switch devices - Google Patents

Description

Aug. 25, 1959 R. L. SNYDER, JR

SWITCH DEVICES Filed May 12, 1954 5 Sheets-Sheet 2 INVENTOR.

jg' zwdz, 5 W hm Aug. 25, 1959 R. 1.. SNYDER, JR

SWITCH DEVICES 5 Sheets-Sheet 3 Filed May 12, 1954 INVENTOR. jzfiaxjj 5y j? g- 1959 R. L. SNYDER, JR 2,901,731

swncu DEVICES United States Patent SWITCH DEVICES Richard L. Snyder, Jr., Moorestown, NJ.

Application May 1-2, 1954, 'Serial No. 429,351 3 Claims. 01. 340 174 This invention relates to electrical switches and selectors for controlling the passage of electrical energy between two'pairs of circuit terminals or between one pair of terminals and a selected pair of a plurality of other pairs of terminals. It relates particularly-to switches and selectors-utilizing reactance elements with saturable magne'tic cores, the saturation of which can be controlled at will. The switches and selectors of the present invention comprise networks of such reac'tance elements, so controllable as to permit or prevent at will the transmission of electrical energy through a single switch and to permit such transmission through a selected path of a plurality of alternate paths between input and output terminals of a selector.

Transmission of electrical energy between pairs of terminals of a switching or selector device or non-transmission thereof, is conventionally controlled by mechanically moving metallic contacts, thereby to make or break metallic conductive connections. Such contacts are usually parts of electromagnetic relay-s. Since such devices employ mechanical motion and mechanical members with mass, they are not suitable for extremely rapid operation, such as required for example, in modern electronic computers. Moreover, they are subject to considerable wear and tear, making frequent replacement necessary when used in selector systems in which rapid and frequentswitching must take place.

Selector systems have been devised using electronic tubes as switching devices. Such systems are not suitable, however, to handle a wide range of power levels of signals. Some modern computers, for example, require selector systems capable of reliable operationover a powerrange in the order of 105 to-one or greater; 7

Switchesand'selectors, hitherto devised using 's-a'turablecore reactance element's, do not provide 'suflici'ent' discrimination, that is, the ratio of energy transmitted for the-saturated and unsaturated conditions of the core, respectively, is not sufficiently great to fill the needs of modern equipment, such as some "automatic computing machines.

An objectof this invention, therefore, is to provide electrical switdiing and selecting devices adapted for extremely rapid operation and for extremely wide control of the degree of transmission or non-transmission of electrical energy.

- A further object of this invention is to provide switching and selector systems capable of reliable and noisefree operation over a range of electrical energy transmitted of about to one, or greater.

' Another object o f the invention is to provide a switching, system in which the inductance values of a plurality of controlled windings in transmission paths of electrical energy are controlled by the electrical current through a single associated control winding.

The'various features of novelty which characterize my invention are pointed out withparticularity in the claims annexed to and forming a part of this specification. For

Patented Aug. 25 i 1 959 ice a better understanding of the invention, its advantages, and specific objects attained by its use, reference's'hould be had to the accompanying drawingsand descriptive matter, in which -I have illustrated and described preferred embodiments of my invention.

Of the drawings:

Fig. 1 shows a basic saturabl'e reactor structure;

Fig. 2 shows a similar structure with modified controlled winding; 7 I 1 Fig. 3 shows'a schematic representation of Fig. -1

Fig. 4 illustrates a reactor structure with a single control winding and several simultaneously controlled'windings;

Fig. 5 shows a schematic representation of the reactor of Fig. 4'; v I

Fig. 6 schematically illustrates a. switch in accordance with the present invention;

Fig. 7 schematically illustrates a similar sw-itchwith with a higher degree of discrimination between on and off positions;

Fig. 8 schematically shows a novel selector system;

Fig. 9 schematically shows a novel binary selector system used in conjunction with a memory device for automatic computing machines;

Fig. 10 schematically shows a selector similar to Fig. 8 with a greater number of alternate transmission paths;

Fig. 11 schematically shows 'a novel selector system for use in a decimal-type system;

Fig. 12 schematically shows a novel selector system for binary operation; and

Fig. 13 schematically shows a modification of the system of Fig. 12.

Fig. 1 shows a reactor element having a magnetic core 1 of relatively easily saturable material, such as certain types of laminated iron or so-called ferrite's. The core is of substantially uniform cross section, except at 2', where it is provided with a window, leaving two magnetic paths of greatly reduced cross sections that can be saturated relatively readily. A control winding 3 is pro.- vided for'saturating the narrow core sections at 2, when an electrical current is passed through the winding. A controlled winding 4 is wound around the two core sections at '2 as shown, in such directions, that a change in' magnetic flux generated by current passing through winding 4- causes voltages to develop across each coil of the Winding in an additive fashion, and equal voltages of opposite phase are induced in winding 4 that cancel each other when changing magnetic flux is generated by Winding 3. Consequently, no net voltage is induced in winding ,4 from winding 3. Conversely, the same holds true.

It is well known, that under the above conditions, the inductive reactance of the controlled winding 4 decreases considerably when the core at '2 is saturated. In practice, I have found that inductance changes of the order of to l are attainable from unsaturated to saturated condition of the core sections at 2.

From Fig. 1, it is seen that, in the absence of'current through control winding 3, the two portions of the controlled winding 4 circulate their flux about the window at 2, a's'indicated by 'the'arrows. No flux is linked with control winding 3.

When control winding 3 'is fully excited and the core sections at 2 saturated, the flux in one section produced by the current through the controlled winding 4 opposes the flux produced by control Winding 3. The opposing flux, which must be overcome, can. be appreciable, because for saturated conditions the current through the controlled winding becomes greatest as its impedance diminishes;

Fig. 2 shows a different type of controlled winding 4a on a structure otherwise substantially identical to that of Fig. 1. In this arrangement the two sections of the controlled winding are connected aiding parallel, so that their flux does not link with the control winding (not shown). When current is passed through the control winding to saturate the core structure, no appreciable voltage can be developed across one of the parallel sections of controlled winding 4a, because this would require the flux to increase further in the saturated direction. Therefore, to produce minimum 'reactance of the controlled winding, current in the controlled winding does not require higher flux from the control winding for saturation. of the core, as it does for the arrangement of Fig. 1. However, this type of controlled winding has the disadvantage of developing a short circuit when the control winding current is rapidly changed, thus preventing sufficiently rapid change in controlled 'winding inductance to make this arrangement satisfactory for very high speed operation.

My invention is not limited to the use of one or the other of these types of controlled windings, since either can be used, the choice depending upon the ultimate application of the switch or selector system in which such reactor elements are to be used.

Fig. 3 shows a schematic representation of reactor elements of Figs. 1 and 2. Core 1 is represented by short parallel lines indicating a saturable core. Control winding 3 and controlled winding 4 are drawn at right angles to each other, signifying that there is no flux linkage between the two windings.

Fig. 4 shows a basic reactor structure having a sat-urable core 1, a sectionalized control winding 3, a plurality of core openings indicated at 2, and four separate controlled windings 4. Fig. 5 is a schematic representation of such a structure.

I have found that a suitable reactor can be made, as shown in Fig. 4, using a cylindrical ferrite core, having a length of one inch, an outer diameter of 0.62 inch, and an inner diameter of 0.430 inch. The openings shown at 2 in Fig. 4 extend over the entire cylinder length and have a diameter of 0.05 inch. The number of openings depends upon the number of controlled windings to be controlled by a single control winding. In the example cited twelve holes are used. By way of example, the controlled windings may have from fourteen to twenty-eight turns, and the control winding one hundred and forty turns.

Fig. 6 shows a circuit diagram of a switch according to my invention. It comprises two saturable reactor elements 10:: and 10b, each having saturable magnetic cores 11a and 11b, control windings 12a and 12b, and controlled windings 13a and 13b. The controlled windings are connected between two pairs of terminals 14 and 15. Winding 13a is connected in series between two corresponding terminals of pairs 14 and 15, whereas winding 13b is connected in parallel relation to terminal pair 15.

Control windings 12a and 12b are connected in an energizing circuit, shown by way of illustration of principle only, comprising a battery 16 and a switch 17. Examples of practical circuit arrangements will be described later.

In operation, control windings 12a and 12b are selectively energized by switch 17 and battery 16. In the switch position shown, winding 12a is energized, saturating core 11a to minimize the inductive reactance of controlled winding 13a. No current flows through control winding 12b and the inductive reactance of controlled winding 13b has its maxitmum value.

With minimum series impedance and maximum shunt impedance, transmission of electrical energy between terminal pairs 14 and is least impeded.

When switch 17 is thrown to the right-hand position, control winding 12a becomes deenergized and winding 12b energized, so that the inductive reactance of controlled winding 13a assumes its maximum value, and

4 the reactance of 13b a much lower value. With a high series impedance and low shunt impedance, minimum transmission between terminal pairs 14 and 15 is effected.

The difference between maximum and minimum transmission with this novel switch is considerably greater than with saturable magnetic devices hitherto known, since those use single controlled windings only, instead of a network thereof. Increased discrimination between maximum and minimum transmission can be obtained by other networks, as shown below. The source of energy can be connnected to either the terminal pairs 14 or 15 and a receiver to the opposite pair, or transmission of energy can be eifected selectively in either direction, as will be shown later.

Fig. 7 shows a switch similar to that of Fig. 6, except that it utilizes a network capable of greater discrimination in that it employs two series and two shunt inductances. Again there are provided two saturable reactor elements 20:: and 20b having cores 21a and 2117, respectively, control windings 22a and 22b, and controlled windings 23a, 24a, and 23b and 241). Control windings 22a and 22b are connected to a source of current 27 by way of a switch 28. V

For the switch position shown, winding 22a is energized and controlled windings 23a and 24a are at their minimum inductance value because of saturation of core 20a. Controlled windings 23b and 24b have their maximum inductance value, since control winding 22b is deenergized and core 21b unsaturated. For this condition, minimum series inductance values and maximum shunt inductance values, maximum transmission of energy between terminal pairs 25 and 26 can take place.

When switch 28 is turned to the right-hand position, then maximum series and minimum shunt inductance values obtain and minimum transmission between terminal pairs 25 and 26 can take place.

Fig. 8 shows a selector for selectively connecting one pair of first terminals to one of two pairs of second terminals by greatly increasing the degree of transmission between these pairs of terminals, while simultaneously reducing to a minimum the transmission between the first terminals and the other pair of second terminals.

Accordingly, there are provided two saturable reactor elements 30a and 30b, with control windings 31a and 31b, saturable cores 32a and 32b, and controlled windings 33a, 34a, 35a and 33b, 34b and 35b, as shown. The controlled windings are connected in networks between terminal pairs 36, 37 and 38, so that windings 33a and 34a are in series between terminals 36a and 37b, and 33b connected between the junction of 33a and 34a and tenninals 36b and 37a, the latter being conductively connected as shown. Controlled windings 34b and 35b are connected in series between terminals 36a and 38b, whereas 35a is connected between the junction of 34b and 35b and terminals 36b and 38a, the latter being conductively connected.

In operation, assume control winding 31a to be energized, core 30a saturated, while control winding 31b is deenergized and core 30b unsaturated. In this condition, the inductance values of series windings 33a and 34a is very low, whereas that of shunt winding 33b is very high, permitting maximum transmission of energy between terminal pairs 36 and 37. Conversely, the inductance values of series windings 33b and 34b is very high, whereas that of shunt winding 35a is very low, permitting minimum transmission of energy between terminal pairs 36 and 38. Hence, terminal pair 36 is efiectively connected to terminal pair 37 and effectively disconnected from terminal pair 38.

In one embodiment of my invention, I found extremely effective discrimination to be obtained when varying windings 33a and 34b from 10 to 2000 micro-henries, 34a and 35b from 5 to 1000 micro-henries, and 35a and 33b from 20 to 4000 micro-henries, from saturated to unsaturated condition of the cores.

asomsi Obviously, energy can be equally well transmitted selectively from 36 to either 37 or 38, or .from either 37 or 38 to 36. Since these networks are bi-directional and this quality is made use of in accordance with my invention, I have avoided the use of the terms .input and output terminals, since the same terminals are used in short succession in both capacities, as will be pointed out in connection with Fig. 9.

Fig. 9 shows the schematic circuit diagram of an embodiment of my novel selector system used in conjunction with an automatic computing machine. In such machines it is necessary to store intermediate solutions of a problem for later use by the machine. The information is often .stored in the form of magnetic patterns in a relatively high coercive force term-magnetic surface of a rotating drum produced by positive and negative current pulses in windings on magnetic recording heads positioned close to the magnetizable surface of the drum. These pulses may have a duration in the order of between 0.2 and microseconds, spaced in time by intervals in the order ofbetween 3 and 30 microseconds. v

This circuit includes a recording amplifier, generally indicated at 40, adapted to receive from part of a computer (not shown) pulses to terminals 41 and'42, indicative of one or zero as is conventional in binary computing machines. Such pulses produce current pulses of one direction or opposite direction in a transformer secondary winding 43, applied to terminal pair 44 of my selector system. One of the functions of the selector is to connect selectively terminals 44 to one of four magnetic recording and reproducing heads 45, 46, 47 and 48 cooperating with a magnetic drum 49, driven by a motor 50, for recording pulses received at 41 and 42 on one of the four tracks on drum 49.

Another function of the selector system is to disconnect the heads 45, 46, 47 and 48 from terminals 44 and connect a selected one to terminals 51, connected to a reproducing amplifier generally indicated at 52. The output of this amplifier at terminals 53, is applied to a portion of the computer (not shown) requiring the stored information.

Selection of recording or reproducing of information and selection of the specific magnetic head is effected by the program section of the computer, generally indicated at 54, transmitting properly timed pulses of proper polarity to the control grids of tubes 55, 56, 57, 58, 59 and 60, whose anodes are connected to control windings and to a common battery 61.

The selector system comprises three sections 62, 63 and 64, each with two saturable reactor elements. Sections 62 and 63 are the same as shown in Fig. 8 and described in connection therewith. When program section 54 places a positive voltage on the control grid of tube 56 and a negative voltage on the grid of tube 55, terminals 44 and 65 are connected and terminals 51 disconnected from 65. The system is in recording position.

The next step is to connect terminals 65 to one pair of terminals 68, 69, 70 or 71, selectively. This is done in two steps in selector sections 63 and 64. Section 63 connects terminals 65 to terminals 66 or 67. Section 64 connects terminals 66 to 69 or 70 and terminals 67 to 68 or 71. Assume a positive voltage on the grid of 57 and a negative voltage on the grid of 58, then terminals 65 are connected to terminals 67.

Section 64 comprises basically two sections 63, except that they are combined into only two saturable reactor elements with only two control windings, instead of four elements. Assume a positive voltage on the grid of 59 and a negative voltage on the grid 60, terminals 66 will be connected to 70 and terminals 67 will be connected to 71. However, since there is no connection between 65 and 66, the connection is now made from 65 via 67 to 71 and magnetic head 45 for recording information on drum 49.

A positive voltage on the grid of 55 and negative voltage on the grid of 56 will-cause the recorded signal .to be reproduced and appear at terminals 53.

Fig. 10 illustrates how a number of selectors of the type of Fig. 8 can be combined into a single selector unit with only two saturable reactor units. This selector comprises two saturable reactor elements v and 81,, each having a saturable core 82 and 83, and a control winding 84 and 85, respectively. This selector connects selectively each one of five first terminals 86, 87, '88, 89 and 90' to one of two pairs of second terminals, there being a total of ten pairs, numbered 91 through 100. If control winding 84 is energized while no current flows through 85, all terminals 86, 87, 88, 89 and 90 are connected to terminals 91, 92, 93, 94 and 95, respectively. Inspecting terminals 86, 91 and 100, and connecting controlled windings, it is seen that controlled 'windings 1'01, 102 and 103 form the same series and shunt network between terminals 86 and '91 as shown in 'Fig. 8. Similarly, windings 104, 105 and 106 form a network between terminals 86 and 100. The same types of network are formed between other terminals. Obviously, similar selectors can be built for any number X of pairs of first terminals and 2X pairsof second terminals.

Fig. 11 shows how various selector units can be combined into a decimal selector system. This system comprises sections 110, 1 11, 112 and 113, with one pair of first terminals 114 and ten pairs of second terminals 115. Each section has two control windings, their terminals indicated by numeral 116. Sections 110 and 111 are as shown in Fig. 8. Section 112 is the same as shown as section 64 of Fig. 9, while section 113 is shown in detail in Fig. 11.

Of the two pairs of second terminals of 110, one pair is connected to section 111, and can be selectively connected to any one of four of the five pairs of first terminals of section 113. The remaining pair of terminals is connected directly to one of the second terminals of section 110. By properly energizing and deene'rgizing control windings of the various sections, terminals 114 can be connected to any one pair of terminals 115.

Because of the attenuation of energy sutfered by transmission through sections 111 and 112, as contrasted to the absence of attenuation in the direction between sections 110 and 113, section 110 is made unsymmetrical, that the attenuation of 111 and 112 are built into one of the networks. In this manner energy transmitted between 114 and any pair of terminals 115 suffers the same attenuation regardless of direct connection between sections 110 and 113 or transmission by way of sections 111 and 112.

Fig. 12 shows a selector system comprising selector units to 126, each as shown in Fig. 8. These units are arranged in N=3 tiers, the system having a pair of first terminals 127 and 2 =8 pairs of second terminals 128. Each unit has one pair of first terminals and two pairs of second terminals. Each tier has 2 pairs of first N terminals and 2 pairs of second terminals. Second terminals of one tier are connected to first terminals of the next higher order tier. Each unit has a pair of control windings, whose terminals are indicated for unit 120 as 130 and 131. The networks of controlled windings in the units, can be exactly the same as in Fig. 8, or of any other suitable design employing more or fewer controlled windings, as will be obvious to those skilled in the art, Without departing from the spirit of my invention.

This system has the advantage of using identical units, but has the disadvantage of requiring seven separate pieces of information, to energize seven and deenerg'ize another seven associated control windings.

Fig. 13 shows another system for accomplishing the same purpose, using three selector units 140, 141 and 142, that is only one unit per tier. Unit is of substantially the same designas shown in Fig. 8. Unit 141 is of the same design as section 64 of Fig. 9. Unit 142 is of the same design as shown in Fig. 10, except for omission of one pair of first terminals and associated two pairs of second terminals as Well as associated controlled windings. For example, unit 142 is obtained from Fig. 10 by omitting terminals 86, 91 and 100, as Well as windings 101 to 106, inclusive.

The system has a pair of first terminals 143 that can be selectively connected to any one pair of second terminals 144 by selectively energizing and deenergizing the appropriate control windings of each unit. Each unit and each tier has two control windings, the terminals for unit 140 being indicated at 145 and 146.

This system has the advantage of requiring only three pieces of information, to energize three and deenergize another three associated control windings. It has the disadvantage of requiring three difierent types of selector units.

While I have illustrated and described the best form of embodiment of my invention now known to me, it will be apparent to those skilled in the art that changes may be made in the form of the apparatus disclosed without departing from the spirit of my invention as set forth in the appended claims, and that in some cases certain features of my invention may be used to advantage without corresponding use of other features.

Having now described my invention, what I claim as new and desire to secure by Letters Patent, is:

1. An electrical switch comprising a pair of first terminals and a pair of second terminals, a first and second mechanically static electrical reactance element, said elements exhibiting non-linear controllable reactance characteristics, control means associated with each of said elements for varying the magnitudes of said reactances, an electrical circuit path, between said first and said second pairs of terminals including said first element in series relation to said circuit path and said second element in shunt relation to said circuit path, and circuit means including said control means for selectively effecting simultaneous increase of the reactance of one of said elements and decrease of the other of said elements, thereby to control the electrical impedance between said first and second pairs of terminals.

2. An electrical switch comprising a pair of first terminals and a pair of second terminals, a first plurality and a second plurality of mechanically static electrical reactance elements, said elements exhibiting non-linear controllable reactance characteristics, first control means associated with said first elements for simultaneously and identically controlling the reactances of said first elements, second control means associated with said second elements for simultaneously and identically controlling the reactances of said second elements, an electrical network connecting said first and second pairs of terminals including said first elements in series relation to said pairs of terminals and said second elements in shunt relation to said pair of terminals, and circuit means including said control means for electively effecting simultaneous increase of the reactance of one plurality of elements and decrease of the reactance of the other plurality of elements, thereby to control the electrical impedance between said first and second pairs of terminals. I

3. An electrical switch comprising a pair of first circuit points of reference and a pair of second circuit points of reference, av first and second mechanically static electrical aperiodic reactance element, said elements exhibiting non-linear controllable aperiodic reactance characteristics, control means associated with each of said elements for varying the magnitudes of said reactances, electrical circuit paths between said first and said second pairs of circuit points of reference including substantially solely said first element in series relation to said circuit points of reference and said second element in shunt relation to said circuit points of reference, and circuit means including said control means for selectively efiecting simultaneous increase of the reactance of one of said elements and decrease of the reactance of the other of said elements, thereby to control the electrical impedance between said first pair and said second pair of circuit points of reference.

Progress Report No. BL3 of the Computation Labo ratory, Harvard Univ., by Robert C. Minnick (pages 5.6, 5.13-5.16, and 5.18 relied upon).

Images