Патент USA US2404250код для вставки
July 16, 1946. Y ,1. A. RAJCHMAN 2,404,250 COMPUTING SYSTEM Filed Jan. 22, 1944 3 Sheets-Sheet l w11 ¿wam @am haw ma lul. el! /ld W . ' /aa /41 /a /2/ l /// _ 0% 0 momma’ mac-'5 > /lPdr T0523 ì INPI/T Tl/BES FMPrmFfmG/d ¿v9/MY orf/e :avr/m rases Ä 7TOR/VEY July 16, 1946,. v.|- A.' RAJCHMAN 2,404,250 COMPUTING SYSTEM Filed Jan. 22, 19514 3 Sheets-Sheet 3 .Maman <É1bß INVENTOR. 2707x011/ \ “día Y HTTÜR/VEY 2,404,250l Patented July`16, 1946 UNITED _STATES PATENT "oEEicE y ». _ 2,404,250 coMrU'rTNG sys'rEM Jan A. autumn, Princeton, N. J., assigner 'to Radio Corporation of America, a corporation of' Delaware Application January 22, 19‘44, Serial No. 519,299 6 Claims. (Cl. 235-61) ` 1 2 . whether there is a 4:22 or not, the fourth whether there is an 8:23 or not, etc. This invention relates to computing systems of a type disclosed in a copending application Se ' It is obvious that fractions and fractional numrial No. 511,729 flied Nov. 25, 1943. It has for bers can be expressed in the binary system in a its principal object the provision of an improved computing circuit and method of operation ß manner similar to the decimal fractions by using a “binal point” analogous to the “decimal point." whereby the sum of a plurality ofl numbers is A table of a few fractions would be: readily derived. The two inventions are dis .tinguished from one another in that the pres 0 .0000 ent invention is directed to the problem of addi 1% .0001 tion, whereas the invention of the copending ap ‘0 1A; .0010 plication is directed to the problem of multiplica 13g .0011 tion. Certain features are common to both inven ‘A .0100 tions. The _broader claims appear in the co 1°C .0101 pending application, the claims of the present ' % .0110 application being directed more particularly to u 11. .0111 the features which make possible the addition of V2 .1000 two or more numbers. 19s .1001 All the computations are performed in terms of numbers. The present computing device is there fore of the numerical type. as contrasted with '.0 ldevices using continuously variable physical quan % .1010 H .1011 % .1100 tities, such as voltage, current or phase, as variable H .1101 "/a H .1110 .1111 of computation. The whole computation is made in the binary system of numeration so that any number is expressed as a sum of powers of two u in which the coemcients of the terms are zero or one. These are the-only two digits of the 1 1.0000 ' 1„1c 1.0001 .1% 1.0010 - binary system. In this system, a number is expressed thus‘. where the coeiiicients ak are either one or zero. The numbers can be written in the usual digital representation as shown for the first seventeenl numbers in the following table: .0 1 2 a 0, 1 10 11 4 - 100 5 e v s 9 10 11 101 11o 111 1000 1001 1010 1011 12 _13 14 ~ 1100 1101 1110 só 13m/„4l 1101.011011 For any number the first digit from they right of the binal point signiiies`whether there is a 1/2=2-1 in the number or not. the second whether there is a V4=2-'-', the third whether there is a 35 l/ß=2-3 or not, etc. « This system of numeration was chosen -be cause Amost electronic computations are more easily performed in it than in any other system. This unusual method of expressing numbers does not involve any practical diiliculty so long as the input and output of the computing device are converted automatically to, control some physical apparatus, such as an anti-aircraft nre» control _ system. Under such conditions, no cipherlng or 46 deciphering to or from the decimal numeration is _K involved. All the operation is made in a direct, system in which the binary number is expressed by a sys tem of as many‘potentials as there are digits inA 5o it, each potential having one of two definite values V1 and V2 corresponding respectively to the digits 15 ` 1111 zero and one. All these potentials exist simul 1s 10000 taneously -on a system of conductors each carry-l ing a potential corresponding to one digit of the For any number the first digit from the right, or ñrst “digital position," signifies whether there u number. Thus, for example, to express the first> seventeen numbers, iìve conductors would be is a 1-=2° in the number of not, the second digital required. The number 9 would be expressed by place whether there is a 2=21 or not, the third l giocato 3 the in any column has an odd number of pairs. an excitation of the live conductors: VrYaViviVs. since it can be written as 01001. In? e a device, two or more such systems of odd number- of groups of four or an odd num ber of groups of eight. Thus in the above ‘ex potentials are combined and 'a new system of po ample the operation can be followed by starting tentiils'isderivediromthem. Theresultofthe from the right column and going through all computation is the stationary unal value of these output pot'entiais'antl` dependsA only on the sta columns successively as follows: iive ones, write one, carry over one two columns to the‘left (since there is one group of four'in five and one is odd) ; three ones, write one, carry over one to the tionary value of the input potentials. regardless of the manner in which they were reached. A sudden change in one or more input digits will. nex`t column to the left (since there is a pair ai'ter short transientacause the output poten in three); eight ones. write zero and carry over tials to reach their correct stationary values. so that the operation of the direct computing device one three columns to the left (since there is.one group of eight in eight); three, write one and carry over one to the next column; two, write zero and carry over one to the next column; six, may be considered as "continuous" It does not involve any `trigger elements with inherently stable states or any other "holding” devices, nor write zero and carry over one to the next column does it necessitate any dennite sequence, timing. and one two columns to the left, one write one. or clearing pulses. Therefore. it is not a counterV v and ilnally one write one. This procedure may of any'sort and does not involve impulses. It is Y be summarised by stating that the digits of any» basically the fastest type of numerical device, one digital position and digits carried over from since no time is wasted in the proper sequencing of preceding digital positions to that particular posi operations. . . ' ' tion are added, that is to say the number of ones . As pointed out above, in the direct binary sys - is merely counted. Then the result of that count tem of computation with ñxed constraints, the4 ' is expressed in the binary system, and its lowest output potentials or currents are in direct rela tion to the input potentials so that for each place or digit of the inputs and outputs there must be digit determines the corresponding digit of the sum. its next digit, determines the carry over to theenext digital position, its further next digit ' provided a conductor. ` determines the carry over to the further next digital position, etc. Of course, the expression g The sum of two binary numbers A=10111 and B=10110 (A=23 and B=22 in the decimal numer of a number in the binary system is obtained by determining whether the number of Vunits in it is odd or even for the lowest digital position, whether the number of pairs of unitsin it is odd or even for the next position, the number of groups oi.' four is odd or even for the further next position, ation)A can be made as is done in the usual ele f mentary arithmetic. 1 1l carryover 1o111 s v as + 10110- B ' +32 etc. 101101 A+B ' _ v ' This arithmetical process is performed by the i 4s _ computing circuit ci the presentinvention. 'I'his addition proceeds by adding successively thedigitsofAandBofthesamedisitalposi tion from right «to left, taking into account the “carry over” number which results when the sum of the digits in any digital position is equal to two or more. Thus in the above example. the operation can be followed by starting from the right and saying successively: one and zero equal Important features xof the present invention. 40 are (1) a plurality of input tubes (one for each ' digital position of the numbers to be added) each / connected to operate as a cathode follower so\ I that it is either non-conductive or conducts a standard unit (about 4 ma.) of current. (2) a, plurality of carry overtubes (for transferring » vunits of current from lower digital positions to a one, write one: one and one equal two, write aero, and carry over one; one, one and one equal three, write one and carry over one; one, aero and zero equal' one. write one; etc. higher digital position) 'each connected in the - same manner as the input tubes, «(8) a plurality of carry over control tubes, (4) a resistance network so interconnecting the input, carry over ' ' control and carry over tubes as to ensure that the More generally the sum of any number of num bers can be made as is done in the usual decimal digits oney are carried over to the proper higher~ arithmetic, provided the “carry over" digits are properly taken into account. For example the sumSofsixnumbersAaAaAsAiAsandAs digital positions, and (5) means for registering or indicating in their proper order the various digits of the resulting sum. ' The input tubes for any given digital position of the numbers to be added are connected in can be calculated by writing the ynumbers in a column and adding as usual. The following parallel with one another and. through a com-_ mon resistor, to a source of plate potential. Also example illustrates the procedure: näl à }carry over digits A A1 ________________ -_ -\ 100101 37 A: ________________ _.- 001101 13 As ________________ _- 101111 47 A4 ................ -_ 010101 21 v A1 ................ .__ 101110 46 As .......... --_..._--.. `100111 39 Total _____ __.- 11oo1o11 20a ' ~ This addition proceeds by adding successively AthedigitsofAnAz . . . Asofthesarnedigitalposi \ tion from right toleft, taking into account the connected in parallel with each digital position group of input tubes and supplied with plate voltage through the same common resistor are whatever carry over tubes that are required to transfer digits one to this digital position'from lower digital positions. To the control grids of the input tubes are applied either --180- (when the corresponding digit is zero) or zero voltage (when the corresponding digit is one). >The f tubes conduct current only when the zero volt 70 age is applied to their grids. Each conducting input tube is utilized to represent a digit one and each non-conducting tube is utilized to rep resent a digit zero. As the parallel-connected input tubes and the “carry over” digits which >occur when the sum 75 parallel-connected carry over tubes. if any, for 2,404,9so I a given digital position become conducting one after another, the plate voltage of all these tubes is reduced by discrete steps because all the tubes ' 6 means may be utilised to app - these digital rep resentative potentials tothe input leads. ~ The condition of the two ~ ‘l of input tubes Ilto "and Ilto2i is'dete -« bythedigits draw current through >the same plate resistor. This voltage reduction in discrete steps is uti 5 of the two numbersto be ad. Each of these tubes is conducting a standard amount of cur lized to control the conductivity of the indicator rent of about 4 ma. when a potential (0 v.) rep tube and the carry over control tubes for this tative of the digit -one is applied to its con given digital position. The carry over control trol d and is biased oi'f when a potential tubes, in turn, control the carry over tubes by which digits are transferred to high digital po 10 (_180 v. representative of the digit zero is ap plied to its control grid. These standard unit sitions. Adjacent the resistors and leads inter currents are supplied through a common resistor connecting these various tubes are legends indi 42 and are determined by the fact that each of cating suitable values which may be utilized in the tubes is connected to operate as a cathode fol assembling and operating the circuit. lower. Under these conditions, the anode poten In the operation of the circuit, the indicator tial oi' all the input tubes decreases by a standard and carry over control tubes are conducting amount for each tube that is made to conduct the when no input or carry over tube corresponding standardunitsof4ma.andeachofthetubesmay to the' given digital position is conducting. When be considered as representing a digit one or a digit „one of the parallel-connected input and carry ‘ over tubes is conducting the indicator tube is 20 zero. For converting these various digits into a bi biased ot! so that an indication for that digital position is produced. When two of the parallel connected tubes are conducting. the ilrst carry over control tube is biased off, a carry over tube nary number which is the sum of the two num bers, a group of carry over tubes 22 to 26 and a group of carry over control tubes 21 to 32 are in the next higher digital position is made con 25 provided. The resulting sum is registered or in dicated by a group of indicators 22 to 3l which ducting, and a potential applied from the cathode of this carry over tube biases of! the indicator tube. Further steps in the operation are here may include a neon lamp or the like. The man ner in which these results are accomplished will be more easily understood in connection with ever, it should be noted that (1) the conductivity 30 Fis. 2. Fig. 2 shows the details of that part o_f the cir of the indicator tube for the given digital posi cuit which appears in the heavy lines of Fig. 1. tion is controlled jointly by the parallel-con It will be noted that the input tubes Il and II nected tubes and by the higher digital position inalter explained in detail.A At this point, how are connected to the same terminal of the re carry over tubes and (2) the carry over control tubes each include a diode element which en 35 sistor 42 as the carry over tube 2l which has its control grid connected to the carry over control ables the grid of the corresponding carry over tube Il for applying a positive potential when . tube to assume a predetermined voltage. a one is to be transferred from the second digital The invention will be better understood from the following description considered in connec tion with the accompanying drawings and its scope is indicated by the appended claims. Referring to the drawings: Figure 1 is a diagrammatic representation of a computing circuit arranged in accordance with position to the third digital position which'is rep resented by the input tubes I2 and' I 9. All the f carry over tubes 22 to 28, like the input tubes Il to 2|, are of the cathode follower type so con nected as to conduct a standard unit (4 ma.) of current. It is apparent that the potential at the lower terminal of the resistor 42 is reduced by a pre determined amount when one of the tubes 2l, Il or Il takes current, by twice this amount when in the various connections. two of these .tubes take current and by three Figure2isawiringdiagram ofapartofthe 50 times this amount when all three of these tubes circuit of Fig. 1, and . take current. These different voltages are ap Figure83and4aresimilartol1‘iga1and2 plied through the resistors 4I and 44 to the first with the exception that the circuit is so extended or control grids of the indicator tube I1 and the as to add six numbers each having six digital carry over control tube 2l. Potential is applied positions. ' Upon detailed consideration of the circuits of 55 also to these grids from a -150 v. lead through resistorsßandß. Tothesecondorscreen Figs. 1 to 4, it becomes obvious that these cir grids of the tubes Il and I1. potential is applied ' cuits are readily extended to deal with any de sired number of numbers having any number of the invention for adding two numbers (A and B) l 45 of six digital positions. circles being used to in dicate the electron discharge devices involved digital positions. ' Connectedinshunettothetubel‘lisaneon. The circuit of Fig. 1 includes one group of in 60 tube 41 for indicating when this tube is not con ducting (a condition existing when a digit xero put tubes Il to i5 to which are applied potentials isin the third digital position of the sum of the representative of the various digits of a number two numbers being added). A and another group of input tubes Il to 2| to . The carry over tube 24 of the fourth digital po which are applied vpotentials representative of the various digits of a number B. In each of 65 sition has the upper end of its cathode~ resistor connected through a resistor 4l to the first or thesefgroups, the lowest digital position is at the control grid of the indicator tube 21. The con top and highest digital position is at the bottom. trol grld of the tube 24 is connected to the diode This is indicated by the binary numbers placed element of the tube Il and through a resistor 4l above the various input leads. When zero volt age isapplied to one of'these input leads, the 70 to the anode of the tube 2l so that the tube 24 conducts current only when the tube Il is biased digital position which it represents contains a oil. The purpose of the diode element of the tube one. When _180 volts are applied to an input 3|) istoestablish atthegrid of theearry overtube lead, the digital position which it represents con 24 a predetermined potential which is interme tains a zero. Switches 40 and 4l (Fig. 2), an other computing circuit or any other suitable 75 diatethose ofthe +550 v. and -600 v. leadswhen 2,404,250 the tube 30 becomes non-conducting and no Plate current is drawn through its anode resistor by the tube. ` , . The manner in which the circult'operates to convert the digits established by the tubes 25. I3 5 and I9 into a binary number is- indicated by the and 99-I00 function to .transfer a digit one tov following tabulation: higher digital positions is obvious from the f_ore going explanation. Input tubes conducting ’ Additional indicatorltubes | 0| and |02 are used, together with the tubes 39 to 39, to indicate the sum of the six numbers. A tenth digital po sition indicator tube |03and .carry .over tube |04 None .............. _; ...... _ 1 2 3 have been indicated in dotted lines for the rea son that they- may be omitted since the corre From this tabulation, it is evident that the tubes 30 and 31 are conducting and the tube 24 is sponding digit is always zero. ' This is true be biased off when none of the tubes 25,_i|2 and I9 is conducting. When one .of the tubes 215, ilv or. at most 101111010A when each of the numbers is I9 is conducting, the potential at the lower end of As shown more ’clearlynin Fig. 4, thegcominon cause the sum of six numbers of six digits can be 111111. ,_ 4 . ’ . the resistor 42 is reduced suiliciently to bias of! 20 anode lead of the. carry over tubes 15 and 25 and the tube 31 thereby lighting the lamp 41 and in the input tubes |3,‘I9, 53, 59, 55 and 1| is con nected through resistors |05 and |05, respective dicating a. digit one in the third digital position ly, to the ?rst or control grids of the carry over ‘ of the binary number. When two _of the tubes '25, I9 or I9 are conducting, the potential Jat the lower control tubes 18 and 80 (aswell as tothe control end of the resistor 42 is reduced sufficiently to 25 grids of tubes 90 and 31'). Potential is also ap plied to these control- grids from a ---150 v. lead bias oil the tube 30. This has two results. It makes the tube 24 conducting so that a digit through the resistors |01 and |00. Poten one is carried over to the fourth digital position. tial- is applied from a +18 v. lead to the When the tube 24 conducts. a positive potential second or screen grid 'of .the indicator tube is applied to the control grid of the tube 31 so 30 31j and from a +45> " volt lead 'to the sec that this tube takes currentand the lamp 41 is extinguished. When all of the tubes 25, Il and I9 conduct the potential at the lower end fof the resistor is sufficiently negative to bias »oil both tubes 30 and 31 so that the carry over tube 244 35 remains conducting and the lamp 41 is lighted; Under these conditions, a binary number of 1100 is established in the part of the circuit detailed* in Fig. 2. How the complete sum' of two num bers represented by potential applied to all 'the input' leads is established is readily understood yfrom the foregoing explanation. Figs. 3 and 4 illustrate how the circuit of Figs. 1 and 2 may be extended to produce the sum'oi‘ ondcontrol grids 'of the carry over control tubes @0, 10 and 80.' Potential is applied to the anodes ofthe‘tubes 31, 30, 1liv and 00 from a +550 volt lead. IPotential is applied 'to the grids of the carry over tubes- 24, 1'9'and 0I from a -600 v. lead respectively through'resistors |09, I|0 and III. The grids of _the- tubes 24, 19 and 0|> are con nected directly to the diode elements of the tubes 30, 19 'and 00 and through the resistors 49, ||2 and -I Il to the anodes- ofthese tubes. >When the> tubes 30,19 and 00 are biased off, the correspond .ing vdiode -elelirient functions-to establish a pre determined potential at the grid ofthe tube 24, 19Yor 0|; Adjacent eachyresistor of' the circuit is six numbers each having six digital lpositions, 45 -a -legend indicating a suitable value which may the same reference numerals being fapplied to be' utilized in‘assembling the circuit. corresponding parts of all the ilgure's. ` ' It will be noted that -~this extended circuit in-l p As in the case` of Fig. 2, the input _and carry over tubes are _oi' the cathode follower type so cludes a group of input tubes 50 to 55, 55 to 6|, connected that each is either out oft' or draws a 52 to 61 or 68 to 13 for each additional-number 60 standard unit (about 4 ma.) of current. ' With to- be added. Since there are six input'tubes theseA connections, the potential at the upper |542 |-55-6 |--91-'-~13, I4-20-54f-60-60ï end of the common -resistor"- is dependent on 12, I3«-|9-53-59--65--1I, I2--|0-52-59-A 54-10, I'I--I1-5I--51-63--59 or |0--^|6j-50--- how mamr of theïtubes 15, 25, I9. I9, ’53, 59, 65 and 1| are conducting and decreases in equal 55-52-68'for each] digital position of the num- 5 el steps'äas one after another= of the ltubes become bers to be added, additional carry over tubes and conducting. ' I carry over control tubes are required. Thus a The values of the various resistors and the ap carry over control tube 14 and a carry over tube plied voltagesare so'ehosen- that, when one of 15 are required forV transferring 'a digit one- to the tubes 15, 25, I3, I9, 59, 09, 55 or _1| is con the third digital position when four or more of ducting, vthe-voltage atv the uppe'?'termlnal ofthe the ñrst digital position -input tubes |5--2|-`--55-- common resistor 42' is so reduced that the tube 5|-51-13 are conducting. A carry over conf' , ,311s biased'oñ, lighting the lamp 41 and indicat trol tube 19 and a carry over tube 11 are re-v ing-a binary number of- 100. ' , quired for transferring a'digit one-to the fourth~ When' two- of these tubes -are conducting. the tube’ SII>> isbias'ed ofi, the -tube 24'becomes con ducting thereby transferring a digit one to the next""higher or fourth digital'position, and a volt age"applied from the cathode- of the'tube 24 through- a resistor | I4 to the ?rst grid of the tube 70 l1“ makes this tube conductive and extinguishes digital when four-or more o1'- the second digital y ' position input tubes |4--20--54-50--55----12 'or three of these tubes and the' carry over-tube“ are conducting. A carry over control tube 18 and a carry- over tube 19 are required to carry a digit -one to the iifth digital position when four-and vup tofeight of the tubes 15-25-I3-I9-53--59ä-â-55-L-H are conducting and a carry over control tube 50 and a carry over tube 8| are required for vtransfer the' lamp 41. e ' When three tubes are conducting, the voltage atethe upper end of the resistor 442' is reduced so that the tube I1 is again biased ci! and the lamp ring a digit one to the sixthv digital when all of 1l 41‘is lighted. At this point, 1000 is registered at 2,404,250 10 and means for applying to other of said tubes the tube 24 and 100 is registered at the tube 31 making a binary total of 1100. When four tubes are conducting, the voltage at the upper end of the resistor 42 is reduced further so that the tube 18 is biased of! and the carry over tube 19 is made conducting. A poten tial applied from the cathode of the tube 19 through resistors H5 and IIS to the grids of the tubes 31 and 30 makes these tubes conducting so that the lamp 41 is extinguished and the tube 24 is biased ofi. `A binary count of 10000 is now reg istered in the tube 19. control potentials representative of digits to be , transferred from lower to higher digital posi tions, means for establishing at the anodes of the‘tubes of each group a potential depending on the number of conducting tubes in that group, and means responsive to the potentials so estab lished ior indicating the sum or' said numbers. 2. An adding circuit including a separate group of tubes for each digital position of the numbers to be added, means for applying to cer tain or said tubes control potentials representa tive of the digits of the numbers to be added, and means for applying to other of said -tubes control When ñve tubes are conducting, the tube 18 is biased oiî, the tube 19 is conducting and the tube 31 is biased ofi. Under these conditions, a bina 15 potentials representative of digits) to be trans I'erred from lower to higher digital positions, ry number of 10000 is registered in the tube 19 means for establishing at the anodes of the tubes and a binary number of 100 is registered in the of each group a potential depending on the num tube 31 making a total of 10100. ber oi conducting tubes in that group, and means When six tubes are conducting the tubes 30 including a resistance network responsive to the and 18 are biased oil', the tubes 24 and 19 are potentials so established for indicating the sum conducting, the tube 31 is conducting, a binary ofy said numbers.` 1000 is registered in the tube 24 and a binary 3. An adding circuit including a separate 10000 is registered in the tube 19. The total is group of tubes for each digital position of the now binary 11000. In this- case, the tube 31 is made conducting by potentials applied from the 25 numbers to be added, means for applying to certain of said tubes control potentials repre cathodes of the tubes 24 and 19 through re sentative of the digits of the numbers to be sistors I|4 and H5 to its control grid. added, and means for applying to other of said When seven tubes are conducting the tubes 31, tubes control potentials representative of digits 39 and 18 are biased on, the tubes 24 and 19 are conducting, and the total binary number 30 to beltransferred from lower to higher digital positions, means for establishing at the anodes registered is 11100. of the tubes of each group a. potential depending When all the tubes are conducting, the tube on the number of conducting tubes in that group, 90 is biased on, the tube 8| is conducting, the and means including a plurality of electron dis tubes 24 and 19 are biased ofi.’ and the. tubes 31, 30 and 18 are conducting so that a binary 100000 35 charge devices controlled by the potentials so established to indicate the sum of said numbers. 4. An adding -circuit including a group of This entire operation may be expressed in" input and carry over tubes, means for applying tabulation form as follows: is registered in the tube 9|. to said input tubes control potentials repre Tubes con- Tube Tube Tube Tube Tube Tube Tube Total duet 80 78 30 37 81 79 24 40 sentative of digits of the same digital position of the numbers to be added, means for applying to said carry over tubes potentials representa tive of digits to be transferred to said digital 0n.. On._ On._ On._ Oil.. 011 Oil.. 0 On-. On-_ On.. OH-, Ofi-. Oil.. Ofi-. 100 position from lower digital positions, means for On._ 0n.. Oil.. 0n.. OIL. On._ 0n.. 1,000 On._ 0n-- Ofi.. Ofi.. Ofi.. O11.. On._ 1,100 45 establishing at the anodes of said tubes poten 0n-. Oil-. On-. On.- Oil-. 0n.. 0d.. 10,000 tials dependent on the number of conducting 0n. 0i1_. 0n-. O11.. Oil-. On-- 011.. 10,100 tubes in said group, and means responsive to On._ O11.. Oil.. On.. Ofi.. 0n.. On-_ 11,000 On._ 01T.. 011'.. 011. Ofi.. On._ 0n. 11,100 said potentials for indicating when the number O11.. 0n-. 0n.. On-. On._ Oi’l-. 0 __ 100,000 of said conducting tubes is odd. 5. An adding circuit including a group of It is apparent from the foregoing description 50 inputand carry over tubes, means for applying that an important feature of the carry over to said input tubes control potentials represen system is the resistive couplings controlling the tative of digits of the same digital position of potential of the grids of the carry over control the numbers to be added, means for applying and indicator tubes. This system of couplings to said carry over tubes potentials representa operates in such a way that the effective number 55 tive of digits to be transferred to said digital of voltage steps on any grid is equal to the num position from lower digital positions, means for ber of conducting tubes l to 8 (to be converted establishing at the anodes of said tubes-poten to the binary system) minus the sum of the tials dependent on the number of conducting products of the carry over digits of higher digital positions by the corresponding power of two. 60 tubes in said group, and means responsive >to It is evident that such eiïective number of _volt said potentials for transferring digits to higher age steps is precisely the correct binary digit. digital positions. tube 28 and the input tubes I3 and I9 of Fig. 2. devices potentials representative of said first 6. An adding circuit including a iirst group From the circuit of Fig. 4, it is seen that the which has an electron discharge device for each carry over tubes 24, 19 and 89 are at the same D. C. level as the input tubes i3, i9, 53, 69, 65 05 digital position of a first number. a second group which has an electron discharge device for each and 1l so that they are suitable for group digital position of a second number, means for ing with the input tubes of the higher digital applying to said ilrst group of electron discharge positions. The same is true of the carry over number, means for applying to said second group I claim as my invention: 70 potentials representative of said second number, 1. An adding circuit including a separate and means for converting said potentials into group of tubes for each digital position of the potentials representative of the sum of said ?rst numbers to be added. means for applying to cer tain oi said tubes control potentials representa tive of the digits oi the numbers to be added. 75 and second numbers. ` JAN A. RAJCHMAN.