中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

46 records (took 0.019 seconds)

    32.
    发明专利
    未知

    公开(公告)号：DE1174541B

    公开(公告)日：1964-07-23

    申请号：DEJ0022561

    申请日：1962-10-27

    Applicant: IBM

    Inventor： ROTH ROBERT INGERSOLL ,  FLEISHER HAROLD

    IPC: G06E1/04 ,  G06F7/38 ,  G11C11/44 ,  H03M7/12

    Abstract: 978, 659. Digital electric calculating. INTERNATIONAL BUSINESS MACHINES CORPORATION. Oct. 29, 1962 [Oct. 30,1961], No. 40721/62. Heading G4A. A plurality of multidigit binary coded decimal numbers are added simultaneously using current steering techniques. The circuits particularly described use cryotron gates and cryotron adding trees as disclosed in Specification 923,770 which sum n binary digits by marking 1 out of n + 1 output lines and recode this output into binary form. The decimal numbers N1 to N36 (Fig. 1) provide inputs to sets of adding trees 102 to 105. Each decimal digit is in the 1-2-4-8 code and, for example, all the "1" order bits in the units order digits are summed in the trees 111, 121, 131, and 141. Tree 111 sums the bits in numbers N1 to N9 and marks one out of ten outputs; tree 121 receives this output and also signals representing the "1" order bits in the units order digit of numbers N10 to N18, marking one out of ten output lines and possibly a carry line to adding tree 125 which operates on "1" order bits in the tens order of the decimal numbers, i.e. each bit has a decimal weight of ten. Adding trees 131,141 are identical with tree 121. Each D decoder now sums the weighted values of the marked outputs from its corresponding adding tree and generates carries. The D decoder 152 for the tens decimal order is shown in Figs. 5a to 5c. Decoder 152 converts the one out of ten markings into a 1-2-4-6-8 representation of their digital sum and produces carries into an E decoder of the next higher order (not shown). Suppose that the sum of the "1" bits in the tens order is 7; of the "2"bits: 4; of the "4" bits; 3; and of the "8" bits: 5. The output of circuit 145 (Fig. 5) then represents 7 tens and is registered in decoder 152 as 00111; the output of circuit 146 represents 8 tens and is registered in decoder 152 as 10000; the output of circuit 147 represents 12 tens and is registered as 00010 with a carry of one hundred; the output of circuit 148 represents 40 tens and is registered as 00000 with a carry of four hundred. The five bit numbers are added by counting to give a sum 10121 represented by marking terminals 554S1,555S0,553S1,552S2 and 551S1. The carries are represented by marking terminals 541C, 542D, 543C, 544C, 545C and 546D. As an example consider the output from circuit 145. Line 421S7 has current flowing in it and sends cryotron gates resistive to stear current into the right limbs of loops 501, 502 and 506 and the left limbs of loops 510, 514. This in turn steers current into the lower limbs of loops 534, 525 and 529 and the upper limbs of loops 533, 537, which steers current into the left branches of adding trees starting at terminals 551, 552 and 553 and the right branches of trees starting at terminals 554, 555. E decoder 162 for the tens order is shown in Figs. 6a to 6c. Its inputs are the 1-2-4-6-8 coded sum from the D decoder 152, the carries from the previous order D decoder 151, and the carries from the previous order E decoder 161. Adding tree 601 sums the "1" bits and has as inputs the "1" bit sum from the D decoder, three "1" bit carries from the previous order D decoder (corresponding to the "one hundred" carries in Fig. 5c) and three "1" bit carries from the previous E decoder. One out of the eight lines 601S0 to 601S7 is marked and in circuit 605 this marking is decoded into a three bit binary number of which the highest order bit is an input to the "4" bit adding tree 603, the next order bit is an input to the"2" bit adding tree and the lowest order bit is the final sum "1" bit of the tens decimal order and is registered by the state of the current in the loop FS1. Adding trees 602 to 604 sum the "2", "4" and "8" bits to mark one out of a set of outputs. Decoders 606 to 608 code this marking into some or all of the following: (a) Markings representing digital values 20, 40 or 80; (b) Markings representing carries of 100 to 200 into the next order E decoder; and (c) Markings representing digital values 60 or 120. The markings (a) and (c) are, together with the "6" bit sum from the D decoder, inputs to an asymmetrical adding tree 609 of which the output is marking on one out of thirty-eight lines. Finally, decoder 610 translates this marking into "2", "4" and "8" bits and carries of one and or two hundred. An application of the adder to binary to coded decimal conversion is described in which coded values of 2‹ to 2 35 are stored in the registers N1 to N36 and only the relevant registers are summed, those registers whose orders in the number being converted contain zeroes being treated as empty (Fig. 7, not shown). Specification 862,178 also is referred to.

    37.
    发明专利
    未知

    公开(公告)号：DE1815841A1

    公开(公告)日：1969-08-21

    申请号：DE1815841

    申请日：1968-12-19

    Applicant: IBM

    Inventor： FLEISHER HAROLD ,  JEROME HARRIS THOMAS

    IPC: G11C13/04

    Abstract: 1,190,184. Read-only stores. INTERNATIONAL BUSINESS MACHINES CORP. 14 Jan., 1969 [15 Jan., 1968], No. 2043/69. Heading G4A. A data processing system includes an optical memory plane storing a matrix of blocks of bits, and a plurality (e.g. three or two) of matrices of lenses, arranged in series to pass light from an illuminated block, the product of the numbers of lenses in the matrices equalling the number of blocks. The required block of bits, stored in transparent/opaque form, is illuminated with linearly-polarized laser light through a digital light deflector, the matrices of lenses passing the transmitted light to an optical AND gate at which each " transparent " bit of the word to be read out changes the polarization of light from a digital light deflector used to select the word so that the light will pass through a lens array and analyser to an array of photodetectors. A bit can be recorded on the memory plane by passing light, positioned by a digital light deflector, in the reverse direction through the lens matrices, each of which has an associated matrix of electro-optic switches and analyser to direct the bit to only one block, diffusing screens and an automatic photographic emulsion developing system being also provided (in situ). The memory plane could alternatively be a Lippmann film in which case read-out is by reflection using laser light of a frequency selected by an electro-optic frequency selector and positioned by a digital light deflector, or the plane could be metal reflecting spots on glass.

Patent Agency Ranking