GB528059A - Improvements in calculating machines - Google Patents

Abstract

528,059. Calculating-apparatus. SOC. INDUSTRIELLE D'INSTRUMENTS DE PRECISION. April 21, 1939, No. 12044. Convention date, April 23, 1938. [Class 106 (i)] Electrical calculating apparatus for performing calculations of the general form: AÎB, DŒAB, or #AB, comprises a series of groups of circuits, the circuits of each group, excepting the last, controlling the circuits of a succeeding group, the groups comprising a data input group, a group representing tables of partial products, and groups for totalizing selected partial products, including circuits for summing in the normal system of notation and circuits for summing in a different system of notation, comparison of the two results determining the amount to be transferred to a higher denomination. In the embodiment described, for calculations of the form DŒAB or #AB, A and B are each limited to two digits and D is limited to four digits. The factors are entered from record cards by feeler members which pass through perforations in the cards and establish connections between a source of voltage and characteristic relays controlling multi-contact switches (e.g. of the type described in Specification 487,260, [Group XXXVII]), related to the data punched in the card. The units and tens digits of the factor B are set up in a straightforward manner (Fig. 4, not shown) while the units and tens digits of the factor A (Figs. 5 and 6, not shown) and the digits of the term D (Fig. 9, not shown), are subject to combination coding in order to reduce the number of switches required. The multiplication of two two-digit numbers (see Fig. 3) takes place in four stages, corresponding with the maximum number of digits in the product,' under the control of a multicontact stepping switch. At each stage a circuit is completed through the switches set up by the record card readings and circuits representing the appropriate partial products. The partial products a x b are selected from representations of four Pythagorean tables (Fig. 1, not shown) the units digit by adding the digits obtained from two units tables and the tens digit similarly from two tens tables. This splitting of the tables reduces the number of digits represented and, hence, the number of switches in the apparatus. Thus, digits 0, 1 and 0, 2, 4, 6, 8 occur in the two units tables, respectively, and 0, 1, 2, and 0, 3, 6, in the two tens tables. By grouping the factors a in pairs differing by 5 another set of tables (Fig. 2, not shown) is which leads to a further reduction in the number of switches required. Transfer is accomplished by using two sets of selective adding circuits, one operating on a decimal base of notation and the other on the hedecadal system (base 11). The difference between result digits obtained in the two notations provides the tens transfer digit, as described in Specification 483,906 which is referred to. The four digits of the term D are set up in a relay-type register (Fig. 9, not shown) from the record card. In the course of a computation #AB the relay register may be used for registering the progressive total. When the capacity of the machine is exceeded (i.e. when #AB+D = 10,000 or over) a relay is energized to operate a warning signal. Other safety devices are also described. In a modification the stepping switches are replaced by chains of relays.