| 摘要 |
1,219,187. Programmed control of machine tools. PRATT & WHITNEY Inc. March 18, 1968 [April 17, 1967], No.13013/68. Heading G3N. A numerical control system for moving a plurality of parts along reference ares between existing points and desired points comprises means for moving each part at a rate determined by its position error and the largest position error. As shown, in a numerical control system for positioning parts 22-24 of a machine tool by pulseresponsive servos 33-35, movement commands are read from a tape 37 by a reader 38 decoded in a decoder 39 and changed into binary form by a translator 40. The reader, decoder and translator are controlled by a tape reader control 41 for advancing the tape. The commanded end ofspan positions are fed from the translator to axes command registers 43-45 and an arc command register 46, for circular interpolation. Movement of a machine part is determined by pulses from electro-optical quantizers which are fed to axes actual position registers 51-53. The command and position registers are recirculating registers containing the same number of bits, one of which indicates sign. The instantaneous commanded positions of the parts are held in registers 54- 56 the contents of which are compared with the contents of axes command registers 43-45 by comparators 57-59. The outputs of the comparators, which are indicative of the existing errors between the instantaneous commanded positions and the commanded end positions are applied to interpolators 60-62 providing pulses to the servos 33-35. A clock pulse generator 64 feeds a binary frequency generator 65 providing a plurality of pulse trains to a feedrate adjust control 66, set manually, for a pulse generator 67 comprising a multiplier responsive to a programmed number, from the translator 40, stored in a register 68. The feedrate pulse frequency from the generator 67 is applied to a control network 80 which determines the rate at which command pulses are iterated by interpolators 60-62. Signals indicative of the interpolator output pulses are selectively fed to the control 80 by a function gate 88 enabled by an error magnitude control 93 determining the largest of the instantaneous position errors from comparators 57-59. During interpolation the existing position error is determined with respect to the commanded end position and pulse rates produced corresponding to such error. For a linear interpolation, first order equations are derived, for two or three axes, determining the resultant vector in dependence on the axes errors from the commanded positions. For circular interpolation, the end points are read into registers 43, 44 and the co-ordinates of the centre-point of the span read into registers 45, 46. |
