| 摘要 |
1,175,423. Programmed control of machine tools. CINCINNATI MILLING MACHINE CO. Jan. 5, 1967 [Jan. 17, 1966], No.809/67. Heading G3N. Apparatus for the automatic inter-polated control of a machine tool, wherein sets of signals define a unit vector, at each terminal point at the intersection of programmed path segments, are sealed by offset correction factors and algebraically summed with the co-ordinate dimensions of the terminal points to define points of inter-section of segments of a path parallel to the programmed path and spaced therefrom at a selected perpendicular distance. As shown, a numerical control system for the co-ordinate slides 15-17 of a milling machine is responsive to recorded information on a punched tape 18. The tape has eight channels on which a programme is recorded by a word-address code in which each block of information includes three dimensions in binary coded decimal form and a set of binary coded decimal compensation factors. The tape 18 is fed through a device 29 in which the code is read and converted into a binary form which is issued serially, least significant bit first. The cutter compensating values, set up manually on decimal switches in a unit 31, represent the difference between a nominal cutter diameter and the actual cutter diameter and are sampled and applied to the converter portion of device 29, a switch determining whether the actual value is larger or smaller than the nominal value. The binary coded cutter diameter difference value is than transmitted to a cutter compensation computer 32 where it is multiplied successively with the cutter compensation factor for each co-ordinate to produce three products, having a sign determined by the switch in unit 31, which are inserted into a differencing unit 30. The unit 30 derives new corrected co-ordinate positions which are compared with the previous read and corrected positions and binary difference signals generated. The difference signals are serially shifted out to a circuit 33 where they are normalized by eliminating the leading '0's in the most significant bits until the largest number has 1 in its most significant bit place. The normalized binary numbers are then transferred into circuits 34-36 which perform simultaneous inter-polations for each difference dimension. The inter-polation is effected by iterative additions, the carry or overflow signals being fed to accumulator circuits 37-39 where they are electrically combined with the previous accumulation to generate a number which proceeds from the point previously read on the tape to the next point. A sign signal from the circuit 30 determines whether the accumulation is additive or subtractive. Certain digits of the accumulated number are periodically and cyclically connected to a circuit 40 wherein they are multiplied by a fixed factor. The output of the multiplier is fed to a digital to analogue converter 41 changing the binary information to a voltage analogue signal impressed on an alternating signal carrier. These signals are fed to resolvers 42-44 connected to circuits 45-47, which are operated periodically and cyclically with the connection of accumulators 37-39 to multiplier 40, to sample and rectify the signals from the resolvers 42-44 and feed D.C. hold circuits 48-49 controlling, through amplifiers 51-53, motors 54-56 positioning slides 15-17. The tape may include.a programmed feedrate which is converted in the reader into a binary number which is conveyed to a feedrate control circuit 57 producing a train of gate pulses proportional to the programmed feedrate. The control circuit includes provision for an over-ride, that proportionally reduces the programmed feedrate. A train of gate pulses is derived by a computer circuit 58 such that the rate of the pulse train varies in accordance with the reciprocal of an approximation of the vector sum of the co-ordinate span lengths. The trains of pulses from circuits 57, 58 are fed to an inter-polator drive circuit 59 producing discrete blocks of pulses at a repetition rate that is a function of the two pulse trains. Each discrete block of pulses produces one serial addition in each of the inter-polator circuits 34-36. It is stated that the decimal switches 31 can be replaced by an arrangement responsive to a measurement of the strain on the cutter, whereby the programmed path position can be compensated for cutter deflection. |
