DE1200417C2 - PROCEDURE AND DEVICE FOR ROUTING ERROR CORRECTION IN NUMERICAL CONTROLS - Google Patents

Description

FEDERAL REPUBLIC OF GERMANY

Int. CL:

G 05 b. 19 / 20-

GERMAN

PATENT OFFICE

PATENT LETTERING

German KL: 42 rl, 19/20

Number: File number: Filing date: Display date:

P. 12 00 417.0-32 (S 92730) '
August 20, 1964 September 9, 1965

AusnaK-im ;: January 17, 1974

l'niciiisi'hcifl wckhl \ on da Auslegest,:! «:; ·, from

In numerical machine controls, a distinction is made between absolute and incremental Setpoint specification. With an absolute setpoint specification, the machine commands for moving to the desired position are always set to a specific one Fixed point, while the machine commands are always used with the incremental setpoint specification starts from the position reached in the previous program step. If you put some tolerance for the retraction of the machine part to be moved into the specified position is based on the latter type of control the risk that after several program steps this for the individual program step still acceptable inaccuracies to a no longer acceptable overall add up errors. In principle, three types of errors can occur in a program step: The machine the machine can exceed the prescribed setpoint position due to its inertial mass does not reach the specified setpoint position due to premature braking, and it can temporary machine vibrations occur in such a way that the direction of movement of the machine temporarily deviates from the programmed direction.

As a rule, a digital binary counter is used as the setpoint generator in numerical controls, into which the desired setpoint position is entered as a counter reading, this counter reading being reduced by distance pulses supplied by an actual value transmitter. The count zero then serves as a criterion that the desired setpoint position has been reached by the machine. The aforementioned path errors in the individual program steps are usually recorded by a bidirectional path counter, so that the path covered too much or too little in the individual program step can easily be taken into account in the next program step. A bidirectional counter, however, practically doubles the effort compared to a counter that only needs to count in one direction, which is important in that the capacity of the counter serving as a setpoint position memory must be designed for the maximum distance that can be predetermined by the machine.

The invention is based on a known method to correct an in the course of or after the completion of a single program step numerical machine controls occurring, digitally recorded path errors by means of one in two Directions working correction counter, which before the next program step a path error Method and device for correcting path errors in numerical controls

Patented for:

Siemens AG, 1000 Berlin and 8000 Munich

Named as inventor:

Dellert, German, 8520 Erlangen;

Kubatsch, Roland, 8500 Nuremberg

takes on the corresponding number of pulses and the count of which is counted down to zero, with one that can be counted empty by path pulses, the target-Tst difference setpoint position memory feeding the feed drive, its content by a corresponding amount after a path error has occurred can be corrected and with a pulse generator, the correction counter and the setpoint position memory Can supply impulses. Such a procedure is in the dissertation of, K. Jüstel »About Structure and operational safety of numerical point controls ", Technical University of Aachen, 1963, described. It is true that with this method the effort for the setpoint position memory can be reduced reduced. than not all, depending on the extent of the expected machine vibrations Counting loads, but z. B. only one counting decade for both upward and downward counting need to be upgraded. Since the two other types of errors mentioned above are basically the same Are of the order of magnitude, a corresponding number in front and in front of the correction counter must also be provided down counting decades.

The invention has the task of counting up and down through the total of the existing To halve the effort involved in counting steps compared to the known method. This is resolved Task through the entirety of the following features: a) In the event of machine vibrations, i.e. when they occur of travel impulses in the wrong direction, the travel impulses are processed by an evaluation circuit, which detects the occurrence of travel impulses in the wrong direction, to a setpoint value

309 683/393

Down counter used for position memory is blocked and at the same time released to the correction counter to first read the content of the correction counter to increase and count it empty again after another change of direction, whereupon the path pulses are fed back to the down counter alone;

b) In the event of an overflow, the excess displacement pulses are saved in the correction counter as soon as the down counter has been counted empty, for which purpose an overflow circuit releases the path pulses for the correction counter;

c) if the setpoint position is not reached, the Pulse generator, depending on the standstill of the position pulses and a still existing signal * 5 of the down counter, pulses to both counters until the down counter is empty, whereby the path error is stored in the correction counter;

d) when adding the path error stored in the correction counter, depending on the direction of the next program step, the path impulses of the next program step are initially only fed to the correction counter until it is empty, and then depending on the Signal that the correction counter has been counted empty, only the down counter;

e) when subtracting the one saved in the Korreklurzäliler Path error, depending on the direction of the next program step, is the Feed drive generating path pulses is initially blocked, then pulses from the pulse generator fed to both counters, the content of the down counter decreased for so long until the correction counter is empty, and then the feed drive is released.

The invention and its embodiments are illustrated below with reference to the figures will.

In Fig. 1 the schematic diagram for the control is drawn in a coordinate of a working machine. Before the beginning of each program step, a suitable information storage device 1, in the example shown, a punched tape, enters the desired value step according to size and direction into a decoding device 2. The amount of the predetermined displacement PS of the machine part to be positioned relative to the position Pn reached in the previous program step is entered as a corresponding counter reading into the setpoint position memory labeled P. A signal ZS ₁ then appears at an output of the position memory P , as a criterion for the fact that the counter reading of the position memory is not equal to zero. This signal can be used to set the feed motor for the corresponding coordinate in motion, whereupon a digital actual value transmitter influenced by the moving machine part, for example a magnetic disk 3 to which a Hall generator 3c / is assigned, one pulse / " , delivers. The pulses / "pass an evaluation circuit 4 and pass through one of the gate circuits T ₁ to T ₄ either to the position memory P or to the correction counter K. While the position memory P is only intended for counting in the reverse direction, every pulse arriving at its input thus reducing its count, the correction counter K allows counting in both directions, so that the pulses arriving at its input either increase or decrease its count in accordance with a signal at its + and - marked inputs. Normally, the distance pulses I ₁₀ pass through the gate T ₁ to the position memory P. In a direction indicated by 5 evaluation circuit whose operation will be described in more detail later, be programmed coordinate direction (K +, K-) and motion 'direction of feed (B j, B) , the first-mentioned signals of the decoding device 2, the last-mentioned signals being taken from the evaluation circuit 4. If in the course of the program step the programmed coordinate direction and the actual direction of movement of the feed do not match, this is an indication that a so-called oscillating movement of the machine part is occurring. Then the signal R appears at the output of the evaluation circuit S, whereby a blocking of the gate T ₁ and at the same time an opening of the gate T _{1 is} effected via the negation stage 11. The actual value pulses I _w then run on the correction counter K for the duration of the return. If the direction of rotation of the pulse disk 3 and thus of the machine part to be moved is subsequently correct, the correction counter is counted empty by the path pulses I _10. If the counter reading of the correction counter K has reached zero, which can be evaluated by the signal ZSk , then the position memory P is released again for empty counting by the pulses I ₁₀ by means of the gate circuit T ₁ .

If, towards the end of a program step, the machine part to be moved overflows beyond the programmed setpoint position, this manifests itself in the fact that the counter reading of the position memory P is zero and still is. Pulses I ₁₀ occur, which could not cause any further change in the count of the position memory P. In this case, the signal U appears at the output of the overflow circuit labeled 6, which opens the gate circuit T ₃ and allows the path pulses I ₁₀ that still occur to reach the correction counter input, whereby its counter reading is increased from zero by an amount that corresponds to the overflow path is proportional. As a result, this path error is saved as a counter reading in the correction counter.

If, after the end of a program step, the machine part to be moved does not reach the specified setpoint position, this operating state is characterized by the standstill signal St and the signal ZSp, which indicates a counter content. With these two signals, the signal N is generated by a further evaluation circuit 7, which opens a gate circuit T _6. Then pulses h from a freely oscillating pulse generator 8 can reach both the correction counter K and the position memory P. The pulses h count the position memory P empty, while they increase the count of the correction counter K from zero. When the counter reading of the position memory P reaches zero, which manifests itself in the disappearance of the signal ZSp, the signal N at the output of the evaluation circuit 7 disappears, and the pulses h are blocked by the gate circuit T _9. Again, the path error caused by the overrun is as

Corresponding counter reading in the correction counter K.

As shown in the foregoing, the path error resulting therefrom is transferred to the correction counter J ^ in the event of an overflow as well as an overflow. Depending on whether a movement in the same coordinate direction or in the opposite coordinate direction is provided for the subsequent program step, an addition or a subtraction of the new setpoint value supplied by the information storage device for the distance pulses I _{w has} to take place to take account of these path errors. For this purpose, two evaluation circuits labeled 9 and 10 are provided, each of which has stored the old programmed coordinate direction and compares it with the new and with the type of error that has arisen (overflow or overrun). If an addition is to take place, the signal A arises at the output of the evaluation circuit 9, which opens the gate circuit J ₂ , so that the path pulses I _w initially count the correction counter K empty; if its counter reading reaches zero, the gate circuit T _{2 is} blocked and the gate circuit T _{1 is opened} for the distance pulses I _w , so that the normal empty counting of the counter P takes place again. The addition takes place in that the actual value pulses I _w do not initially lower the count of the position memory, but must first count that of the correction counter K empty.

If the consideration of the error that occurred in the previous program step requires a subtraction from the new nominal position value, then the signal S appears at the output of the evaluation circuit 10, which opens the gate circuit T _5. The pulses h supplied by the pulse generator 8 now arrive simultaneously at the inputs of the correction counter K and the position memory P and lower the counter reading of the latter until the counter reading of the correction counter K is zero. At the same time, care must of course be taken to ensure that the feed drive is blocked for the process of blanking the correction counter K , so that no displacement pulses I _{w are} generated. The evaluation circuits 5 to 10 can be implemented using methods of logical linking technology known per se using commercially available modules. In this respect, there is no need to go into more detail about their internal mode of action. It is advantageous here in the interest of a small number of types, predominantly building blocks of one type, e.g. B. so-called Nor gate to use.

In Fig. 2, for example, the internal structure of the evaluation circuit denoted by 5 is shown. A signal should appear at its output R when the machine part to be moved oscillates, that is, a brief, unintentional reversal of direction of movement occurs. For this purpose, six Nor gates labeled! 2 to 17 are provided. A NOR gate is a logic module, at the output of which a signal only appears if there is no signal at any of its inputs. In the following, the absence of a signal with a 0-signal and the presence of a signal with an i-signal will be referred to. The NOR gates 12 and 13 are interconnected in a manner known per se by mutual, cross-feedback feedback to form a bistable multivibrator. If a coordinate direction in the negative direction has been programmed for the respective program step , the signal marked K- should be a so-called 0 signal. The same applies to the direction of movement of the machine part indicating signals B and B + as well as for the count of the counter K correction signal representing ZS _k. _.

If the programmed coordinate direction and the actual direction of movement do not match, then at an output of the Nor gate 16 or 17 a!, - signal and at both outputs of the Nor gates 14 and

ίο 15 a 0 signal occurs. Since so that all the inputs of the NOR gate 12 are filled with 0-signals, a /.-Signal, which can be interpreted as an indication of developing machine vibrations appears at the terminal designated R. The acquisition of the signals labeled B + and B- , which characterize the actual direction of movement of the machine part, can be obtained with the arrangement according to FIG. 1 from the direction of rotation of the machine coupled to the machine part to be moved

ao pulse disk 3 can be obtained.

In Fig. FIG. 3 shows an example of how the signal St is processed in the absence of any path pulses I _{w supplied by the pulse disk 3.} The in F i g. 3 is part of the evaluation circuit 4 in FIG. 1. It in turn consists of two mutually cross-coupled back, working as a bistable multivibrator Nor gates Ii and 19, which two AND gates 20 and 21 are connected upstream. These are through a differentiating capacitor 22 of the. Distance pulses I _w applied. The rising edges of the displacement pulses I _{w are} suppressed by a discharge diode 23. The output of the nor gate 19 is connected to the input of the AND gate 21, and that of the nor gate 18 is connected to the input of the AND gate 20. In this way it is achieved that the actual value pulses / «, ·, which are made up of the nor-gates Ii and 1 © /, which are applied to the capacitor 22, alternately toggle into their two positions. The monostable multivibrators 24 and 25 connected to their output are triggered periodically, and.

their breakover times are chosen in such a way that their output L signals overlap each other in time.

The method according to the invention also opens up the possibility of controlling in several coordinates get by with a single position memory, which is then successively on each to be controlled Coordinate is switched. With such a

■. Switching of the position memory is always the responsibility of everyone

Correction counter assigned to the coordinate with the corresponding actual value transmitter connected; so that any z. B. by unintentional vibrations path impulses caused by the shutdown machine part are not lost, but as error

pulses stored in the correction counter and then from the next in the manner previously described Program step can be taken into account.

Claims (3)

Patent claims: I. Procedure for correcting an in the course of or after the completion of an individual program step with numerical machine controls occurring, digitally recorded path error by means of a correction counter that works in two directions and that is set before the next program step takes on a number of pulses corresponding to the path error and then its counter reading counted down to zero depending on the direction of the next program step is fed the target / actual difference to the feed drive with an empty countable by distance pulses Setpoint position memory, the content of which increases by a corresponding one after a path error has occurred Amount can be corrected, and with a pulse generator, the correction counter and can supply pulses to the setpoint position memory, characterized by the entirety following features: a) In the event of machine vibrations, i.e. if travel pulses (/ ".) in the wrong direction occur, the travel impulses are blocked by an evaluation circuit (5) that detects the occurrence of travel impulses in the wrong direction, to a down counter (P) serving as a setpoint position memory and at the same time released to the correction counter (K) in order to first increase the content of the correction counter and to count it empty again after a renewed change of direction, whereupon the displacement pulses are fed back to the down counter alone; b) In the event of an overflow, the excess displacement pulses are saved in the correction counter as soon as the down counter is empty, for which purpose an overflow circuit (6) controls the path pulses enables the correction counter; c) If the setpoint position is not reached, the pulse generator (8) sends pulses (7 _λ ,) to both counters (K, P) , depending on the standstill (Si) of the path pulse and a signal (ZS ₁₁ ) still present from the down counter, until the down counter is empty, whereby the path error is stored in the correction counter; d) when adding the path error stored in the correction counter, depending on the direction of the next program step, the path pulses of the next program step are initially only sent to the correction counter. fed until it is counted, and then, depending on the signal (ZS _k ) that the correction counter has been counted empty, only to the down counter; e) when subtracting the path error stored in the correction counter from the direction of the next program step, the feed drive generating the path pulses is initially blocked, then pulses (/,.) from the pulse generator (8) are both counters (K. P) supplied, the content of the down counter being decreased until the correction counter is empty, and then the feed drive is released. 2. Device for performing the method according to claim 1, characterized in that the counting inputs of the down counter (P) and the correction counter (K) via logical Torbzw. Gate circuits (T ₁ to T _1. ) Are connected to the actual value transmitter (3) and the pulse generator (8), the gate circuits being open depending on the type of error and the direction of two successive program steps. 3. Device for performing the method according to claim 1 for generating the standstill signal the path impulse, characterized by one with two crosswise one below the other fed back Norgattern (18, 19) built bistable flip-flop, which by the displacement pulses (/ ".) Periodically alternating between their two positions is tilted and a monostable multivibrator (24, 25) is connected to each of their outputs, their combined output signals change when the two flip-flops are periodically flipped overlap (Fig. 3). 1 sheet of drawings