DE2127731A1 - Position controller - Google Patents

Description

PATENT LAWYERS -.

Dr. rer. nat. DIETER LOUIS Z I 2 / / 0 I.

Dlpl.-Phys. CLAUS PÖHLAl /
Dipl.-Ing. FRANZ LOHRENTZ

8500 NORNBERQ 12 026 30 / h

KESSLERPLATZ I

CIIiCINNATI MILACRON INC., Cincinnati / Ohio, USA

Position controller

The invention "relates to a position control system of US Pat Type "in which a motor unit responds to command pulses which are issued by a command circuit, the motor unit executes a displacement or feed step of a fixed size with each command pulse, which is why the feed rate and accelerating the motor unit according to the command pulse sequence or frequency and the rate of change corresponds to the frequency. In particular, the invention is concerned with a position controller a motor unit for moving an associated output member to a desired position, with a source for a first analog signal continuously representing the distance to be covered to the desired position so / ie with a command circuit, which issues command pulses for the motor circuit, whereby the frequency of the pulses is added first.

109850/1365

and then decreases, so that the output member of the motor unit is accelerated and decelerated accordingly, so as to achieve the desired Position to achieve. The motor unit. Can be a stepper motor, another type of electric motor or a include electro-hydraulic unit.

Although the device according to the invention is particularly advantageous can be used in the multi-axial control of machine tools, it is easy to see that the motor unit a.Also can be used to drive any given load and the invention can also be used for uniaxial operation can be used.

An almost universal requirement in numerical control systems is that the feed rate and, consequently, the frequency of the command pulses should always be within the limits given by the curve shown in FIG. 1 of the accompanying drawing. First, the speed may have a maximum! Speed Yy. do not exceed. Second, the acceleration and deceleration must not exceed values determined by the inclination of the parts of the curve _{indicated by A M} and D ^, at least above a speed Vj to which the acceleration or from which the deceleration is substantial Can be carried out instantly «These limits are determined taking into account the need for a slip in electric motors

109850/1365 ^

OfHOtNAL

to avoid, the maximum allowable tool speed for a machining pass, etc. The particular determining conditions ■ vary from system to system, but the shape of the curve is common to most systems. V ™ can easily be changed, for example when it is controlled by an operator or by a program. Likewise, Α .., and D .. can sometimes be variable. A ^ and D ™ are usually, but not necessarily, the same size.

If there is a curve with the shape according to FIG. 1, there is the problem of deciding when the deceleration should be initiated if the predetermined end position must not be exceeded. In any case, the speed must reach the value V ₁ just before the desired position; the motor unit then crawls into the end position. Some known systems set a certain braking distance from the end position and initiate the deceleration when only this distance has to be covered. This deceleration distance is made large enough to allow enough time even under the most unfavorable conditions. Although such an approach is safe, it means that the last part of most operations is carried out with prolonged creeping, which is time-consuming and irritates the operator. Short work steps are carried out completely in creep speed four den.

It has also already been suggested that several (e.g. 3)

109850/1365

Provide different stopping distances, which are 5 different Corresponding speed ranges, with the stopping distance in each given instant in accordance is selected at the current speed. This of course reduces the degree of Creep, however, by no means turns it off completely.

It has also already been proposed to initiate the delay, as soon as the distance to be covered decreases to a value proportional to the current speed. However, it will be shown below that this measure does not limit creep to a minimum.

The aim of the invention is to create a simple analog circuit which makes it possible to limit the duration of the creep to a very small value, the exact stopping distance being selected as a practically continuous puncture of the present speed. If desired, the circle allows V _{M to be changed} during an operation without any detrimental effect. The circle will always keep the time required for an operation close to the theoretical minimum and, with the exception of a tiny period of time, will avoid creeping, which cannot be completely eliminated in any practically feasible system.

To solve this problem, according to the invention, a control

109850/1365

device of the type mentioned proposed which is characterized by a circuit for deriving a second analog signal that is non-linearly at least approximately in accordance with the frequency of the command pulses changes with the square of the frequency, as well as by a comparator, which compares the two analog signals and generates a delay signal which influences the command circuit in such a way that the frequency of the command pulses decreases, when the two analog signals reach predetermined mutual values.

As will be explained below, the two linear sections in which the amplitude can be a function of the frequency illustrative curve can be approximated as following a quadratic law.

The command circuit preferably has a voltage-controlled oscillator and the control voltage is supplied by a time-constant circuit which determines the acceleration and deceleration speed. It is possible to let the control speed build up in the absence of the deceleration signal (positive or negative, depending on what leads to an acceleration of the oscillator) until it sioh. decreases at a value, preferably an adjustable value, which determines the maximum command pulse frequency. If the arrangement is such that the delay signal is overdriven,

109850/1365

when the frequency drops to the initial value mentioned, will a state of dynamic equilibrium exist at this value as long as the delay signal is present.

Further features, details and advantages of the invention follow from the following description of a preferred embodiment with reference to the drawings and from the Subclaims. Show it:

FIG. 1 shows the curve already mentioned at the beginning;

Figures IA and IB

further explanatory curves;

Figure 2 is a block diagram of the main features of an embodiment form of invention;

FIG. 3 further parts of this embodiment;

FIG. 4 shows the characteristic of an integrating amplifier used in FIG. 2, and FIG

FIG. 5 shows a circuit diagram of the voltage-controlled oscillator (VCO) of FIG.

As already mentioned, Fig. 1 shows the dependence of the

109850 / 1.365

speed of time. If we only turn our attention to the last part of this curve, where the delay occurs, and the speed to reach the position

5, we get that in Pig. Curve 2 shown in IA. This is a parabolic curve or curve following a quadratic equation. V / e used this quadratic dependency the invention a dependency according to a quadratic equation between the distance still to be covered and the speed, as indicated by curve 4 in Pig. IB shown is. For example, if the distance to be traveled decreases at speed V ^ -, as shown by the line

6 indicated, the delay signal is given at point 7. However, if the distance to be traveled is a

lower speed Vj, as indicated by line 8 reduced, the delay signal is given at point 9.

If a quadratic equation following curve 4 is replaced by a linear relationship 5, so ™ is the delay signal .am point 7 ^1, where in the first Pall V, which is not removed from the end position considerably further than the point 7. In the second Pail, ie at the speed V ^, however, the deceleration ^{signal is given at point 9 1} , which is significantly farther away from the end position than point 9. As a result, the amount of creep increases significantly, as shown by curve 3 in Fig. 1A , and the time

109850/1365

which is necessary for ending the delay increases, as indicated in FIG. 1A, by the difference between t ₁ and t ₂ . Such a time difference, when summed up over a sequence of setting movements, leads to a considerable reduction in performance in a system in which the linear dependence 5 der.Pig. IB is used.

In the illustrated embodiment of the invention, the numerical control is carried out in two coordinate directions, namely the X and Y directions. The actual values of X and Y are stored in registers IO and 11 (Fig. 2). The final values X ¹ and Y ¹ are initially entered into registers 12 and 13, for example from a tape program. The subtractions X'-X and Y ¹ -Y are then carried out on elements that are not part of the present invention, so that the differences 4x and iY remain in registers 12 and 13. Command pulses are now emitted by an oscillator 14 to fill the X and Y registers, whereby the ^ X and ^ Y registers are emptied at the same time (X and Y are apparently increased and accordingly ^ X and ßj decreased). For the sake of simplicity, the present description does not take into account the fact that the quantities can in reality also be negative. The sign can be taken into account in any known manner. Digital comparators 15 and 16 determine ßX = O and ÄY =, 0 and

109850/1365

the signals / \ X = O and ΔΥ = O are used to open gates 17 and 1.8, so that as soon as jdx or ÄY reaches the value 0, the corresponding gate closes and further filling or emptying, as mentioned above, not taking place. The gates 17 and 18 also require a signal i ¹ which controls them in the operating state and which is only emitted when Ax and AY have been entered in the registers and the system is ready to carry out the programmed operation.

Command pulses passed by gates 17 and 18 control X and Y motor units 19 and 20, respectively. These can be stepper motors, _{Vj corresponding to a command pulse train of 300 Hz and V M} a pulse train up to 5 kHz.

In the embodiment shown, a single oscillator supplies the command pulses for the X and Y directions. From this it follows that when operations are carried out in two coordinate directions, the motor units accelerate with one another and can then be delayed together in order to complete the shortest work step (ÄX or Δυ) in each case. This is just a matter of a unit to complete the longer work step can be accelerated and decelerated again. However, it would be taken for granted possible, if desired, to provide two completely separate systems for the two axes, in particular

109850/1365

- ίο -

have two voltage controlled oscillators.

In the present Pail, however, we only have a single oscillator 14. This is, as will be explained in more detail below, constructed so that, in the absence of a delay signal DEO, it accelerates up to a frequency which is determined by a control voltage V ^ which the frequency V _M in Pig. I.

corresponds, while in the presence of a delay signal DEC a delay to Vj takes place. In addition, guaranteed a built-in time-constant circle that follows the inclinations Ajyj- and Dj, It's supposed to be the way that Signal DEC obtained will be explained.

The values of Λχ and ΔΥ are alternately sensed to determine whether any of them are so small that the delay should be initiated. It is not necessary, the most essential Units (bits) of Ax and ÄY to be checked, as for such large values a delay is never required. In addition, it is possible (although not necessary) to do the least disregard significant information values (bits) as less accuracy is required when it comes to it is about determining when to delay than the actual attitude control. In the present embodiment X, Υ, ΑΧ and Αϊ are 16-digit (16-bit) numbers. However, only the 8th-12th Digit (from least significant end) used in the circle to be described below. From-

109850/1365

- li -

selected 5 hits from ÄX and AY should be designated with As ".

A time circuit 25 transfers As into a register 21 alternately von Ax and Ay. Only as an example it should be mentioned that a full test cycle can take 640 / us, with such cycles can be applied to X and Y alternately. In contrast, registers 10, 12, 11, and 13 can circulate scrolling registers (recirculating shift registers) that work with the assigned computing units in 10 / us cycles. The five bits of AES can then be entered one after the other in the corresponding order in the register 21 starting from Ax or AY, where can be proceeded in a known manner. In the same way, parallel transmission between static registers Find use.

The five stages of the As register are connected to a digital-to-analog converter, which consists of resistors 22 and an output amplifier 23, the output signal V "of which corresponds to the distance still to be covered (in the coordinate direction just checked). The command pulses from the oscillator are applied to a non-linear integrating amplifier 24, the output signal V ™ of which is approximately proportional to the square of the feed rate. V ^ is passed to a comparator 26, the output signal OVD of which indicates that the output signal of the integrator falls below a reference level V ₁ and that, as a result, the oscillator falls to a frequency below 300 Hz

109850/1365

decreases and is used as will be described below " to pull the frequency back up.

Vg (distance still to be covered) and Vj, (square of the feed rate) are fed to a differential amplifier 28 via resistors 27 of suitable mutual dimensioning, whose output signal SD. indicates that Vg-V ™ on UuIl or another threshold level has fallen at which the delay must be initiated.

Referring to Mg. 3, the signal DEG is derived from SD as follows. Two D flip-flops 30 and 31 monitor the delay signal SD with respect to X and Y, respectively. If both flip-flops are in the state Q = I, an OR gate 32 outputs the signal DEC, which also in the presence of F, the inverse of the The aforementioned feed signal F, is output in order to slow down the oscillator during the waiting time before a work cycle is carried out. The signals AX = 0 and AY = O are applied to reset inputs B. of the flip-flops 30 and ^ 31 in order to prevent the oscillator from being delayed simply because either Az or AY was originally set to Hull or has reached this value.

! The delay signal SB is sent to the D inputs of both plipflops 30 and 31 given. However, the flip-flops only change their current state if they are controlled via the C input will. The flip-flops are fed from the X and Y directions

109850/1365

ordered, controlled by the time unit 25 arriving pulses towards the end of the 640 / us period. In this way, at the beginning of an X cycle, the five bits of interest (information units) of Ax are entered into register 21 as S. After a period that enables the signal S \ D to develop, the X-direction pulse controls the flip-flop 50 only when the signal Έ is present at the gate 54. During the next cycle, As is taken from Ay and it then controls the Y-direction pulse the flip-flop 51, if F is present, and it opens a gate 55. Furthermore, a gate 56 is provided, which allows the signal DEC to pass only in the absence of OVD.

The following procedure therefore results. Originally is 3? available and therefore DEO appears. The oscillator 14 slows down to Vj and OVD appears. Gate 56 closes and DEC disappears. Now the oscillator runs up slowly, which causes the OVD to disappear again. The gate 56 opens and the signal DEC is given back to the oscillator. This results in a state of dynamic equilibrium in which the oscillator frequency fluctuates around V ₁ .

When the signal i ^{1 is} applied and f disappears, DEC also disappears. None of the plip-flops 50 or 51 has the state Q = I, since either SD does not exist in the corresponding cycle or AX = 0 or AY = 0. 3? For the following consideration it is assumed that the initial condition

109850/1365

is present. As a result, the oscillator runs up and reaches the value V ^ or a lower value, if the operation is only short. Under certain circumstances SD appears in an X cycle and switches the flip-flop 30. DEC appears again, whereupon the oscillator frequency falls until the value V ₁ and 'the equilibrium state described above is reached. The motor units 3.9 and 20 crawl with Vy to ^ X = 0 when the flip-flop 30 is reset to Q = O. DEO disappears again and the oscillator accelerates again to complete the Y pre-stroke or operation. If SD appears in a Y cycle, then flip-flop 31 is toggled and DEC appears. As soon as V γ is reached, the system crawls until AY = O applies, the flip-flop 31 is reset and the operation is completed.

Before the oscillator 10 is now described in more detail, we return to FIG. 2, from which it can be seen that the integrating amplifier 24 has a positive feedback path which is formed by three diodes 29. When V ^ ₁ exceeds the forward voltage drop of these diodes, the positive feedback path causes the rate at which V ™ increases with frequency to increase. The integrating amplifier therefore has a characteristic curve corresponding to FIG. 4, in which V ^ ₁ is plotted against the command pulse frequency. The characteristic consists of two linear sections, the second of which has the greater gradient. Fig. 4 leaves

109850/1365

realize that the characteristic curve thus obtained is surprisingly good follows a quadratic law that is dashed

.is shown in Fig. 4. Tests have shown that this approximation actually gives very good results with very little creep, regardless of the value of V- ,, at where the delay begins.

The oscillator (also called VCO) 14 is stronger in FIG. 5 shown in detail. The VCO has emitter-coupled Transistors 40, where the size and phase position of the Output via a common capacitor 33 and independently on the one hand via the resistor 41 and on the other hand via the resistor 42, to which the transistor 43 and the resistor 44 are connected in parallel, can be determined. The resistance 41 is a pest value, so that there is a fixed pulse width for the command pulses. Because the conductivity of the transistor 43 via the control voltage V " is changeable, the oscillator frequency can be changed. The output signal of the oscillator is via a transistor 46 buffered, the duroh diodes 47 and 48 is limited, and is then available on a line 49, which also in Pig. 2 is so designated.

The control voltage Y " is generated by the voltage across a capacitance 50 (for example 50 / Ul ¹ ) which, in conjunction with a resistor 51 (for example 33kXl), has the required time constant

109850/1365

for the slopes V ^ and V ₀ (Fig. 1) results. In the absence of a DEC signal, the capacitance 50 discharges from the + IOV line to the -10V line via the series resistors 51 and 52, the former being significantly larger than the latter. However, the discharge is stopped at a potential corresponding to the value V ^ due to the emitter-base effect of a transistor 53, where Vj, is somewhere in the range of 5V-r 10V. The end potential of -10V is sufficient compared to this range to result in a sufficiently linear section V «corresponding to V.

The capture potential V ™ at the base of the transistor 53 is derived from a manually adjustable potentiometer 55 in the illustrated embodiment. In another embodiment, however, it could also be derived from a program carrier via a digital-to-analog converter. A constant-time circle 56 prevents V _M from changing too suddenly.

If the delay signal DEC is applied, a transistor arrives 57 in oil-conducting state, whereby another Traneistor i? 8 becomes conductive and the capacitor 50 is approximately linear '' a directional potential of +24 V is charged. However, the potential Vq does not reach 10 V because of the above Effect of the OVD signal interrupting the DEG like this

109850/1365 i

is required in order _{to keep V 0} at a value which _{corresponds to the aforementioned frequency V T} (the higher positive values V _n assumes, the less the transistor 43 conducts and the lower the oscillator frequency becomes). A catch diode 62 connected to the +10 V line supports this limitation in the case of positive values of V _n .

The oscillator also has two overshoot devices. When used in machine tools, V _{w will} normally be adapted to a machining operation and setting movements can be carried out much more quickly. For this purpose, an inverse EASCH signal, ie EASCH, switches the transistor 59 into the blocking state and the transistor 60 conductive, if a rapid movement is desired, which leads to a decrease in the base potential V ^ of the transistor 53 '. V _c now drops in order to increase the oscillator frequency.

The second device is used if the motor unit has an optionally operable frequency divider or several frequency dividers (eg 1:10 and 1: 100) for dividing the command pulses in order to obtain slow and very slow feed rates in a known manner. When switching from a feed rate range to a higher range, it is necessary to switch to a lower value of · V _{1 with} respect to the oscillator frequency, with such a switch only as described above

109850/1365

can take place at a speed determined by the time-constant circuit 50, 51. However, if a FAST DEC signal is applied, if such a feed rate changeover is carried out, the transistors 61 and 63 become conductive and Y _Q is pulled up to +24 V via a resistor 64 which is significantly smaller than the resistor 51.

For the sake of simplicity, a 2-coordinate or -axis system was used described. An extension to three or more axes is easily possible for the Pachmann. As mentioned above, each axis could also have its own control system instead of sharing a single control system. The invention is also applicable to systems of training in which the speeds along an axis are proportional to be maintained at the desired displacement along this axis, thereby eliminating the machining operations along all of them Axes are extended over the same time interval. In such a system, the deceleration control would be optional apply to the axis to which the greatest displacement has been transferred.

109850/1365

Claims (8)

Claims 1. Position control device with a motor unit for moving an associated output member into a desired one Position, with a source for a first, continuously going back to the desired position Analog signal representing the distance to be laid as well as with a command circle, which command pulses for outputs the motor circuit, the frequency of the pulses first increasing and then decreasing, so that the output member of the Motor unit is accelerated and decelerated accordingly in order to achieve the desired position, marked by a circuit (24) for deriving a second analog signal which is non-linear with frequency the command pulse changes at least approximately in accordance with the square of the frequency, as well as by a comparator (28) which compares the two analog signals and generates a delay signal which the Command circuit influenced in such a way that the frequency of the command pulses decreases when the two analog signals are predetermined achieve mutual values. 109850/1365 2. Control device according to claim 1, characterized in that that the amplitude of the second signal depends on the command pulse frequency according to a curve that consists of consists of a plurality of linear sections, e.g. two linear sections through which a quadratic law is approximated. 3. Control device according to claim 1, characterized in that the circuit for deriving the second analog signal has a non-linear integrator (24, 29). 4. Control device according to claim 3, characterized in that the integrator has an amplifier (24) with capacitive Having feedback and a further positive feedback path with diodes (29), the further path being effective as soon as the output signal of the integrator exceeds a predetermined level. 5. Control device according to one or more of the preceding claims, characterized in that a digital register (21) which is connected to a digital-to-analog converter (22, 23) is used as the source for the first analog signal. 6. Control device according to one or more of the preceding claims, characterized in that the loading 109850/1365 error circuit a voltage controlled oscillator (14) and has a time-constant circuit (50, 51) for outputting a control voltage for the oscillator in order to increase the acceleration and pre-determine deceleration rate. 7. Control device according to claim 6, characterized by a circuit (55) for generating an adjustable 3? angvoltage level for limiting the control voltage to a value that determines the maximum command pulse frequency. 8. Control device according to one or more of the preceding claims, characterized by a circuit (26, 36) for overriding (switching off) the delay signal as soon as the command pulse frequency has reached a preselected one Value goes down. 109850/1365