Positioning servo system

[Detailed Description of the Invention]

This invention relates to the high positioning servo system of Staffiness. The direct-current servo motor is widely used as the source of power in an industrial robot, NC machine tool 1, the terminal unit of information machines and equipment, and other automation apparatus. In these uses, thing , fixed firmly is required so that a hand, a tool, a printer head, a roller, etc. used as the load of a motor may be moved to the position of a wish at high speed, in the position of a wish, and it may be made to stop for a short time and after a stop may not deviate from the position. In order to satisfy these demands, the position speed servo which detects the position of load and speed and controls the current value which flows into a motor according to the result is used. Moreover, ' power cuts the present invention to the high position speed servo system of Staffiness, using a small servo motor. 1) Explain movement when C (direct current) servo motor and load are connected. Drawing 5 is a perspective view showing movement which connected Loading with servo motor M. DC servo motor M rotates according to the power committed on the current which flows into the magnetic field made with a permanent magnet. The armature current of servo motor M is set to ia. Torque which arises in servo motor M is set to LL. Since torque is proportional to current, it makes KT a proportionality factor, and it is torque (it can do [ writing it as t == KT ia (1), or ].). Surrounding moment of inertia of the load axis of load is set to JL. The motor has connected the slowdown machine with Load via Accelerator in many cases. It is considered as motor M and moment of inertia JM converted into circumferences of a load axis, such as a slowdown machine. The rotation position of the circumference of the axis of load is denoted by angle O. Let L be time. Equation of motion J=JM+J' It is (3). By giving armature current ia suitably, load rotation angle 0 is controllable. The command value of a rotation angle is set to OC. When command value OC is given, what kind of armature current ia should be sent through a To get closer sake for 0 at OC? This is a problem. The conventional positioning servo system is explained. Drawing 6 shows the lineblock diagram of a position servo control system. It is connected to Loading in the detector which detects a position or strange positions, such as an angle. here, since an angle is made an issue of as an example, angle detector A provides and comes out. The straight line displacement on the X-axis, gamma axis, and the Z-axis can be similarly dealt with by the present invention besides an angle. Servo drive circuit B receives signal 0 of angle detector A mosquitoes, carries out comparison with position command value OC, and determines armature current ia. Drawing 7 is a servo drive circuit block diagram concerning a conventional example. In order to determine armature current ia from current position signal 0 and position command value thetaC1, adding machines 1 and 2, multiplication machine 3.4, 5, and differentiation machine 6 grade are used. 10 11 L Mechanical coupling is shown. Multiplication machine 3 hangs (-1) on signal 0, and sets it to (-0). Adding machine 1 adds position command value thetaC (-0), and makes (Oc-theta). Differentiation machine 6 carries out time differentiation of the signal O. (d0/dt) is obtained. Multiplication machine 4 hangs (-Kv) on this. Kv calls it a speed gain. adding machine 26 -- ' which adds (Oc-CI) and -Kv (dO/dt). Multiplication machine 5 hangs KF on the above-mentioned result. It is called KFzeta Yo feedback gain. armature current ia comes out and is given. The first paragraph [ 2nd ] shows 3rd paragraph +1 resistance for stability. (1) (2) An equation of motion is obtained from (4) equations. however T2 -- ohm 0 -- [ (8) It comes out. Hereinafter, argument of the movement is carried out by the differential equation normalized in this way. the time of the signal which changes as an angle command value in the shape of [ as shown in Drawing 8 ] Step * Ump being given -- rotation angle 0 power f -- it is considered how it changes. The height of a step is given to H and a step signal with T=Q+. A formula (5) can be solved easily. They are what considered it as D Oohara meter and made time changes of theta the graph, or Drawing 9. They are a step response and Called about this. In Drawing 9, a vertical axis is (OAI) and a horizontal axis is omegaOL (=T). Parameters D which show the value of resistance are 0.5 and .0.707. The time of 1.0 is illustrated. coefficient-of-resistance Rika -- when small, while going too much greatly, returning from a desired value and A substitute(ing) vibration -- a desired value -- Come close. A desired value will not be exceeded if D is large. However, Approach speed is slow to a desired value. It is that approaches a desired value most quickly at the time of 1= 1, without going too far. When allowing slight going too far, Confrontation of D= 0.707 is the quickest. About the value of D, a desirable value is decided uniquely in this way. It is speed gain 1 (it is necessary to determine v5 feedback gain ICF) so that the value of this D may be given. In (7) of the definitional equation of D, KT and J are a motor, load, or a constant that will be decided if given, and it cannot set them up freely. However, the expression of relations between Izetav and Ki is only decided only by the value of D having been decided. In order to determine the value of Kv and Kp uniquely, one more types are required. Stiffness of a servo system is adopted by more evaluation M Semi and second. Staffiness S is load to the directed position -- The squid -- it is the quantity showing whether it fixes firmly 1 This. A following formula defines Staffiness S. s Yu. (9) Here, [S is the stability in a stationary state when only deltatheta displaces 0 from command value Oc and a long time is fully maintained. If this is large, load will be supported firmly in a predetermined position. (1) (4) types to Stiffness S is S =KFKT. It is given by 00. feedback gain KF is thick -- it can kick -- Staffiness S is so large that it is large. K) may be enlarged. However, when KF is enlarged, it will increase in the maximum of armature current under servo operation, and the maximum rating value of a motor will be exceeded. Therefore, feedback gain KF cannot be enlarged without any restriction. Thus, feedback gain KN, i.e., Staffiness S, is restricted by the maximum rating of motor current in the conventional servo system. in order to obtain large Staffiness S, a large-sized servo motor with the large maximum rating must be used -- it becomes. It cannot be said that comparatively big load is controlled by a small motor. The performance of a small motor is demonstrated to the utmost, and it cannot do in * To. When using a DC servo motor for the drive of a robot arm, small size and a weight saving are required very much of a degree. Simultaneously, in order to acquire high position accuracy, sufficiently large Staffiness is required. In the case of the conventional servo system, it is difficult to fulfill such opposite conditions. * One of the methods which increases Stiffness at the time of stillness is inserting in a motor drive circuit the limiting circuit which restricts the maximum of motor current, using the motor of the same maximum rating. In order to enlarge Staffiness, feedback gain KF is enlarged enough [ effectual ] and, moreover, armature current ia of a motor is restricted within the limits of ±(setting a certain steady value below the maximum rating to IM) IM. Drawing 10 is a graph which shows the input-and-output characteristic of a limiting circuit. Horizontal axis X is an input and is given by the right-hand side of (4) types. Vertical axis f1 (xi is output barrel current ia.) Limiting circuit function f t (xi showy definition is carried out.) As variable X, the quantity of the right-hand side of (4) types is used. Namely, X =lCF (the form of (b) (jc-0- (braking paragraph)) is taken.) The present invention has the new feature also in a braking paragraph besides using a limiting circuit. As a braking paragraph, it is linear expression % type % of speed. Quantity dOdo which was proportional to the square of speed differentiation and had numerals contrary to speed differentiation as a braking paragraph in the present invention -One stone 1'clt ' Q4 is used. Motor current i1 is dOdo in the present invention. 1 a-fl (it gives by Kv(theta C-theta-KV Completion 1 a, l) alpha0.) The definition of limiting circuit function Ex(x) is as <>upsilon. It is that of alpha (1) If it substitutes for (2) and substitution of (6) and (8) is performed, an equation of motion will be made dimensionless, d2 theta do do - = f 2 (set to 1dT2 dT dT.) (Oc-theta-2"-1-1) Normalization limiting circuit function E 2 (Xl is 0upsilon') A definition is come out and given. P r =K F KT 1(v 2J It is alphaeta. Next, a drawing explains the composition of the positioning servo system of the present invention. This is provided with the servo drive circuit which determines motor current a] for giving equation of motion alpha 0. Drawing 1 is the whole positioning servo system lineblock diagram concerning the example of the present invention. Loading is connected with servo motor M. Negative (angle detector A and speed detector C are connected with l'iJ I.) Angle detectors A are an encoder, a resolver, The potentio meter, etc. speed detector C -- an octopus -- it is a generator etc. Speed detector C The and rotation angle speed (d theta/dL) are obtained. Absolute-value circuit 13 makes absolute value theta/d [ of 1d ] [1. Multiplier 14 (dO/dt) ldO/dtl is calculated. To the upper value, Calculator 15 squares Kv. Subtraction machine 16 carries out comparison subtraction of command value Oc and the present angle theta. Subtraction machine 17 is dOdothetaC-theta-Kv-1from multiplication machine 15 and subtraction machine 16-1. (IQdt dt is calculated.) Amplifier 18 hangs feedback gain KF on the value of alpha 8. IJ Mitta 12 has the current restriction operation expressed to 00 formulas, and gives the current decided by theta:Formula to servo motor M. Drawing 2 is the whole positioning servo system lineblock diagram showing other examples. A different place from Drawing 1 is the point of there being no speed detector C, differentiating angle signal theta of angle detector A with differentiation machine 19, and having obtained speed (dO/dL). Drawing 3 is a block diagram of servo drive circuit B of the example corresponding to Drawing 2. it gives and appears in the element with which the same numerals as Drawing 7 are common. 1.2 is an adding machine. 3, 4.5-1, and -Kv. The multiplication machine which squares I<F. 6 is a differentiation machine. 10 11 shows mechanical coupling. 12 is a limiting circuit. 13 is an absolute-value circuit and 14 is a multiplier. The feature of the present invention is in the place which restricts current by a limiting circuit, using the square of speed as a braking paragraph. If it does in this way, a braking operation will already decrease in Beginning and going too far (overshooting), and Staffiness S will be obtained in the feature that it is large. By a computer, non-linear differential equation a0 is solved. Drawing 4 shows the result of numerical computation. The size of a step is set to the <H-2 [ same ] (Radians) as a precedent. Size"/K F of a limiter is set as the value of 0 and 05 which produced big going Past in a precedent. Size l' of resistance calculated about 4.5.10. Normalization time omegaot and the vertical axis of a horizontal axis are normalization rotation angle theta/II in Drawing 4. At the time of I"=4, braking runs short, and 0 misses a desired value, and it dies, and is liking To. It descends or passes through the neighborhood of re-Hiyuki past The, several times, or a desired value up and down. Convergence is slow. Although it dies too much and quantity decreases at the time of I'''-5, it converges slowly, still vibrating the neighborhood of a desired value. At the time of 1'=1.O, it dies too much more, quantity becomes small, and the amplitude of vibration is also small. It converges more promptly. Thus, convergence will become quick if braking coefficient F is enlarged. P can be arbitrarily set up also by operating only feedback gain Kv. Staffiness S is I (based on F.). The influence of KF influences the speed of convergence through time-basis omega0 [ useless ] which does not lead but normalizes 1'. If normalization equation 00 was used, this point cannot understand enough. A formula from the first is alpha 9. It comes out. Another normalization equation must be made. I will calculate a solution about the following differential equation in preparation before that. It is the following non-line type equation. F is a constant. In order to make it correspond to 00 formulas, as for positive one and SF, in (dy/dX), positive and (dy/dx) should just consider that F is negative, when negative. (dY/dX) is set with new variable P. 1-= Stand ■ It is upsilonx. The second degree differentiation It can change. (b) man ° Kick with a formula. Variable Q Q=P2 Foundation -- a definition is given. Han type It is changed into the line type equation to say. General solution However, A is a constant of integration. 1.21) Co., Ltd. -- from -- obtain the integral equation to say. This is a case where (dgamma/dx) is positive. (d) When '/dx is negative, it may replace by F-> (-F). The inside of Q root sign must always be positive. The value of y of the point that speed 9 is set to O is calculated noting that the inside of the root sign of ' (b) and (C) is 0. Drawing 11 is a graph which expands and shows a part of oscillating portion of Drawing 4. Although a horizontal axis is X and a vertical axis is y, X corresponds to time and y corresponds to angular displacement here. however, y -- zero itself -- it is not (0-Oc) -- it corresponds. (C) Correspondency can be known from a formula. It becomes a (b) type between Cd(s) of which a (b) type with 9>0 The consists between ab(s) and of which a (C) type consists in 9 To 0 between bc(s). Here, b, c, d, and ..... show the point used as :l-0, i.e., an extremum. For example, the denominator of a (C) type is considered between l c. At the time of of, it is y=o, and M- is set to 0 and is always negative at y+0. Drawing 12 M - When + Be"Y It is a graph which is divided into the members of O and shown. If B is smaller than the value of a ring, the graph of 0 upsilon and six proverbs has two intersections. y coordinates of an intersection hit Extreme and gamma coordinates of c. here -- Ml being set to 0 --" of a (b) type -- it is because it is set to ?/dx20. 03 F2 It sets. epsilon is a positive slight quantity. Under approximation called 2Fgamma Toughness 1 which substitutes for a Han type and calculates y coordinates of an intersection, it is y2. Two Fepsilony-epsilon=0 It deals in 2 called (Ro). This is solved and it is b=epsilonF+FStone earT. (From) C-epsilon F-1 second F2+=. (From) The ratios of b and c are (C) and 1 To lay 1 from (From). It is clear that it is smaller than @1. Next, the thing same per field of 9>0 is considered. Between cd(s), the inside of the root sign of a (b) type is made into M+, and the value of gamma which sets this to 0 is calculated. This is c and d. M = - Engineering+ -- A e-'gamma (From) If A is not less than 172F2 like +F2F2, there will be no field where M+ just becomes. Drawing 13 is the inside of M+. A e-'Y It is a graph which shows ■. 8 When it sets with 0upsilonF2, this which two intersections produce is c and d." -- delta - approximation which 4 upsilon is substituted for (From) and called 2Fy Toughness 1 -- y2+2Fdeltay-delta=002 -- solve this c --F Delta-F stone 7 Adventure theta■ d = -F delta+rho] A cow ■ is obtained. Inequality % Formula %() Established(ing) is clear. Thus, although the value of y vibrates to the upper and lower sides of y=Q, the ratio of the absolute value of an extremum is always smaller than 1, and it turns out after all that it converges to gamma-0. Drawing 14 is a diagram for the state of convergence to be shown [ of y->0 ]. Straight line M shows 00 formulas and straight line N shows 0upsilon type. A horizontal axis is y. index function (the left going up) Kt Into -- 2 * ...... a Mission type -- index function (the right going up) LX and L2 ...... is a curve which shows a (b) type. Suppose that movement between a and b corresponds to several kiloliter of boxes. The value of y of intersection PI of Kl and M is b. The point on straight line N of the same y coordinates as Pl is set to QL. Index function L1 which passes along Ql is made. It is considered as graph L1 and other intersections Q2 of straight line N. The value of y of Q2 is C. The point that the value of y is C on straight line M is set to F2. Index function graph 1 (2 is subtracted.) which passes along F2 Other intersections with straight line M are set to I'3. X coordinates of P3 are d. It turns out that Ri different appearance is used and it is completed as y= 0 by the value of the extremum of The and y. The speed of convergence is based on F. Convergence is so quick that it is large in F. The interval of X coordinates of the point used as y =O is constant in Drawing 11. This is proved. For example, it can approximate with a (C) type between bc(s). The denominator of the left side It can rewrite. the integration of this -- Or et al. -- it asks easily. The left side of 0@ is pi. Therefore, the difference of X coordinates between bc(s) is pi. the same -- cd and ...... all the differences of X coordinates are pi also between. This fact is proved [ Drawing / 4 ] although a horizontal axis hits X. The cycle is constant although theta vibrates up and down. Now, the normalization equation is considered after the Han type. (b) The next conversion may be performed in order to equip a formula and 0@ type. Gamma-theta-Oc When theta9F was large, attenuation of y in vibration of - time was large. F is proportional to Stiffness S (= KF KV). The measure of time is 17 beta from Q type. It turns out that it is proportional. After all, the direction which enlarged Staffiness S is a double A meaning, and it turns out that convergence to command value thetaC of 0 becomes quick. if Staffiness S is enlarged, the portion into which a formula like a (Ho) type is materialized becomes narrow, and, in almost all cases, the greatest limiting circuit current±IM is flowing into" the servo motor -- Eye -- This -- . Since it drives in a motor by the maximum output, it reaches to an extreme and comes out, and a step response is [ to / near the command value ] quick. An effect is described. Same then there being limiting circuit current IM decided by maximum rating current and current torque coefficient KT are the same in a motor. Stiffness decided by 00 formulas is I (it can do greatly by enlarging F.). It compares with the conventional method and the present invention is I (thick To 1. The also has little line 7 Too much, and can make F large without unreasonableness [ Stiffness / S ].). According to the present invention, even if it is a small direct-current servo motor, Stiffness at the time of stillness can be enlarged. since Stiffness is large -- accuracy -- it can position highly. A king It is made to move to the demand of the servo system stated to 1 beginning, and the position of a wish of (1) load at high speed, and is made to stop smoothly in the position of (2) wishes for a short time. (3) Fix after a stop firmly so that it may not deviate from the position. The inside of what is said and (2) Especially (3) can be satisfied. Present invention A fal robot's joint driving device, the (b) NC machine tool, the positioning servo system of the (c) DC servo motor, It has a wide scope in the positioning servo system by a fal electricity oil pressure servo, etc.

[Brief Description of the Drawings]

Drawing 1 is the whole positioning servo system lineblock diagram concerning the example of the present invention. Drawing 2 is the whole positioning servo system lineblock diagram concerning other examples of the present invention. Drawing 3 is a fluke figure of the servo drama If circuit of the example corresponding to Drawing 2. The graph of the servo system, of the present invention Drawing 4 indicates a step response to be. II = 2 and "/KF = 0.05 r' is an example about three kinds of values of 4*5.10. Drawing 5 is a perspective view showing movement which connected load with the servo motor. Drawing 6 is a position servo control system lineblock diagram. Drawing 7 is a block diagram of the servo drive circuit concerning a conventional example. The graph Drawing 8 indicates the step human power which gave position command value OC in the shape of a step to be. The graph which shows the step response of the servo system which requires Drawing 9 for a conventional example. I) = the case of Q, 5, 0.707, and 1 was shown. Drawing 10 is a crowd power characteristic graph of a current limiting circuit. A horizontal axis is input X and a vertical axis is output ft (it is xi.). The graph Drawing 11 indicates typically the temporal response of the step response of the servo system of the present invention to be. Horizontal axis X corresponds to time and vertical axis gamma corresponds to the difference (theta-Oc) of present value theta and command value thetaC. b c, d, and ...... show an extremum. a Take the place in l positive/negative. Drawing 12 is a graph which shows (M-) the value in the root sign of the denominator of the integration of gamma in the field where nine in the 11th figure is negative. It is the portion which gave the slash, and a field which just carries out the inside of a root sign, and gamma coordinates of both ends give an extremum. Drawing 13 is a graph nine in the 11th figure indicates the value of value M+ in the root sign of the denominator of the integration of y to be in a positive field. The portion which gave the slash is a field which carries out right [ of the inside of a root sign ], and y coordinates of both ends give an extremum. A diagram for Drawing 14 to show dynamic (y->0) change which the servo system of the present invention reduces the absolute value of an extremum according to Si's positive/negative, and converges to a command value. Straight line M (-- graph of L+-!->) Straight line N is F. 2F2 1 The graph of (-10-). K1, K2, and ...... are C-zFyF. The index function graph of 2F2, Ll, L2, and ...... are C2Fy. Index function graph. Ql, C2, and ...... LX and L2 -- the intersection of ..... and straight line N. P+ and P2, ..-... K1. and K2. -- the intersection of ...... and straight line N1. 1.2 ..... An adder 3.4.5 ...... Credit Calculator 6 and differentiation machine N1 ..... Servo motor L ..... Negative Load A .... Angle detector B ..... Servo drive circuit C ..... Speed detector 12 ..... A reminder 13 ...... Absolute-value circuit 14 ...... Multiplier 15 ...... Credit Calculator 16 17 ...... Decrease Calculator 18 ...... Amplifier Artificer Bamboo Honaku applicant for a patent =[ Sumitomo Electric Industries, Ltd. ] 78 Drawing 5 Drawing 6 (- Drawing 7