JPH0334082B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a feedback control device that performs proportional-integral control, and is particularly applicable to high-speed feed drive motors for blades such as machine tools, arm drive motors for industrial robots, etc. The present invention relates to a proportional-integral circuit suitable for responsiveness, overshoot prevention, small speed ripple, and high control rigidity.

[Background of the invention]

A conventional feedback control system is shown in FIG.
The detection circuit unit 8 detects the target value V _REF and the control result, ie, the control V _O , and outputs a feedback signal V _F. Next, the comparison section 1 calculates the difference V _E (=V _REF -V _F ), and the amplification circuit section 2 amplifies it. This output V _EO
is amplified by the proportional circuit section 3, and is amplified by the integral circuit section 4.
It is integrated by These outputs V _P and V _I are added in an adder circuit section 5 to output a signal V _A. Note that the proportional circuit 3, the integrating circuit 4, and the adding circuit 5 are referred to as a proportional-integral circuit section 9. If the controlled object 7 is an electric motor, the signal V _A becomes a torque command (or current command),
The current is amplified by the electric amplifier circuit 6, controls the controlled object 7, and outputs the controlled variable V _O. FIG. 3 shows the operation of the block diagram shown in FIG. 1, and assumes that the controlled object 7 is an electric motor. Therefore, the control amount V _O follows the target value V _REF . Time t ₀ to t ₁ will be explained with reference to FIG. 3-1 is the target value
When V _REF changes stepwise and the control amount V _O and feedback amount V _F accelerate and rise in a ramp-like manner, the output V _E of the comparator 1 decreases from a high positive value as shown in 3-2. Next, the output V _EO of the amplifier circuit section 2 is V _E
In order _to amplify _the
become that way. Next, V _EO is multiplied by an arbitrary coefficient in the proportional circuit 3 to become 3-4. In integration circuit 4
Integrate V _EO and rise like 3-5. Next, the adder circuit adds V _P and V _I of 3-4 and 3-5, but the saturation voltage of the adder circuit (+V AMAX on the positive side and -V _AMAX on the negative side)
V _AMAX ), so it does not rise and reaches +V _AMAX at time t _A as shown in 3-6. The broken line is
Indicates the arithmetic addition of V _P +V _I. Next, at time t= _t1 , when the target value V _VREF decreases stepwise from positive to negative, the output V _E of the comparison section 1 and the output V _EO of the amplifier circuit section 2 become △V _E and △V _EO, respectively. ΔV _P decreases, and the output V _P of the proportional circuit 3 also decreases. The output V _I of the integrating circuit 4 becomes discharged because the input V _EO becomes negative, and decreases as shown in 3-5. The outputs of the proportional circuit and the integral circuit are added in the adding circuit 5,
The proportional circuit power V _P and the integral circuit output V _I are shown in Figure 3-3-4.
3-5 is not saturated individually, so
The adder circuit output V _A , that is, the current command, falls below the point of the broken line V _P +V _I by △V _A = △V _P as shown in 3-6. Therefore, the adder circuit output signal V _A does not become a negative voltage at time t= _t1 , as shown in 3-6, but becomes a positive voltage. The positive voltage continues until time t _B , but at time t ₁ V _A
is positive, the torque to be controlled is on the positive side, causing a response delay, that is, overshoot. This situation is shown by the broken line 3-1. Furthermore, the third
Figure 3-7 shows the ideal adder circuit output V _A , and from the saturation voltage +V _AMAX to t = t ₁ , △V _A = △V _P
V _A becomes a negative torque command.
At this time, the controlled output V _O is t as shown in 3-8.
= t It is decreasing from ₁ . In this case, there is no overshoot.

As mentioned above, FIG. 1 has the disadvantage that overshoot occurs due to the relationship of the saturation voltage of the adder circuit. Next, FIG. 2 will be explained as another example. FIG. 2 shows an example in which the proportional-integral control circuit section 9 of FIG. 1 is configured with an analog circuit. Target value V _REF
and feedback value V _F are connected to the negative terminal of operational amplifier OP1 through resistor _R1 . The positive terminal of OP1 is connected to zero potential through a resistor _R2 .
Also, a resistor Rf and a capacitor are connected to the negative terminal and output terminal.
Cf are connected in series. In addition, operational amplifier OP
1 is operating linearly, the negative terminal of OP1 is at zero potential, so the current if flowing through the resistor Rf and capacitor Cf in Figure 2 is if = i ₁ - i ₂ = 1/R ₁ (V _REF −V _F ) …… becomes. Therefore, the output V _A is V _A = - (Rf if + ∫ ^t _p 1/Cfif dt) - [Rf/R ₁ (V _REF -
V _F )+1/R ₁ Cf∫ ^t _p (V _REF −V _F ) dt]... If V _REF −V _F =V _E , then V _A =−Rf/R ₁ ×V _E +1/R ₁ Cf∫ ^t _p V _E dt) ... (The first term in the above equation is called the proportional term, and the second (The term is called an integral term) From the equation, the proportional circuit output V _P and the integral circuit output V _I shown in Figure 1 are as follows: In Figure 2, the voltage across the resistor Rf is V _P and the voltage across the capacitor C _f is V _I. ,
Since Rf and Cf are connected in series, the output voltages are added.

Therefore, FIG. 2 shows a control system similar to that of FIG. 1. In this case, as shown in Figure 1, the saturation voltage of the proportional part and the integral part are not determined individually, but are determined by the proportional-integral circuit output V _A added as shown in the equation, so the proportional part When is large, the saturation value distribution will be different so that the integral part will be small. In the opposite case, the integral part becomes large. Next, the operation shown in FIG. 2 will be explained with reference to FIG. The overall control system is the same as in Figure 1, the controlled object 7 is an electric motor, and it is assumed that a constant load is applied to the electric motor.4-1 calculates the target value V _REF as
When the rotational speed is increased by a small amount from a certain constant rotational speed, the control amount V _O and the feedback amount V _F accelerate as shown by the solid line between time t ₃ and t ₄ . As shown by the solid line 4-2, the proportional section output V _P rises rapidly at t=t ₃ and becomes zero at t=t ₄ when the target value V _REF and the feedback amount V _F become equal. As shown in 4-3, the integral part is integrated at t= _t3 and not integrated at t= _t4 , resulting in a peak value. The output of the proportional integral part, that is, the current command V _A is the addition of V _P and V _I , so it looks like the solid line 4-4, but the target value is reached at t = t ₄ .
V _REF and the feedback amount V _F are equal, but if the load on the motor is constant, the current command V _A must be equal before and after t= _t3 . However, at t = t ₄ , the proportional part V _P becomes zero, but the integral part V _I integrates the acceleration torque from t ₃ to t ₄ , and becomes the value V _I1 at t = t ₄ . Therefore, it is overcharged by V _I1 − V _I. Therefore, V _I1
Until the amount of -V _I is discharged, a current larger than necessary flows as shown in 4-4, so the acceleration is further accelerated after t= _t4 , resulting in overshoot as shown by the solid line in 4-1. Next, in Figure 2, reduce the capacitor Cf,
When the integration time is shortened, the rise becomes even steeper and the amount of overshoot becomes larger, as shown by the broken line 4-3. For this reason, V _P and V _A also become like the broken lines 4-2 and 4-3, so the controlled variable V _O also approaches the target value faster, as shown by the broken line 4-1, but
It becomes vibrating. Next, in a similar circuit shown in Figure 2, the low speed rotation speed, for example, the speed range is 1:1000~
Figure 5 shows an oscilloscope when rotating at the lowest speed of 10000. As shown by the solid line in FIG. 5, at the lowest speed, speed ripples occur in a steady state due to speed fluctuations due to slot ripples caused by the stator winding of the motor, as well as non-uniformity in the torsional rigidity of the load. The broken line indicates a smaller value for the capacitor Cf, and as mentioned above, the overshoot at startup will be larger, but since the integration time is shorter, it has the ability to absorb speed ripples. Therefore, when Cf is small, the speed ripple in steady state is extremely good. As mentioned above, if there is an overshoot due to the saturation voltage problem in the configuration shown in Figure 1, and if the capacitor Cf for the integrating circuit in the circuit shown in Figure 2 is large, the response will be slow.
Due to the slow response, the speed ripple at low speeds becomes large. On the other hand, when Cf is small, there is a drawback that the overshoot becomes even larger.

The above problem requires precision of several microns.
Speed control and positioning control of the electric motor for feeding blades in NC machine tools, or the arm drive motor of assembly robots, which are becoming faster and more precise year by year, have a high frequency of forward and reverse rotation, and the speed response frequency is 30 to 100Hz. When high-speed response is required and stable speed control characteristics without speed ripple are required over a wide speed range of 1:1000 to 10000,
Solving these problems has become an essential condition.

[Purpose of the invention]

The purpose of the present invention is to provide high-speed speed response that prevents overshoot, with small speed ripple even at the lowest speed in the speed range of 1:1000 to 10000. The object of the present invention is to provide a proportional-integral control circuit in which speed fluctuations (control rigidity) are extremely small.

[Summary of the invention]

As shown in Figure 1, if overshoot occurs due to problems with the saturation values of the proportional circuit 3, the integration circuit 4, and the saturation value of the adder circuit, when the proportional-integral control output V _A attempts to exceed the maximum value V _AMAX , This can be solved by initializing the integrator circuit output signal V _I to be (V _AMAX - V _P ), and in the case of Fig. 2, the target value
When the difference V _E between V _RER and the feedback amount V _F is large, increase the integration time to prevent overshoot, and when V _E becomes small, that is, the target value V _REF and the feedback amount V _F approach each other. This system realizes a system with quick response by automatically switching the constants of the integrating circuit by reducing the integration time when the system changes.

[Embodiments of the invention]

An embodiment of the first aspect of the proportional-integral circuit section of the present invention will be described with reference to the block diagram of FIG. Figure 6 is the first
It corresponds to the proportional-integral circuit section 9 in the figure, and the same reference numerals as in FIG. 1 are completely the same circuits. 1st
The difference from the diagram is that the maximum positive limiter value of the adder circuit output V _A (current command) is set at D ₁ and the maximum negative limiter value - D ₂ is provided, and the proportional limiter value is set at the proportional circuit output shown in Figure 1. A circuit 11 is provided, and a three-position operation nonlinear circuit 10 is provided which compares the adder circuit output V _A with +D ₁ and -D ₂ , and when this output V _D is negative,
Reset the output V _I of the integrator circuit 4 to be D ₁ −V _P , and when V _D is positive, set the output V _I of the integrator circuit 4 to −D ₂ −
It is designed to be V _P. For example, if +D ₁ is set to +10V - D ₂ = -8V, the output V _D of the three-position nonlinear circuit 10 will be set to V I = +2 V when V _D is negative, V _P = +8 V, and V _I = +2 V when V _D is positive. If V _P = -6V, V _I = -2V is set. When V _D is zero, the configuration is such that it has no effect on the integrating circuit 4.

Next, the operation shown in FIG. 6 will be explained with reference to the time chart shown in FIG. In FIG. 8, as shown in FIG. 3, the target value V _REF is changed stepwise at t=to as shown in 8-1, and the controlled variable _VO increases at an accelerated rate. t=
From to to t ₃ , 8-2 V _E , 8-3 V _EO , 8-4 V E
V _P is exactly the same as 3-2, 3-3, and 3-4 in FIG. V _P indicated by 8-4 indicates a case where the positive limiter D ₁ of the proportional limiter circuit 11 has not been reached. As shown in 8-5, the integrating circuit output V _I is integrated in the same way as in Figure 3, since the adder circuit output V _A does not rise to the positive limiter value D ₁ from time t = to to t ₁ . To go. However, at t=t ₁ V _A = D ₁
Then, due to the operation of the three-position nonlinear circuit 10, the integrating circuit output V _A is set to V _I =D ₁ -V _P , so that it is held at a constant value. That is, the integrating circuit 4 is not overcharged. Next, at t=t ₂ , the adder circuit outputs
Since V _A satisfies V _A <D ₁ , the integrating circuit output V _I increases. Next, if the target value changes to the negative side at t= _t3 ,
The integrating circuit output V _I becomes discharged and decreases until t= _t4 .
When the adder circuit output V _A becomes V _A = -D ₂ at t = t ₄ , the three-position operation linear circuit 10 operates and the integrator circuit outputs.
V _I is set to the value -D ₂ -V _P and held until t=t ₅ . Note that the output of the three-position nonlinear circuit 10 is shown at 8-7.

In this way, the three-position nonlinear circuit 10 limits the limiter of the proportional circuit to -D ₂ to _{D 1} , which is the same as the addition circuit output V _A , when the adder output V _A attempts to exceed the range of -D ₂ to D ₁ . Therefore, the maximum amplitude value, that is, the limit value, of the output of the integrator circuit is automatically changed by the output V _P of the proportional circuit to prevent overcharging of the integrator circuit and overshoot. be.

FIG. 7 shows an embodiment of the second invention, in which the proportional limiter 11 of the output of the proportional circuit 3 is removed from the block diagram of FIG. In this case, the adder circuit 5
Since the output V _A of is suppressed by the first and second reference voltages D ₁ and -D ₂ , when the proportional circuit output V _P exceeds the range of -D ₂ to D ₁ , the integral circuit output Although V _I is set to a value of opposite polarity, exactly the same effect can be obtained regarding the above-mentioned integration circuit overcharge prevention.

Note that a specific embodiment of the first invention is shown in FIG. 8-1.

In Figure 8-1, R ₅ to R ₂₈ are resistances, P, I,
D ₁ and D ₂ are variable resistors for constant setting, D ₁ to _{D 6} are diodes, C is a capacitor, and OP ₂ to _{OP 9} are operational amplifiers. The comparison section 1 and the amplifier circuit section 2 are composed of an operational amplifier _OP2 and _R5 to _R8 , and the output thereof _{is VEO.}
becomes. The signal V _EO constitutes a proportional circuit section 3 including an operational amplifier OP ₃ , R ₉ to R ₁₁ , and a variable resistor P for adjusting a proportionality constant. The integral circuit section 4 consists of OP ₄ , R ₁₂ to _{R 14} , a capacitor C, and a variable resistor I for adjusting the integral constant, and the output of the proportional circuit section is OP ₅ to _{OP 6} , R ₁₅ to
Proportional limiter circuit 11 from R ₂₅ diodes D ₁ to _{D 4}
It consists of The operation of the proportional limiter circuit 11 is from the variable resistor D ₁ for adjusting the negative maximum value to the resistor R ₁₅ .
to the negative terminal of OP ₅ through the positive maximum value adjustment variable resistor D ₂ , which is connected to the negative terminal of OP ₆ through the resistor ₂₄ , and the circuit a of the resistor R ₂₀ and the peripheral circuit b of OP ₅ , The peripheral circuit c of OP ₆ is divided into three circuits, each with characteristics as shown in a to c in Figure 8-2, and the three circuits are synthesized (a +
b+c), the limiter operates at a location other than between D ₁ and D ₂ as shown by the broken line. Since D ₁ and D ₂ can be adjusted with variable resistors, the limiter value is variable. The adder circuit 5 is composed of _R26 to _R28 and OP7, and the three-position nonlinear circuit 10 is composed of OP8 to OP9 and _D5 to _D6 . The operation of the three-position nonlinear circuit 10 is as follows:
Compare the adder circuit output V _A with D ₁ and D ₂ , and if V _A <-D ₁ when the output of OP ₂ is negative, V _A becomes negative and OP ₈ operates as a comparator and becomes “H”. As a result, diode _D5 becomes conductive. Resistors R ₁₂ and R ₁₃ are R ₁₂
>>If R is set to ₁₃ , the output of the integrating circuit is instantaneously discharged until V _A =-D ₁ .

Furthermore, when V _A > D ₂ , the output of OP ₉ becomes "L", the diode D ₆ becomes conductive, and the output of the integrating circuit is similarly discharged until V _A = D ₂ . As a result, the adder circuit output V _A is limited to within -D ₁ to D ₂ .

Next, since the specific embodiment of the second invention excludes the proportional limiter circuit 11, it is obvious from FIG. 8-1, so a description thereof will be omitted.

Next, the contents of an embodiment of the third invention will be shown in the block diagram of FIG. 9, regarding the problems caused by the circuit configuration of FIG. 2 and the operation explanation of FIG. 4 shown in the conventional circuit. 9th
The figure differs from the proportional-integral control circuit shown in FIG. 1 only in that a coefficient adjustment circuit 12 and a coefficient setting circuit 13 are provided at the front stage of the integration circuit 4, and other aspects are completely the same. Figure 9 shows that in the circuit of Figure 2, if the capacitor Cf is increased as shown in Figure 4, the overshoot will be small as shown in Figure 4-4-1, but the response time will be longer; When speed ripple occurs and the capacitor Cf is made smaller, the response time becomes faster and speed ripple is improved, but overshoot increases.

The present invention is based on the difference between the target value V _REF and the feedback value.
If the absolute value of V _E or the amplifier circuit output V _EO is large, equivalently increase the value of capacitor Cf and reduce V _E
Alternatively, when V _EO becomes small, the value of capacitor Cf becomes small, so that the advantage of both is switched. That is, the input characteristics of the coefficient setting circuit 13 are as shown in FIGS. 10 to 13. In FIG. 10, when V _EO exceeds a certain value,
V _EO when target value V _REF and feedback amount V _F are close
(or V _E may also be used) is switched in stages to increase it when it is small. Although FIG. 10 shows one stage, there may be two or more stages. 1st
FIG. 1 shows the characteristics shown in FIG. 10 with hysteresis. FIG. 12 shows continuous switching. In FIG. 13, K=(1V _EO1 +b) ^-n +1, where b>0 and n≧1 (real number). In addition, the coefficient adjustment circuit 12 is configured to output V _EO ×k based on the input V _EO and the output k of the coefficient setting circuit 13. When V _E or V _EO is small, the integral time is When V _E or V _EO is large, the integration time is increased to slowly bring the controlled variable V _O closer to the target value V _REF , and when it approaches the target value, respond quickly to overshoot. This is to control the speed ripple to further reduce the speed ripple.

Next, using the proportional-integral control circuit shown in FIG.
In the control system shown in the figure, in order to study the automatic control system, first the block diagram of FIG. 1 is shown using the transfer function of FIG. 14. An explanation of each symbol is shown below.

K _P ...Proportional circuit gain constant K _I ...Integrator circuit integral number K ₁ ...Conversion constant from integral proportional control circuit output to motor output torque T _I1 ...Motor coil time constant K ₂ ...Conversion coefficient from torque to rotation speed K _N ...Constant S for converting motor rotational speed to feedback amount...Laplace operator T...Motor output torque T _L ...Load torque N...Rotational speed The proportional integral part of Fig. 14 is summarized as shown in Fig. 15. In order to achieve more stable control than the block diagram in FIG. 15, FIG. 16 shows a Bode diagram of the open loop transfer function. The transfer function excluding the proportional integral part is
When G ₁ (S) and the proportional integral part are G ₂ (S), the equation becomes as follows.

G ₁ (S)=K ₁ / (1 + ST _I1 ) × K ₂ / S × KN ...... G ₂ (S) = 1 + (K _P / K _I / (1/K _I ) S ... Transfer function G ₂ ( S) decreases at -20 dB/decade as the angular frequency increases, and -40 dB/decade at the angular frequency wH, which is the reciprocal of the motor coil time constant.
It becomes decade.

In order to stabilize this control system, the open-loop transfer function G(s) = G ₁ (s) × G ₂ (s) is generally changed to -40 dB/ decade
and the angular frequency wL=KI/KP is -20dB/
decade, −40 dB/decade at wH, gain GD of wL and wH, GI ₁ is GD = 12 to 15 dB, GI
= −12 to −15 dB is known to be sufficient. Therefore, the transfer function G ₂ (s) of the proportional integral part is changed to the first one so that the one-round transfer function G(s) becomes the broken line in FIG.
You can decide as shown in Figure 6.

Next, when the circuit of FIG. 2 is expressed using the KP and KI constants shown in FIG. 14, it becomes as follows. Output of Figure 2
Since VA is shown in the formula, VA is expressed as S function.
(s), then VA(s)=KP・VE(s)+KI・VE(s)..., and KP=Rf/ _R1 ...KI=1/ _R1Cf ...

FIG. 17 shows how the Bode diagram changes when capacitor Cf in FIG. 2 is changed. If Cf is made smaller, KI becomes larger, and the angular frequency wL (=KI/KP) becomes larger, as shown by the broken line. That is, in the present invention shown in FIG. 9, the target value VREF and the feedback amount VF
When is large, it is a solid-line one-round derivative, and when it is small, it is a broken-line function. The angular frequency wc at which the gain of the open-loop transfer function G(s) reaches zero shown in Fig. 17 indicates the response angular frequency of this control system, that is, the target value VREF and the controlled variable Vo. It will be faster. Also the first
As mentioned above in Figure 7, the gain GD of the angular frequency wL is stable at 12 to 15 dB, but if Cf is made smaller, GD will become smaller and become unstable, but the target value
When the difference between VREF and the feedback amount VF becomes large, the characteristic instantaneously becomes as shown by the solid line in FIG. 17 where Cf is large, so that it does not become unstable.

Next, the operation shown in FIG. 9 is shown in FIG. 18. The period from t= _t7 to _t8 in FIG. 18 is the same as the solid line from t= _t3 to _t4 shown in FIG. ₄ , but from around t=t8 to after _t9 , the coefficient adjustment circuit 12 coefficient setting circuit 13, the fourth
Since the characteristics are switched to the side shown by the broken line in which the capacitor Cf shown in the figure is made smaller, instantaneous convergence can be achieved and overshoot can be significantly improved. In addition, the 18th
The broken line in the figure shows the operation when the characteristics are not switched around t= _t8 .

Next, another example of the third invention, that is, another embodiment of FIG. 9 is shown in FIG. 19. In FIG. 19, a coefficient adjustment circuit 12 is provided at the output of the integrating circuit 4, and the operation of the proportional-integral control circuit is exactly the same as that shown in FIG.

Next, an embodiment of the fourth invention that solves the problem shown in FIG. 2 will be described. The control block diagram is shown in Fig. 9 and Fig. 1.
This is the same as Figure 9, but the characteristics of the coefficient setting circuit 13 in the figure are set to 0 when VEo or Eo becomes large, as shown in Figures 20 to 23, so that no integration is performed. . In this case, the target value
When the difference between VREF and feedback value VF is large, only the proportional circuit operates, and proportional-integral control is performed when target value VREF and feedback value VF approach. In this case, the integrating circuit will not be overcharged, so the overshoot will be extremely small.

Next, the fifth invention will be explained with reference to FIGS. 24 and 25. This invention combines the first invention and the third invention, so Figure 24 is a combination of Figures 6 and 9, and Figure 25 is a combination of Figures 6 and 19. . The operation in this case is the first invention, the third invention
As mentioned in the invention shown in the figure, with only the third invention, speed ripple and overshoot can be reduced, but the integration time of the integration circuit is reduced equivalently to VE or VEO, but the load fluctuation is faster than that time. If this occurs, the fundamental cause of overcharging the capacitor remains. By combining the third invention with the first invention, the addition circuit output can be increased by a change in the proportional circuit in response to a change faster than the integration time.
Since VA changes and, as a result, it is possible to instantaneously reduce the output of the integrating circuit, it is clear that the effects of the invention can be further improved. A specific embodiment in this case is shown in FIG. 26 with respect to the coefficient adjustment circuit 12 and coefficient setting circuit 13 added to FIG. 6. Other circuits are shown in FIGS. 8-1. The coefficient adjustment circuit 12 is composed of a multiplier IC ₁ and variable resistors O and S for offset adjustment, and its output signal B is expressed as B=A×K/10. Here, signal A is VEO in the case of FIG.
In the case of FIG. 5, it is the output signal of the integrating circuit 4. The coefficient setting circuit 13, an example of which is shown in FIG. 10, converts the input signal VEO into a full wave rectifier circuit, that is, an operational amplifier circuit.
Consists of OP10-OP11 resistors R30-36 and diodes D10-11, the absolute value of VEO |VEO|
get. This signal is configured by a comparator circuit, that is, an operational amplifier circuit OP12 and resistors R37 to 38,
When |VEO| is large, the output of OP12 is “H” |
When VEO｜ is small, it becomes “L”. The output of OP12 is resistor R39 transistor Q ₁ , diode D
When OP12 is "H", transistor Q ₁ is turned on, and the coefficient setting circuit output K is K = (R ₄₁ × PV) / (R ₄₀ + R ₄₁ ), and OP12 is inverted.
When is “L”, transistor _Q1 is off,
K=PV is given. Therefore, when |VEO| is large, K becomes small, and when |VEO| is small, K becomes large.

Next, the sixth invention is a combination of the first and fourth inventions, and the block diagrams are as shown in FIGS. 24 and 25.
Although it is the same as the figure, the characteristics of the coefficient setting circuit 13 are as shown in FIGS. 20 to 23.

The seventh invention is a combination of the second invention and the third invention, and is shown in FIGS. 27 and 28.

The eighth invention is a combination of the second invention and the fourth invention, the block diagrams are shown in FIGS. 27 and 28, and the characteristics of the coefficient setting circuit 13 are
3.

Although the first to eighth aspects of the present invention have been described as specific examples using analog circuits, they may be used as proportional-integral control circuits having a software configuration using a microcomputer.

〔Effect of the invention〕

According to the present invention, the frequency of forward and reverse rotation is high, such as AC or DC motors for feeding cutlery such as lathes or driving the arms of assembly robots, and the speed range is such that a high speed response frequency of 30Hz to 100Hz is required. Stable speed control characteristics with no speed ripple over a wide range from 1:1000 to 1:10000 Furthermore, speed fluctuations are kept as small as possible even at the moment when the cutter hits the material, such as in the drive motor for feeding the cutter on the lathe. The present invention is indispensable for speed control devices that require high-speed response and high precision without overshoot, such as interpolation control that requires high control rigidity and high accuracy of robot arms. From an industrial perspective, it is of great benefit. Furthermore, if the present invention is controlled using a microcomputer, the number of parts can be further reduced, which is economically advantageous.

[Brief explanation of drawings]

FIG. 1 is a conventional feedback control block diagram, FIG. 2 is a conventional proportional-integral control circuit, and FIG. 3 is a time chart for explaining the operation of FIG. 1.
FIG. 4 is a time chart explaining the operation of FIG. 2. Figure 5 is an oscilloscope of the rotation speed when the circuit shown in Figure 2 is operated at the lowest rotation speed. FIG. 6 is a block diagram for explaining an embodiment of the first invention. FIG. 7 is a block diagram for explaining the second embodiment of the invention, and FIG. 8 is a time chart for explaining the operation of FIG. 6. FIG. 8-1 is a specific example of FIG. 6. Figure 8-2 shows the characteristics of the proportional limiter circuit of Figure 8-1. FIG. 9 is a block diagram of the third invention, and FIGS. 10 to 13 show the characteristics of the coefficient setting circuit of FIG. FIG. 14 is a diagram showing the block diagram of FIG. 1 using a transfer function, and FIG. 15 is a block diagram in which FIG. 14 is organized. FIG. 16 is a Bode diagram showing the open-loop transfer function of FIG. 15. FIG. 17 is a Bode diagram showing the open-loop transfer function when the circuit constants in FIG. 2 are changed. FIG. 18 is a time chart for explaining the operation of FIG. 9. FIG. 19 shows another embodiment of the third invention. FIGS. 20 to 23 show the characteristics of the coefficient setting circuit of the fourth invention. FIG. 24 is a fifth invention which is a combination of the first invention and the third invention, specifically, a combination of FIGS. 6 and 9;
FIG. 5 is another example of the fifth invention, which is a combination of FIG. 6 and FIG. 19. Figure 26 is Figure 9, 19
A specific example of the coefficient adjustment circuit and coefficient setting circuit shown in FIGS. 24 and 25 will be shown. FIGS. 27 to 28 are block diagrams of the seventh invention, which is a combination of the second invention and the third invention. 1... Comparison circuit, 2... Amplifying circuit, 3... Proportional circuit, 4... Integrating circuit, 5... Adding circuit, 6...
Current amplification circuit, 7... Controlled object, 8... Detection circuit, 9... Proportional-integral control section, 10... Three-position operation nonlinear circuit, 11... Proportional limiter circuit, 12...
...Coefficient adjustment circuit, 13...Coefficient setting circuit.

Claims (1)

[Claims] 1. A detection circuit section that detects a control amount and outputs a feedback value, a comparison section that takes the difference between both the target value and the feedback value, an error amplification circuit section that amplifies the output of the comparison section, and the error amplification circuit. a proportional-integral control circuit section comprising a proportional circuit for amplifying the output of the section, an integrating circuit for integrating the output of the error amplifying circuit section, and an adding circuit for adding the outputs of the proportional circuit and the integrating circuit, and the proportional-integral control circuit section; A proportional-integral control circuit used in a feedback control system comprising a controlled object controlled by an output of a circuit section, and wherein the detection circuit section is configured to detect a controlled amount of the controlled object. First,
a three-position nonlinear circuit for comparing the output of the proportional circuit with a second reference voltage; and a limiter circuit that limits the output of the proportional circuit to within the range of the first and second reference voltages; When the voltage is equal to or higher than the first reference voltage, the output of the integration circuit is set to a value obtained by subtracting the output of the proportional circuit from the first reference voltage based on the output of the three-position nonlinear circuit, and the proportional-integral control circuit The output of the integrating circuit is limited to be within the range of the first reference voltage, and when the output of the adding circuit is below the second reference voltage, the output of the integrating circuit is limited based on the output of the three-position operating nonlinear circuit. The output of the proportional-integral control circuit is set to a value obtained by subtracting the output of the proportional circuit from the second reference voltage, and the output of the proportional-integral control circuit is limited to be within the range of the second reference voltage. A coefficient adjusting circuit inserted in series with the integrating circuit on either the input side or the output side, and a coefficient setting circuit configured to set the coefficient of the coefficient adjusting circuit based on the output of the error amplifying circuit, The coefficient setting circuit is configured to continuously switch the coefficient of the coefficient adjustment circuit so as to decrease the coefficient when the output of the error amplification circuit is large, and to increase it when the output of the error amplification circuit is small. A proportional-integral control circuit characterized by: