JPH04279907A - Digitally controlled position servo device - Google Patents

Abstract

PURPOSE:To offer a device which shortens sampling synchronism and to obtain a device in which a stabilizing principle is applied to the stabilization of a servo system with dead time by introducing multiple sample lag to a control operation by utilizing the stabilizing principle. CONSTITUTION:The servo system is comprised of a current control loop provided with controller characteristic, a velocity control loop provided with controller characteristic, and a position control loop provided with controller characteristic. The response characteristic of the velocity control loop is set sufficiently faster than that of a low rigidity load, and a digital servo system by sampling synchronism Ts is comprised of the position control loop, where proportional control is performed. In such a state, short sampling synchronism Ts is selected so that control accuracy on a command applied from the outside or disturbance, etc., can be prevented from lowering by aliasing, and also, linear phase lag characteristic is inserted so as to increase the gain allowance of the characteristic of a servo device.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an improvement in a digitally controlled position servo device, and more particularly to an improvement in a digitally controlled position servo device in which the characteristics of a controlled object include a secondary vibration system.

0002

2. Description of the Related Art In recent years, with the spread of high-speed arithmetic processors, servo systems of automated equipment have been rapidly digitized. The sampling period in a digital servo system is a basic parameter newly added when converting a conventional analog servo system to digital, and the characteristics of the servo system vary greatly depending on its setting. In most digital servo systems, the characteristics of the system deteriorate as the sampling period becomes longer. On the other hand, as the sampling period becomes shorter, the cost of the signal detector, signal processing device, etc. that constitute the servo system generally increases. Therefore, the selection of sampling period is often a compromise between servo system performance and cost.

0003

However, it has recently been reported that if the sampling period is set too short, the characteristics of the constructed digital servo system may deteriorate depending on the object to be controlled. In other words, depending on the characteristics of the controlled object, there may be a sampling period that maximizes the performance index of the servo system.

The present invention has been made in view of the above-mentioned problems of the prior art, and its purpose is to clarify the principle of stabilization of a servo system brought about by selection of the sampling period using a method that focuses on frequency characteristics, and to improve this principle. The purpose of this invention is to present a device that shortens the sampling period by introducing a multi-sample delay into the control operation by utilizing this principle, and also to provide a device that applies this principle to the stabilization of a servo system with dead time.

[0005]

[Means for Solving the Problems] In order to achieve the above object, the digital control position servo device according to the present invention is a digital control position servo device including a secondary vibration system having a damping coefficient ζ smaller than 0.3 in a controlled object. In the device, the sampling period Ts is selected to be short so as not to reduce the control accuracy due to aliasing in response to externally applied commands, disturbances, etc., and to increase the gain margin of the characteristics of this servo device. The feature is that a linear phase delay characteristic is inserted into the characteristics.

Further, the digital position control servo device according to the second aspect is characterized in that the linear phase delay characteristic is a dead time characteristic with a delay of n samples (n>1, a natural number).

Further, in the digital position control servo device according to claim 3, the linear phase delay characteristic is a Butterworth filter characteristic, a Chebyshev filter characteristic, etc. having a predetermined order m, and is a filter characteristic of a low-pass characteristic. It is characterized by

Furthermore, the digital position control servo device according to claim 4 is characterized in that the cutoff angular frequency of the low-pass characteristic is selected to be lower than the natural vibration angular frequency of the controlled object according to claim 1.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the drawings. Generally, a servo system in an automated device is composed of a controlled object including a servo motor and a load, and a controller including a detector and a compensation element.

That is, as shown in FIG. 1, a servo system 10 includes a servo motor 12 and a load 14, which are connected by a low-rigidity feed mechanism 16 such as a ball screw or a belt.

A mathematical model for this case is shown in FIG. In Fig. 2, ωn is the natural angular frequency of the low-rigidity mechanism model, ζ is the damping coefficient of the model, kf is the conversion coefficient from rotational displacement to linear displacement of the model, m is the mass of the load, and J is the This is the moment of inertia of the rotating system of the same model.

On the other hand, as shown in FIG. 3, the servo system has a multiple loop structure consisting of a current control loop with controller characteristics Gi, a speed control loop with controller characteristics Gv, and a position control loop with controller characteristics Gp. are increasing. In Figure 3, kT is the motor torque constant (conversion constant between the current applied to the motor and the torque generated by the motor), kE is the motor back electromotive force constant, R is the resistance value of the motor circuit, and L is the inductance value of the motor circuit.

Furthermore, if the current controller, speed controller, etc. are appropriately designed, the position servo system can be simplified as shown in FIG. In FIG. 4, k is the control gain and Tv is the time constant of the velocity loop.

Here, if the response characteristic of the speed control loop is sufficiently faster than the response characteristic of a low-rigidity load, and if a digital servo system with a sampling period Ts is configured in the position control loop and proportional control is performed, the system as shown in FIG. Become. In FIG. 5, ζ and ωn are the damping coefficient and natural angular frequency (rad/sec) of a low-rigidity load. Generally, in a low-rigidity positioning mechanism, ζ is considerably smaller than 1, and ωn is often on the order of several Hz to several tens of Hz.

The value of the natural angular frequency ωn of the secondary vibration system to be controlled changes greatly depending on the physical state of the controlled object, and as a result, the characteristics of the constructed servo system also change greatly. Therefore, in order to provide generality to the characteristic analysis of the servo system shown in FIG. 5, each parameter of the servo system is normalized. Here, all parameters in the servo system of FIG. 5 are normalized (dimensionless) using the natural angular frequency ωn of the controlled object. Table 1 shows the normalized parameters of the servo system.

[0016]

[Table 1]

Using these normalized parameters, the servo system shown in FIG. 5 can be normalized as shown in FIG. The one-sample delay element in FIG. 6 is a model of the time delay of control calculation in the control device,
Also, a zero-order holder is used here. These are commonly found in common digital control devices.

Explanation of Stabilization Principle Based on Open-Loop Frequency Characteristics Open-loop frequency characteristics are convenient for locally grasping the stability of a servo system. In other words, the critical stable state of the servo system can be investigated using the relationship between the open-loop gain characteristics and phase characteristics.

The open-loop pulse transfer function Go(z) of the servo system shown in FIG. 6 is obtained as follows using z transformation.

[0020]

[Math 1]

By substituting z=exp(jτsΩ) into equation (1), the open-loop frequency characteristic of the servo system shown in FIG. 6 can be calculated. Taking the case of ζ=0.01 as an example, when the normalized sampling period τS is increased from zero, the corresponding open-loop frequency characteristic is obtained as shown in FIG. As is clear from FIG. 7, a local minimum value and a local maximum value (resonance value) appear in the gain characteristic curve at frequencies Ωm and ΩM, respectively. Further, as the sampling period τS increases, there is almost no change in the gain characteristics, and only the phase delay increases.

As is well known, the phase crossing frequency (when the phase is −
The servo system is stable if the gain is 1 or less at the frequency of π), and the servo system is unstable if the gain is 1 or less. Looking at the case where the sampling period τS=0 (analog servo), it can be seen that the phase crossing frequency almost matches the resonance frequency. This is the cause of limiting the control gain of the servo system. However, as the sampling period becomes longer, the phase crossing frequency moves away from the resonance frequency. The separation of the phase crossing frequency from the resonant frequency means that
This means that the control gain of the servo system is no longer limited to the resonance value. In particular, when the sampling period τS≈1.7, the phase crossing frequency and the minimum gain frequency overlap, and the control gain of the servo system can be set to be the highest. In other words, when the control gains are set the same, if the sampling period τS≈1.7, the maximum gain margin will be obtained in the servo system.

Furthermore, this stabilization principle can also be expressed mathematically. By the definition of z-transform, the open-loop frequency response of the servo system in FIG. 6 can be expressed as follows.

[0024]

[Math 2]

[0025] However, D(s'), H(s'), G(s
') are characteristics of the controlled object including a one-sample delay element, a zero-order holder, and a second-order vibration element, respectively. In other words,

0
026]

[Math 3]

When Ω<<2π/τs, the above equation (3) can be approximated as follows.

[0028]

[Math 4]

The first half of the right side of equation (8) is the gain characteristic of the zero-order holder, and the second half is the gain characteristic of the controlled object. Further, each term on the right side of equation (9) is the phase of the one-sample delay element, the phase of the zero-order holder, the phase of the integral element in the controlled object, and the phase of the secondary vibration element.

Differentiating equation (8) with respect to Ω gives d|Go|/dΩ
= 0, it is possible to find the extreme frequency of the gain characteristic. Further, in the case of Ω<<2π/τS, the gain characteristic of the zero-order holder is approximately constant, so the denominator of the second term in equation (8) may be differentiated with respect to Ω.

[0031] Therefore, the maximum gain frequency (resonant frequency)
ΩM and the minimum gain characteristic Ωm are obtained as shown in equations (11) and (12).

[0032]

[Math 5]

From this result, it can be seen that the gain characteristics do not depend on the sampling period. Also, (11) to (12)
), that is, the condition (Ω
Equation (13) is obtained where m and ΩM are real numbers.

[0034]

[Math 6]

If the equation (13) is satisfied, the gain margin of the servo system will be improved by appropriately setting the sampling period as described above.

On the other hand, in order to maximize the gain that can be set in the servo system, it is sufficient to match the phase crossing frequency and the minimum gain frequency. Using equations (9) and (12), this condition can be expressed as equation (14).

[0037]

[Math 7]

When equation (14) is solved for τS, the sampling period τsopt at which the highest control gain can be set for the servo system shown in FIG. 6 can be estimated as shown in equation (15).

[0039]

[Math. 8]

Furthermore, by substituting Ωm in equation (12) and τsopt in equation (15) into equation (8) and setting |Go|=1, the stability limit gain κm of the servo system corresponding to τsopt
ax is obtained as in equation (16).

[0041]

[Math. 9]

Ωm in equation (12) and τsop in equation (15)
Considering the numerical range of t, it can be confirmed that the approximation condition of equation (7) is almost satisfied. Table 2 shows the results obtained by the above estimation method and an example of the numerical solution of the true value.

[0043]

[Table 2] ──────────────────────────
────────────
τsopt
κmax
ζ ────────────────────
──────────── True value according to the present invention True value according to the present invention
approximation
Approximate value────────────
──────────────────────────
0 1.82 1.81
0.405 0.403───
──────────────────────────
────────0.02 1.78 1
．． 77 0.404 0.4
03────────────────────────
──────────────0.04 1.74
1.73 0.404
0.402──────────────────
──────────────────0.06
1.70 1.68 0.
404 0.403──────────────
──────────────────────────0.
08 1.65 1.63
0.405 0.404────────
──────────────────────────
───0.10 1.59 1.57
0.406 0.405───
──────────────────────────
────────0.20 1.25 1
．． 22 0.423 0.4
22────────────────────────
──────────────

As can be seen from these numerical examples, the method of estimating the optimum sampling period and the corresponding stability limit gain of the servo system proposed here has a small estimation error and can be used satisfactorily.

The following conclusions can be drawn from the analysis based on the open loop frequency characteristics shown above. In other words, the principle behind the fact that a servo system including secondary vibration elements can be stabilized by setting an appropriately long sampling period is that resonance in the servo system can be avoided due to phase delays caused by time delays in sampling and control operations. bring about.

Application to servo systems with one to many sample delays It is known that as the sampling period becomes longer, the servo system becomes more susceptible to noise and disturbances. The sampling period selection method presented in the present invention must also address this issue. Here, based on the conclusions obtained above, we will present a method for shortening the optimal sampling period by adding a multi-sample delay to the control operation of the servo system.

Here, a servo system in which the control operation is delayed by n samples as shown in FIG. 8 will be considered. However, the sample delay number n is a natural number. From the above analysis results, Figure 8
It can be seen that the sampling period that can set the highest control gain in the servo system satisfies the following relational expression (17).

[0048]

[Math. 10]

The terms on the left side of equation (17) are the phase of the n-sample delay element, the phase of the zero-order holder, the phase of the integral element in the controlled object, and the phase of the secondary vibration element. Further, Ωm is the open loop minimum gain frequency of the servo system in FIG. 8, and can be calculated using equation (12).

When equation (17) is solved for τS, the sampling period τsopt that allows the highest control gain to be set for the servo system in FIG. 8 can be estimated as shown in equation (18) below.

[0051]

[Math. 11]

Further, the stabilization limit gain κmax of the servo system corresponding to τsopt in equation (18) can be estimated using equation (16).

It is understood from equation (18) that as the number of sample delays n increases, the optimal sampling period τsopt decreases. In other words, by increasing the number of sample delays in the control operation, the optimal sampling period can be shortened, and the influence of noise and disturbance on the servo system can be reduced.

The optimal sampling period in the case of ζ=0.01, n=1, 2 is calculated as τsopt=1 using equation (18).
．． It is estimated as 79.1.08. Furthermore, when the control gain is set to κ=0.2 and these conditions are substituted into the servo system of FIG. 8 to calculate its closed loop frequency characteristics, results as shown in FIGS. 9 and 10 are obtained. As is well known, the frequency at which the arias phenomenon occurs is a frequency band centered at Ωi=2iπ/τS (i: integer). From this result, it can be seen that by shortening the sampling period, the frequency band in which the aliasing phenomenon does not occur becomes wider. In the case of Figure 9, Ω1=
3.50, and in the case of FIG. 10, it can be confirmed that Ω1=5.84.

Application to a servo system with dead time A dead time may exist in a controlled object or a control device. The dead time is viewed as a problem as one of the factors that causes a phase delay and impairs the stability of the servo system. Here, we will present a method for stabilizing a servo system with dead time by selecting the sampling period according to the stabilization principle of the servo system clarified in the above explanation.

As shown in FIG. 11, the controlled object has a dead time of length Td (normalized dead time τd=Tdωn).
Consider a servo system in which there is a delay of n samples in the control operation. According to the stabilization principle described above, it can be seen that the sampling period at which the highest control gain can be set for the servo system in FIG. 11 satisfies the relational expression (19).

[0057]

[Math. 12]

When equation (19) is solved for τS, the optimal sampling period τsopt is obtained as shown in equation (20).

[0059]

[Math. 13]

Since the sampling period must take a positive value, the upper limit of the dead time for which an optimal sampling period exists for the servo system in FIG. 11 is (18).
It can be found from Eq. In other words, it becomes as shown in equation (21).

[0061]

[Math. 14]

Summarizing the above results, the method for selecting the optimum sampling period for a servo system including dead time and secondary vibration elements can be obtained as follows. (1) If either ζ or τd does not satisfy equations (13) and (21), set the sampling period as short as possible. (2) ζ and τd
If both satisfy equations (13) and (21), the sampling period is set according to equation (20). Note that the number of sample delays is an integer of n≧0. Furthermore, the sampling period can be shortened by increasing n.

By utilizing the phase delay (linear phase delay) caused by the utilization time delay of a filter that has a linear phase delay characteristic and a transmission gain that has a low-pass characteristic, resonance of the servo system can be avoided and stabilization can be achieved. The principles of the invention that can be realized can be applied to further developments of embodiments of the invention.

If we pay attention to the one-sample delay and the transfer gain (ratio of output amplitude to input amplitude) shown in FIG. 6, we can see that ΩτS/2
When is small, this transfer characteristic is approximately unity. Therefore,
The resonance characteristic of the second term in equation (8) (this is Ω=
It does not affect the transmission gain g of a low-rigidity load, which exhibits a peak of the gain characteristic near ΩM), nor can it reduce the peak of g (near Ω=ΩM). The existence of this peak tends to cause resonance of Ω = ΩM in the behavior of the output Xo, and reducing the value of this peak (g near Ω = ΩM) is the position control characteristic G shown in Fig. 3.
This is one of the challenges of p.

The same situation applies to the servo system including the multi-sample delay characteristic shown in FIG. 8 and the dead time characteristic shown in FIG.

On the other hand, there are known filters that have a linear phase lag characteristic and a low-pass transmission gain, and examples of these include names such as a Butterworth filter, a Chebyshev filter, and an elliptic filter. (Takashi Taniogi: “Theory of Digital Signal Processing, 2
, "Image of Filter Communication", Corona Publishing), Chapter 1.

The characteristics of the transmission gain g' and phase delay φ' are as shown in FIG. FIG. 12 is an example of a Butterworth filter. Other Chebyshev and elliptic filters have similar characteristics. Such a filter is characterized by a cutoff angular frequency Ωc' and m. Even at the resonant angular frequency ΩM', the transfer gain is determined by Ωc' and m. In Fig. 12, if the order m of the filter (corresponding to the order of the polynomial representing the transfer characteristic) is selected low, the characteristics will be g1', φ1', and in this case, ΩM' (corresponding to ΩM in Fig. 7).
The reduction in transfer gain in d1db is small. However, the phase crossing frequency Ω1' can be increased. Also, if the order m is selected high, the characteristics will be g2', φ2', and the reduction in transfer gain in this case is d2d
It's big like b. However, the phase crossing frequency Ω2' must be small.

How to select m may be determined based on the value of the damping coefficient ζ of the low rigidity load characteristic. Generally, when ζ is small, m may be set high, and when ζ is large, m may be set low.

A filter having a linear phase delay characteristic with such a low-pass characteristic as a transfer gain characteristic is shown in FIG.
Due to the effect of including it in , the above-mentioned resonance can be further avoided. That is, the amplitude of the oscillatory behavior of Ω=ΩM of the behavior of Xo can be significantly reduced, and the convergence of this oscillation can be significantly accelerated.

An example of such a filter is described in the above-mentioned document 1 (
Chapters 1 to 4). In addition to this embodiment, implementation means as shown in FIG. 13 are also effective. That is, the implementation form of the filter regarding discrete data is called a digital filter, but when both the input and output are discrete data and finite Ω1' and Ω2' are given as in the case of the present invention, the corresponding There is an acceptable data processing time.

Therefore, in the embodiment shown in FIG.
A known microprocessor (e.g., MPU) 24, filter transmission having a processing speed capable of providing data provided via latch memory 20 to data bus 22 and generating output data within an acceptable data processing time. A program memory (eg, ROM) 26 that stores a known algorithm for realizing the characteristics, a known data memory (eg, RAM) 28 that stores data necessary to execute this algorithm, and a latch memory 30 that outputs processing data. configured.

Note that in FIG. 13, the latch memory 20
, 30 are latched, which is determined as part of the MPU operation for each of the above-mentioned allowable times.

Specific Example of Device A specific example of the digital control position servo device of the present invention described above is shown in FIGS. 14 and 15. Note that portions corresponding to those in FIG. 1 are indicated by the reference numeral 100, and description thereof will be omitted.

The servo system shown in FIG. 14 drives and controls a load (controlled object) 114 by a motor 112, and each has the low rigidity characteristics of the model shown in FIG. Here, θ, XO are state quantities corresponding to θ, x in Fig. 2, d
XO/dt is the temporal rate of change of position XO (i.e. speed)
, and are detected by detectors Dp and DV, respectively. An example of Dp is a potentiometer, and an example of DV is a tachogenerator.

In this embodiment, the motor 11
The second control is performed by a current control section 150, a speed control section 152, and a position control section 154. The current control section 150 includes:
Power amplifier 154 supplies current Ia to motor 112
, a latch memory 156 that outputs current command data Ic''.
, current compensator 158, whose current control characteristics are Gi
It is indicated by. Note that as an example of the power amplifier 152,
A known pulse width modulation type power amplifier is mentioned, which generates power that is controlled according to command data Ic that defines the on/off ratio of the pulse width, and supplies a controlled current Ia to the motor using this power. Further, the current Ia is detected by a current detector Di such as a current (instantaneous) detector using a known Hall element.

Here, the detected state quantity Ia, dXO/
dt and XO are ADs each having a latch memory on the output side.
The state quantities Ia, dXO/
dt, data corresponding to XO Ia', dXO'/dt,
XO' is stored at regular intervals and updated at each timing.

In the above configuration, the constant interval is as shown in FIG.
is the sampling period Ts shown in , and the AD converter 16
The characteristics of 0, 162, 164 are indicated by the sampling symbol + zero-order holder.

On the other hand, the command Xr is converted into discrete data Xr' via the latch memory 166. The latch timing of the memory 166 is determined by the AD converters 160, 162, 164.
This characteristic is expressed by the sampling symbol + zero-order holder.

As described above, the sampling period Ts
Discrete data Xr', Ic'', Ia', dX updated with
O'/dt and XO are processed in the digital data input to output section shown in FIG. The sampling period ts allows the processing time spent in this processing section.

Note that in FIG. 15, parts corresponding to those in FIG. 14 are designated by the same reference numerals, and their explanation will be omitted. The microprocessor MPU' in FIG. 15 connects latch memories 156, 166, AD converters 160, 162,
164 input/output data is updated at time intervals of each sampling period Ts. That is, discrete data Xr', Ia', dXO'/dt, XO are set in the latch memory 166 and the AD converters with latch memory 160, 162, 164, and the data are shifted at the timing required for each processing time. Discrete data Xr', Ia', dXO'/dt, X
0 to the MPU', and set them in the data memory RAM' at different times. Thereafter, these data are processed within the sampling period Ts, and the output data of the current control loop Ic'', the output of the speed control loop (=command data of the current control loop) Ic', and
Output data of the position control loop (=command data of the speed control loop) dXc'/dt is generated. Output data Ic
" is the output data of the processing section in FIG. 15, which defines the operation of the pulse width modulation type power amplifier 152 in FIG. 14.

In this way, the input data of each control loop (current control loop Ic', Ia': speed control loop dXc')
/dt, dXO'/dt: Position control loop Xr', XO
') to process the data and output data (
The data processing algorithm for generating Ic'', Ic', dXc'/dt) is stored in the program memory RO in FIG.
It is stored in M'.

The content of the known algorithm is as follows, for example. Here, the characteristics Gi, GV, and Gp correspond to the characteristics shown in FIG. That is, the characteristic Gi of the current controller exhibited by the algorithm of the current control loop is the current compensator 158, latch memory 156, power amplifier 15.
2 (which property itself is a constant). The current compensator 158 includes the hardware shown in FIG.
Current control algorithm stored in part of ' and RAM
It consists of the data necessary for this algorithm that is stored and updated in part of '. For example, the characteristics of Gi are Ki
It is composed of known characteristics such as (Ai+Bi×integral). Here, Ki is a gain constant, and Ai and Bi are each constants.

Further, the characteristic GV of the speed controller exhibited by the algorithm of the speed control loop is determined by the speed control section 152.
This is realized by The control unit 152 includes the hardware shown in FIG. 15, a speed control algorithm stored in a part of the ROM', and data necessary for this algorithm stored and updated in a part of the RAM'. The GV characteristic is composed of a known characteristic such as KV (AV+BV×integral). Here, KV is a gain constant, and AV and BV are constants.

As described above, when the above current control loop and speed control loop operate, the result is generally as shown in FIG. 5 (however, Gp is substituted for k).

Furthermore, the characteristic Gp of the position control section 154 exhibited by the algorithm of the position control loop includes the control gain k shown in FIG. This is a low-pass filter having a multi-sample delay characteristic corresponding to the time characteristic or a linear phase delay characteristic shown in FIG. Control unit 15
4 is the hardware shown in FIG. 15 and the hardware shown in FIGS. 6 to 8 or 6 to 11 or 6 to 13 stored in a part of ROM'
It consists of an algorithm corresponding to , and data necessary for this algorithm that is stored and updated in a part of RAM'.

As described above, the present invention can realize the original stabilization principle of the present invention by the required algorithm by utilizing simple known means as shown in FIG. 15, and can control the behavior of low stiffness characteristics. This can be done with a much higher gain margin than conventional methods.

That is, FIG. 16 is a diagram showing the behavior of the response of the output X0 due to the vibrational characteristics of the controlled object in FIG. 5, and is for the case of ζ=0.01. It is shown how the response behavior is improved by the system with one sample delay in FIG. That is, according to equation (18), τs=1.79 is obtained with n=1, and the response behavior of the output X0 obtained with κ=0.04 is shown in FIG. It is understood from the figure that the vibrational behavior in FIG. 16 has been significantly improved. However, the existence of vibrations based on small-scale resonance is still recognized.

On the other hand, a Butterworth filter (order m=3) is used as an example of a low-pass filter having a linear phase delay characteristic as shown in FIG. 12, and its cutoff frequency is set to Ω1'=0.3. When the sampling period τs=0.5, the behavior of the response of the output X0 is as shown in FIG. In the area shown in FIG. 18, no vibration is observed based on the resonance observed in FIG. 17.

[0089]

As described above, the digitally controlled position servo device with linear phase delay characteristics using the stabilization principle of the present invention significantly improves the behavior of the response of the output X0 to the inherent secondary vibration element. can do.

[Brief explanation of the drawing]

FIG. 1 is a modeling diagram of a general servo system.

FIG. 2 is a mathematical modeling diagram of FIG. 1;

FIG. 3 is an explanatory diagram of a multiple control loop structure of a general servo system.

FIG. 4 is a simplified model diagram of the position servo system shown in FIG. 3;

FIG. 5 is an explanatory diagram of a servo system to which the present invention is applied.

FIG. 6 is an explanatory diagram of a normalized servo system to which the present invention is applied.

FIG. 7 is an open loop frequency characteristic diagram.

FIG. 8 is an explanatory diagram of a digital control servo system having a multi-sample delay element.

FIG. 9 is a frequency characteristic diagram of closed loop characteristics when n=1 and τs=1.79.

FIG. 10 is a frequency characteristic diagram of closed loop characteristics when n=2 and τs=1.08.

FIG. 11 is an explanatory diagram of a digital control servo system with dead time.

FIG. 12 is a characteristic diagram of a low-pass filter having linear phase lag characteristics.

FIG. 13 is an explanatory diagram of an embodiment of a low-pass filter having linear phase delay characteristics.

FIG. 14 is an explanatory diagram of a specific embodiment of the digital control position servo device of the present invention.

FIG. 15 is a hardware configuration diagram of a digital input to output processing section according to the embodiment of FIG. 14;

16 is a behavior diagram of the output response due to the characteristics of the vibratory controlled object in FIG. 5. FIG.

FIG. 17 is a behavior diagram of the output response in a one-sample delay system improved using the optimal sampling period selected according to the present stabilization principle.

FIG. 18 is a diagram of the behavior of the output improved by a linear phase lag low-pass filter.

[Explanation of symbols]

10 Servo system 12 Motor 14 Load (controlled object) 150 Current control section 152 Speed control section 154 Position control section

Claims (4)

[Claims] Claim 1: In a digital control position servo device whose controlled object includes a secondary vibration system with a damping coefficient ζ smaller than 0.3, control accuracy is reduced due to aliasing in response to externally applied commands, disturbances, etc. In order to avoid this, the sampling period Ts is selected to be short, and in order to increase the gain margin of the characteristics of this servo device, the characteristics include:
A digital control position servo device characterized by inserting a linear phase delay characteristic. 2. The device according to claim 1, comprising:
A digital control position servo device characterized in that the linear phase delay characteristic is a dead time characteristic with a delay of n samples (n>1, a natural number). 3. The device according to claim 1, comprising:
A digital control position servo device, wherein the linear phase delay characteristic is a low-pass filter characteristic, such as a Butterworth filter characteristic or a Chebyshev filter characteristic having a predetermined order m. 4. The device according to claim 3, comprising:
A digital control position servo device, characterized in that the cutoff angular frequency of the low-pass characteristic is selected to be lower than the natural vibration angular frequency of the controlled object according to claim 1.