DE102014108541B4 - Drive system with increased accuracy in setpoint signal transmission (method and system) - Google Patents

Abstract

Method for operating a drive system (100) wherein setpoint signals are transmitted via a setpoint channel (90) in the drive system to a drive (101), and wherein - the drive system (100) reads out an input memory (10) with a higher clock (f3), as input signals (S1) are provided to the input memory of a jittered signal source (1) available; - The read setpoint signals (S3) are passed through a low pass (12), whereby the setpoint signals (S3) undergo interpolation and as filtered setpoint signals (S4) are corrected, or are, wherein the low-pass filter (12) is sized and set that errors are locked out of a time inaccuracy; - The filtered setpoint signals (S4) in the setpoint channel (90) after a further scan (13) with a clock (f1 *), which is smaller than the higher clock (f3), further processed, resulting due to time shifts errors in the scanned Setpoint signal (S5) were avoided, are or are and a bandwidth of the setpoint signal (S5) is limited so that no low-frequency signal parts from an inaccurate or fluctuating clock of the signal source (1) for the drive (101) form.

Description

The invention relates to a method for improving the signal quality of at least setpoint signals that are provided by a non-reliable synchronously operating source available, for. B. a bus system under load, in an electric motor drive system.

In many modern systems, the setpoint values for the drive systems are made available via bus systems. In this case, the bus is used for different information for a plurality of bus subscribers (from or to the subscribers). As a result, the signals are not transmitted in a fixed time frame (synchronously), but transmitted (asynchronously) depending on the bus load with an undefined time offset.

This leads to temporal inaccuracies in the signal transmission (so-called jitter in the setpoint). This is also transferable to the actual value.

A shaftless drive system as a rotary printing machine describes DE 42 14 394 A1 (Asea). This rotary printing machine with drive system has at least two drive groups, as individually driven pressure point groups. The drive groups have a drive control consisting of drive and motor and a drive controller. The drive groups receive their position reference (leading axis) directly from the folder. The drive controllers of the drive groups are also connected via a drive bus. The drive controls are interconnected via a data bus and to an operator and data processing unit. This data bus is used to specify target values and to manage the print groups.

The drive controllers of the drive system enable torque control, speed control (speed control) or position control (angular position control) of the driven axle. In the case of high requirements of angle-synchronous synchronization, as they exist in drive systems in machine tools and in printing presses, usually position controls (angle controls as position controls) are used.

The digital controllers are preferably equipped with fast or powerful digital signal processors. Such digital drive controllers can execute a closed-loop control cycle in a very short computing time, preferably in 250 μs or in a shorter cycle time.

In such drive systems three-phase motors are preferably used. The electrical drive energy is supplied to the individual motor via a power electronics circuit, preferably with frequency converter function. The power electronics circuit is controlled by the digital drive controller.

The actual value transmitter for the individual drive is often mounted on the motor axis.

However, arrangements are also known in which an actual value transmitter is attached to the load driven by the motor. For example, in printing machines, it is advantageous to attach a high-resolution position sensor to the torque-free end of the driven printing cylinder. Then the actual value also passes through the bus system.

Drive systems are known in which the setpoint values of the control must be transmitted as precisely as possible to the drives for precise operation of the individual drives. As is known, this is done by synchronization of the control with the drives, so that there are no fluctuations in the transmission rate (frequency). Then the signal arrives, as mentioned above, in time unadulterated in the drive.

However, this requires a high technical effort in the bus system to synchronize the spatially separated systems to each other and is possible only with closely matched components.

Inaccuracies of the setpoint signal transmission (as setpoint delivery delays) can lead to low-frequency setpoint fluctuations at the receiver. These fluctuations are noticeable in poor production quality, e.g. in clearly visible color variations in the print result in multi-color printing.

From the DE 600 30 932 T2 This results in a TDMA packet transmission system. It works in satellite communication, cf. There Abs. 05 and 06. An oversampling is there in para. 023 and 2 explained. The local TP (low pass) 20 is followed by another memory, from which the sub-scan 26 selects an optimal sample, in the search a search device 25 helps.

The invention is inter alia based on the object of inventing a method for operating a drive system, with which the accuracy is improved at least in the setpoint signal transmission. Furthermore, the (additional) task, the signal quality of setpoint signals, which are transmitted via a not always synchronous or even slow-working setpoint source, cheap to improve (its price be worth).

The invention solves this object (s) with claim 1 (for the desired value).

The command value signals are provided by a setpoint source (eg a bus system) to an input memory of the drive system. Here, a setpoint channel is provided which processes the jittery setpoint signal and delivers it to the drive of the drive system without jitter. One frequency is the supplying reference frequency f ₁ (without jitter).

The setpoint channel reads its input memory at a substantially higher frequency (claim 3) than the signals are provided to the input memory. These read signals are passed through the sampling clock (rate) via a low-pass filter. Here, these signals are corrected by interpolation to eliminate the effects of inaccuracy of time. The filtered signals at the output of the low-pass filter are forwarded without error due to time shifts to the drive and processed there. The sampling frequency can be that of the input, ie f ₁ as the reference frequency.

Jitter in a signal is variations in clock phases, more specifically the phases of the clock pulses in the signal. If, for example, the scanning signal of a sensor is used as the signal, then this should be jitter-free for a high measuring precision.

Another advantage of the invented method is the automatic adjustment of the pickup (the drive system) to the current performance of the bus system.

Yet another advantage of the method is that the application is possible regardless of the specific bus system for all discrete-time operations.

An advantage is the accurate transmission of the actual or setpoint signals even in systems that do not operate in sync with each other, but subject to the said jitter, at least a strong jitter, or work completely asynchronous to each other. This is significantly smaller than the reference frequency f ₁ , usually in the (amount) range of less than ± 10%, in particular (in terms of amount) less than ± 5% of f ₁ .

Dependent claims contain advantageous developments, incorporated herein by reference.

The invention will be discussed in more detail below with reference to the drawings by way of example.

1a is a timing diagram S of the drive system 100 with an asynchronous setpoint signal S (t), discrete time as "signals" (in the plural) understood;

1b is a timing diagram ES of the drive system with a synchronous sampling as ES (t) at synchronous T ₁ sampling;

1c is a timing diagram of the drive system 100 as the difference between S and ES;

1d is a timing chart E0, the waveform being corrected as a function of Δf by interpolation as E0 (t);

1e is a timing diagram S - E0 of the desired value channel 90 as the difference between S (t) and E0 (t);

2 is a block diagram of the setpoint signal transmission;

2a Fig. 10 is a block diagram of an actual value signal transmission;

3 is an embodiment in the form of a block diagram with several variants of the implementation of the "interpolation" as the elimination of the consequences of the jitter, .DELTA.f.

1a to 1e show an operation of an embodiment of the method for the operation of a drive system 100 with a setpoint channel 90 and a drive 101 from controller, actuator and actuator. A bus system 1 does not deliver the setpoint S _{1 deterministically} .

The block diagram in 2 has an input memory 10 for the setpoint S ₁ (with the jitter Δf) An oversampling function 11 is planned. Likewise a low pass 12 (as a filter with the steepness corresponding to at least second order). For sampling is a sampling function 13 for the drive 101 in the drive system 100 intended.

The low-frequency setpoint fluctuations occur as follows: Ideally, the setpoints are generated at the clock rate f ₁ (from the bus system 1 delivered and the input memory 10 made available). Due to the sampling theorem, the target value signal must ensure that only signals from the frequency range 0 ... f _1/2 included. The signals are sampled at a clock rate f ₁ . If everything runs smoothly, the signals of the setpoints arrive in the drive 101 unadulterated.

The delivery time is not constant like 1a indicates that it is subject to a fluctuation f ₁ ± Δf due to the bus transit times. The lowest clock rate is f _g = f ₁ - Δf. This results in the Setpoint signal S additional frequency components. The frequency range of the desired value signal is expanded to about 0 ... f _1/2 + .DELTA.f. The frequency .DELTA.f above f _1/2 is outside the permitted frequency of the sampling theorem, and produces at an f ₁ sampling a low-frequency fluctuation or oscillation.

The mistake in the frequency range _f1 / 2 - .DELTA.f ... f _1/2 + .DELTA.f noticeable. During sampling in the drive with the constant clock f ₁ , additional low-frequency components in the frequency range 0... Due to their low frequency, these components have a disruptive effect on the production result, such as uneven color distribution, material thickness fluctuations, surface changes.

One solution may be the synchronous transmission of the setpoint signals. Another is the improvement of the asynchronous transmission, so that even with the presence of a (strong) jitter a good signal quality results for the setpoint.

The setpoints S ₁ are as usual from a setpoint source, eg. B. the bus system 1 transmitted. The jittered clock is f ₁ ± Δf, the lowest clock rate is f _g . These signals are initially jittered in the input memory 10 powered by the drive system, such as 2 it shows.

Due to the fluctuations of the delivery clock, the setpoint values S _{2 are} in the input memory 10 (at the output of the memory) compared to the original setpoints S ₁ falsified. Originally, before a frequency range f 0 ... _1/2, the falsified signals have a first approximation, the frequency range 0 ... f _1/2 + .DELTA.f. The mistakes especially in the frequency range _f1 / 2 - .DELTA.f ... f _1/2 + .DELTA.f noticeable.

The input memory 10 is therefore read out with a (substantially) higher clock f ₃ > f ₁ , for example the factor 2 or greater, or f ₃ ≥ 4 · f ₁ .

At a the assumed clock f ₃ = 4 · f ₁ ... f is the bandwidth of 0 _3/2, ie 0 ... 2 · f _1, that is significantly higher than the input frequency range 0 ... f _1/2 + .DELTA.f , since Δf << f ₁ .

Thus the readout signal clocked from the input memory S ₃ with the clock f ₃ contains the original set point signal S ₁ with the errors in the frequency range f _1/2 - .DELTA.f ... f _1/2 + .DELTA.f. This signal S ₃ will now be with a (in particular steep edge) low pass 12 filtered so that all components above f _g = _f1 / 2 - .DELTA.f be transmitted only with such a small amplitude that they no longer play a role in the further processing.

At the output of the signal S ₄ to the range 0 ... f _1/2 is - at .DELTA.f. This signal can be used with the sampling rate f ₁ * of the drive 101 be scanned, for example, with the frequency f _{1 of} the drive system 100 , But it is also possible faster sampling with higher frequency, but below f ₃ .

This signal S ₅ has then also the bandwidth of 0 ... _f1 / 2 - .DELTA.f. The interfering error components of the frequency range _f1 / 2 - .DELTA.f ... f _1/2 + .DELTA.f are no longer included in the signal S. ₅

The signal S ₅ goes to the drive 101 of the drive system 100 (Function 13 can also be assigned to the drive). The drive 101 has the controller R, the actuator S and the actuator as eg machine M (as DAM, DSM or GM).

Difficulties with low-frequency fluctuations in the production process appear detected and can be considered virtually eliminated. Amounts not included in the setpoint signal between the samples are taken from the lowpass 12 "Interpolated".

A setting 12a the low-pass filter 12 happens automatically. From the analysis of the transmission times of the input signal S ₁ , the fluctuation Δf of the clock of the supply is determined. This results in the over 12a bandwidth to be set to 0 ... _f1 / 2 - .DELTA.f for the steilflankingen low-pass (an attenuation corresponding to a filter of at least second order, that is of at least 12 dB / octave or 40 dB / decade).

This bandwidth becomes the parameterization of the low-pass filter 12 used.

The same procedure can be used in the transmission of actual values with a small clock, analogous to the 2 when 2a ,

The actual value is first "high frequency" from the input memory 10 ' sampled (more than a factor of 2 compared to the delivery frequency), via a low pass 12 ' passed and with a low clock rate f _{a of} the sample 13 ' output.

In determining the cutoff frequency, the clock f _a and the possible fluctuations in the data transmission, as stated above, are taken into account. As a result, the bandwidth of the actual value signal is ex ante limited so that there is no formation of low-frequency actual value signals from inaccurate clock (in the sense of a frequency).

LIST OF REFERENCE NUMBERS

S

Sollwertsignal (mit Jitter in T_g)

ES

Signal nach der synchronen Abtastung T₁

S – ES

Differenz zwischen dem Sollwertsignal und dem Signal nach der synchronen Abtastung

E₀

Sollwertsignal nach der Interpolation 12

S – E₀

Differenz zwischen dem Sollwertsignal und dem Signal nach der Interpolation 12

T_g = 1/f_g

niedrigster Takt

Zu Fig. 2 (für den Sollwert)

10

Eingangsspeicher des Antriebssystems

11

Oversampling (als Funktion gezeichnet)

12

Tiefpass

13

Abtastung (als Funktion gezeichnet)

Zu Fig. 2a (für den Istwert)

10'

Istwert Eingangsspeicher des Antriebssystems

11'

Oversampling (als Funktion gezeichnet)

12'

Tiefpass für den Istwert

13'

Abtastung (als Funktion gezeichnet)

To Fig. 1

S

Setpoint signal (with jitter in T _g )

IT

Signal after the synchronous sampling T ₁

S - ES

Difference between the setpoint signal and the signal after the synchronous sampling

E ₀

Setpoint signal after interpolation 12

S - E ₀

Difference between the setpoint signal and the signal after interpolation 12

T _g = 1 / f _g

lowest clock

To Fig. 2 (for the target value)

10

Input memory of the drive system

11

Oversampling (drawn as a function)

12

lowpass

13

Sampling (drawn as a function)

To Fig. 2a (for the actual value)

10 '

Actual value Input memory of the drive system

11 '

Oversampling (drawn as a function)

12 '

Low pass for the actual value

13 '

Sampling (drawn as a function)

Claims (10)

Method for operating a drive system ( 100 ) where setpoint signals via a setpoint channel ( 90 ) in the drive system to a drive ( 101 ), and wherein - the drive system ( 100 ) an input memory ( 10 ) with a higher clock (f ₃ ), as input signals (S ₁ ) the input memory of a jittered signal source ( 1 ) to provide; - The read setpoint signals (S ₃ ) via a low pass ( 12 ), whereby the setpoint signals (S ₃ ) undergo interpolation and are corrected as filtered setpoint value signals (S ₄ ), were or are, the lowpass ( 12 ) is sized and set to lock errors from a time inaccuracy; - The filtered setpoint signals (S ₄ ) in the setpoint channel ( 90 ) after another scan ( 13 ) with a clock (f ₁ *) which is smaller than the higher clock (f ₃ ), are further processed, wherein errors due to time shifts in the sampled setpoint signal (S ₅ ) have been avoided are and are and a bandwidth of the setpoint signal (S ₅ ) like this is limited, that no low-frequency signal parts from an inaccurate or fluctuating clock of the signal source ( 1 ) for the drive ( 101 ) form. The method of claim 1, wherein the low pass ( 12 ) is provided with a damping according to a filter at least second order. Method according to one of the preceding claims, wherein the input memory ( 10 ) is read out with the higher clock (f ₃ ), wherein the two clocks (f ₁ , f ₃ ) at the input memory ( 10 ) differ by at least a factor of two to provide an oversampling function ( 11 ) to effect. Method according to one of the preceding claims, wherein the filtered signal (S ₄ ) at the output of the low pass ( 12 ) with a sampling frequency (f ₁ *) of the drive ( 101 ) which is less than the higher clock (f ₃ ). Method according to claim 4, wherein the signal (S ₅ ) at the output of the last sampling function ( 13 ) Having a range 0 ... f _1, / 2 - is formed .DELTA.f without low-frequency fluctuations in the range .DELTA.f. Method according to one of the preceding claims, wherein a setting ( 12a ) of the low pass ( 12 ) automatically via an analysis of transmission times of the input signal (S ₁ ) to the input memory ( 10 ) is carried out by the fluctuations (.DELTA.f) of the delivery clock of the setpoint from the signal source ( 1 ) be determined. The method of claim 6, wherein the range 0 ... _f1 / 2 - .DELTA.f (at a parameterization 12a ) of the low pass ( 12 ) is set. Method according to one of the preceding claims, wherein an actual value is first sampled with a first frequency (f ₃ ) ( 11 ' ), then another low pass ( 12 ' ) is filtered out and read or output with a second clock (f _a ) ( 13 ' ). The method of claim 8, wherein the second clock (f _a ) has a lower frequency than the first frequency (f ₃ ). The method of claim 8 or 9, wherein in determining a cutoff frequency of the second clock (f _a ) and variations in the actual value transmission are taken into account.

Images