Method and device for controlling a stepping motor

The invention relates to a method and an apparatus for practicing the Method of driving a stepping motor according to the preamble of independent claims.

Stepper motors are today in a variety of applications for positioning in technical equipment used. They allow precise positioning step oh ne position feedback. If high requirements for positioning asked accuracy, is generally an operation of stepping motors in micro step mode required. The user is also at high Verfahrge speeds, the neces to rapidly reach more distant locations are sary high demands on the concentricity and the low vibration of Motor.

The use of stepper motors in microstepping operation is, for example, in

<ul><li>- "Electrical fractional horsepower motors" of H. Moczala et al, Expert-Verlag, 1993, pp. 261-263, and</li><li>- "Stepper motor drives" by F. Prautzsch, Franzis-Verlag, 1988, pages 70-76,</li></ul>

described. The bases are to be set out briefly.

Under microstepping operation is defined as the supply of the stepping motor with sinusoidally stepped motor currents instead of rectangular block-shaped Mo torströmen the full- and half-step mode. The values ​​of the stepped motor streams are in a stored sine table as current values. The individual consecutive table values ​​are used by the motor control sequentially read and generates the motor currents. Each Tabel lenwert corresponds a temporary motor current value of a limited hours th level. In each case a new motor current value per phase causes the motor to a micro-step further to drive. The sine table is constructed so that by Reading the successive table locations always a fixed increment given is.

The number of full steps of a stepper motor has been designed for. A whole Sine period of the motor current corresponds to a rotation of the stepping motor to four full steps. Therefore, the number of places in the table for a full Sine period the number of micro-steps related to four full steps. The Number of table locations per sine period therefore determines the pitch of Micro steps and thus the position resolution of the stepping motor.

The rotational speed of the motor by increasing or decreasing the Step frequency, so by changing the time intervals between the steps, set. The desired for a rotational speed of the stepping motor Required step frequency is calculated as the product "Number of full steps of the stepper motor times the number of micro-steps by number of ge wished revolutions per second ".

In the well-known stepper motor control systems is in a memory example as an EPROM, stored sine table showing the digital power contains setpoints for the motor current. When the stepping motor in a specific Target position is to approach, a CPU calculates the reaching the Zielpositi on required micro-step sequence with each constant pitches in the form a series of stepped sinusoidal motor current values ​​and for the desired Rotational speed, the required pace.

Accordingly, the length of the previously calculated by the CPU train of micro- below are the necessary digital desired current values ​​from aufein other following table locations of the sine table with the calculation of the CPU Neten pace read.

The read-out digital desired current values ​​of the motor currents are of a D / A converter is converted to analog desired current values ​​at the output of the D / A converter as a stepped approximately sinusoidal analog signals with Basic form appear. From the output signals of the D / A converter is in a final stage of motor current (actual value) as a temporal sequence of graded Mo generates torstromwerte. This generated motor current values ​​have corresponding the data read from the sine table digital desired current values, a staircase like steps, which corresponds approximately to a sine curve. On the same manner for each of the phases of the stepper motor for each microstep generates the corresponding phase-shifted motor current. Each new s motor current per phase leads the stepping motor from a micro-step.

The up to a target position calculated number of sinusoidal stepped Mo torstromwerte corresponds to the number of the micro-steps to reach the Motor target position are required. The respective target position reached is on End of a micro-step. The possible resolution of the target position is the size of the micro Shrill and thus by the number of places in the table of given sine table. angefah Shall positions between two microsteps be reindeer, this is not possible with the given sine table. A crate tion of the resolution of the target position is only by a fine divided sine table and thus to achieve through a higher number of smaller microsteps. This requires the CPU to calculate a longer sequence of steps. For different position resolutions therefore have different finely divided Si nustabellen are stored in memories. An increase in the Drehgeschwin speed of the stepper motor is only possible to by an even higher pace Lich. This also requires additional higher CPU capacity.

A disadvantage of the known control systems are the high resolution and simultaneously high rotational speed of the stepping motor required very high step frequencies and the resulting changes to large CPU capacity, because the CPU produces the step frequencies and calculates the step sequences. Accelerated motions yet also require a steady changer alteration of pace. Low accelerations and high speed enco speeds in microstepping operation at higher resolutions require very long sequences of steps and no longer technically feasible, extremely fast Än ments of time between steps. Therefore, high Drehgeschwindig speeds of the stepper motor at reduced positioning resolution, so additional Lich stored smaller sine tables with fewer table values ​​possible. Reducing the positioning resolution and thus lower frequencies however, lead to bothersome loud motor running and vibrations. To calculate the necessary for generating acceleration or deceleration profiles Consequences of motor current values ​​is also high CPU capacity erforder Lich. Alternatively, before a certain number of acceleration profiles the operation can be calculated and stored in a memory. This limits because of the limited space to traversing this low Number one stored acceleration profiles.

It is therefore an object of the present invention to provide a method and an on device for performing the method for controlling a stepping motor specify which at low frequencies and low CPU intensive the highest resolution and excellent concentricity work. Furthermore, should a large number of acceleration profiles and speeds in ge ringer CPU capacity to be adjustable.

This object is achieved according to the in the characterizing part of the dissolved stated independent claims Features. More ago some embodiments of the invention are the subject of the dependent claims.

The inventive method does not work with a variable Schrittfre frequency and a fixed step size, but with an ever-constant step frequency but a variable increment. According to the invention of an egg in nem EPROM stored sine table, which for the digital desired current values contains the motor current, with an always constant sampling per Sinuspe Riode the sine table at least three arbitrarily selected current values read. is from this current command values ​​with arbitrary values ​​jumps by low-pass filtering a smooth sinusoidal curve of current values ​​for the motor current and reconstructed from the actually sinusoidal motor electricity generated.

Compared to the known methods are not consecutive digital desired current values ​​from the sine table read, but the trainees read digital desired current values ​​corresponding to the desired Drivability of the engine selected. Since successive table locations determine the width of a micro-step, is any place in the table of Sinusta belle assigned to a particular motor position. By selecting the read table spaces will therefore vary the increment. Because of selected digital desired current values ​​all with an always constant Abtastfre frequency, are therefore read out at the same rate, is this variation of the step size at the same time the rotational speed of the Stepper motor varies. For large step sizes is given by the constant Sampling inevitably a higher rotational speed than small Increments.

Different constant rotational speed of the stepping motor, he will testifies by constant sampling every nth value from the sine Tabel le is read out, where n is an integer. For n = 1, each current is read reference value of the sine table and thus the lowest frequency of generated sinusoidal motor current, which is the lowest rotational speed of the stepping motor is adjusted. For n = 2, each second current command value is calculated from the sine table read. Because at constant sampling the Schrittwei te is doubled in the sine table, so that the frequency of the sine generated shaped motor current and twice we for the case n = 1. This also the rotational speed of the stepping motor twice as large.

The number n thus represents a temporal compression factor for the read Sine table, and thus the adjustment value for the frequencies of si generated nusförmigen motor currents. The larger the number n, and thus the step width is between the read table locations of the sine table, the higher are the frequencies of the generated sinusoidal motor currents and the more rotates faster, the stepper motor. As long as the increment is set to be constant, is also the rotation speed of the stepping motor constant.

By stepwise increasing or decreasing the step size in the Sinusta belle can be realized accelerations and decelerations. For SET ment an accelerated movement of the motor are therefore kon a constants sampling selected digital desired current values ​​from the sine table read, the distances between the read standings increase the sine table continuously. To set a delayed loading movement of the motor at a constant sampling selected digital Desired current values ​​are read from the sine table, wherein the distances between the read table locations of the sine table which decrease progressively.

At sufficiently high CPU capacity the necessary digital could also calculates current values ​​for the sinusoidal signal continuously from the CPU who to instead read from the EPROM. According object is here but a stepper motor control with low CPU's capacity.

For position and speed control of the stepping motor is a CPU used. It calculates a desired target position of the step motor a series of intermediate positions, the succession to reach the Target position to be approached. The increment can to any vari be ated. It is through the gaps between the tables to be read out places where in the sine table. Because of the constant sampling frequency rotates the stepping motor with big steps at great speed, when to dress NEN steps with little speed.

At any calculated intermediate position by the CPU per phase of the motor the read out from the sine table digital current setpoint determined. Then are from the CPU sequentially according to the calculated from their Zwi ate position sequence associated digital desired current values ​​from the Sinusta belle read at a constant sampling frequency.

The data read from the CPU digital desired current values ​​are now in analo implemented ge values ​​and from a low-pass filter, a smooth sinusoidal The course of the desired current values ​​reconstructed from which the motor currents to Erre account the phases of the stepping motor can be derived. These are in the Un dependent claims various possible embodiments listed. In one of the possible embodiments of the method are selected from the Sine table read-out digital desired current values ​​to a D / A converter out. At the output of the D / A converter appear stepped analog current Setpoints. Its basic form is for small rotational speeds, thus small Increments, approximately sinusoidal. For high rotational speeds, ie for large step sizes, have the stepped analog desired current values ​​large Amplitude jumps and no resemblance to a sine. All stepped Current values ​​appear at the constant sampling frequency. These can with only one specifically tuned to the sampling analog Lowpass filter under certain conditions given here to the smooth Sinusoid reconstruct.

Namely, since the sampling frequency is constant, the sampling theorem (see also applies popular mathematics textbooks or "Electrical fractional horsepower motors" of H. Moczala et al., Expert-Verlag, 1993, pages 278-281). Thereafter, time changing signals by sampling at a constant Abtastfre frequency from the sampled values ​​be completely reconstructed if the Sampling at least twice the value of the highest rekonstruie having-generating frequency.

For the inventive process, this means that the sampling frequency more has to be chosen than twice the size to be generated as the largest Sinus rate, which of the largest adjustable rotational speed Stepping motor in full step mode corresponds. That is, the sampling frequency, wherein with only four samples per cycle of the sine table of engine maximum speed moves. With the same low sampling frequency then all other, so smaller speeds possible namely by scanning of more than four set values ​​per sinusoidal period.

It follows that in the ET to a complete sine period of at least three Samples from the sine table are required as support points for the filter, the sinusoid to be completely reconstructed. In the full step mode a Two-phase stepper motor, so with four steps for a full sine period, but are already four points available, because the sine table in the EPROM is sampled at four points. This is already in full step operation the Forde always fulfilled tion of the sampling theorem. For the microstepping operation with corre accordingly more samples is satisfied the requirement anyway.

To reconstruct a sinusoidal curve of the current values ​​for the Controlling the stepping motor is used, a suitable low-pass filter. Out the stepped analog desired current values ​​reconstructed this low-pass filter a smooth sine curve. The low pass filter has to filter out all frequencies which do not of generated sine frequency correspond. For this, the filter must Cutoff frequency of the low-pass filter is less than half the sampling frequency of the CPU to get voted. In order to simplify the low-pass filter can be a higher constant sampling be used as the lowest possible. At the Output of low pass filter then appears a smooth sine wave electricity Setpoints, which is passed to the input of an output stage that the resulting generated motor current.

In another possible embodiment of the method are selected from the Sine table read-out digital desired current values ​​instead of a D / A Converter fed to a pulse width modulator. This produces for each is Henden digital desired current value from the sine table at its output high-frequency pulse train with a constant pulse width modulation frequency, in fol lowing briefly called PWM frequency. The analog average of these produced Pulse train is in each case the level of the input signal proportionally.

The generated by the pulse width modulator high-frequency pulse trains are on out the input of a low pass filter. The constant PWM frequency of the Pulse width modulator generated pulse trains must be greater than the sampling frequency the CPU, but at least the same size, and thus be more than twice as large as the cut-off frequency of the lowpass filter. The low pass filter forms at all incoming high-frequency pulse sequence the analog mean and filtered out both the PWM frequency and the sampling frequency out. Thus rekon structed the low pass filter directly a smooth analog sine wave for power Target values ​​corresponding to the read out by the CPU desired current values. At the Output of low pass filter therefore appears a smooth sine wave electricity Setpoints, which is passed to the input of an output stage.

For each phase of the stepping motor can be read from the current Setpoints the sine table in one of the ways described, either through a combination of D / A converter with a subsequent low-pass filter or by a combination of pulse-width modulator followed Tiefpaßfil ter, the corresponding phase-shifted, actually sinusoidal waveforms the current values ​​produced and performed on a respective output stage. The outputs of the Power amplifiers provide phase-shifted sinusoidal motor currents.

With previously known control systems drives a stepping motor according to the sine shaped stepped motor current values ​​in small jerky steps his Target position. At high speeds and to reduced positi onsauflösung, so bigger steps, performs an operation with previously known to controls to disturbingly loud operating noise of the engine and disturbing Vibrations. In contrast, moves a step motor with a drive by the inventive process due to the fact sinusförmi gen motor currents without jumps and so quiet and vibration to his Target position.

The method is suitable for positioning for all permanent-magnet magnet stepper motors, ie also for hybrid stepper motors and linear stepper drives, and generally for all motors with at least two coils and pha senversetzten motor currents, such as synchronous motors. Equally an application of the method to multiphase stepper motors possible in the example for each phase from the sine table the necessary digital Desired current values ​​are read out with the required phase shift and therefrom for the different phases of the motor according to the invention it ford variable-phase motor currents are generated. With a lei erful CPU, it is also possible for the digital desired current values ​​of the sine calculate table continuously, instead of being taken from a sine table.

Opposite known stepper motor controls in microstepping The inventive method works for all speeds and all Fine positioning is always the same low sampling frequency. Therefore, with the new procedure at a constant sampling frequency only by default one very finely divided sine table be the position resolution be increased arbitrarily, even with constant maximum speed. There According to the invention a single finely divided sine table for setting all Increments sufficient eliminates the erforder the prior art methods Liche memory for storing various sine tables for various dene increments.

The new process is characterized by a very large number of possible Speeds and acceleration profiles with arbitrarily high Positionsauf solution from with optimized concentricity. This is the new method clearly superior to the previous step motor drivers, because these high speeds could only reduce the position resolution or by increasing the pacing rate with the previously discussed disadvantages be achieved.

The size of the constant sampling frequency determines the number of Rechenope tions of the CPU per unit time and thus affect the required CPU Capacity. Since the inventive method at all speeds and resolutions with the same low, constant sampling works is also always requires only the same small CPU capacity.

The invention is based on two embodiments using the schemati 's drawings explained in more detail.

Show it:

<b>Fig.</b> 1 diagram of a first device of the invention with digital -Analog conversion by a D / A converter;

<b>Fig.</b> 2 diagram of a second device according to the invention with digital Analog conversion by a pulse width modulator.

In <b>Fig.</b> 1 is in an EPROM <b>1</b> a sine table stored, the digital, the contains current values ​​for the motor current to be generated. The number of Table spaces determines the potential position resolution of the motor. A CPU<b>2</b> reads at a constant sampling frequency per sine period of at least three selected digital desired current values ​​from the sine table in the EPROM <b>1</b> out.

In the present example it is assumed for simplicity that a kon constant rotational speed to be set for the motor. For this purpose, the digital desired current values ​​from the sine table with equal intervals read between the read out of the standings from the sine table. In the present example it is believed that the sine table with a large KISSING increment is read. Thus, the frequency will produce the the motor current high and thus the rotational speed of the stepper tors be high.

The CPU of the <b>2</b> read from the sine table digital desired current values be a D / A converter <b>3</b> out. At the output of the D / A converter<b>3</b> he seem stepped analog desired current values ​​which in separately <b>Fig.</b> 1 schematically are shown. They have an approximately sinusoidal in the illustrated case, shaped course on. As per period of the sine few samples be made, the analog output signals of the D / A- converter <b>3</b> huge jumps on.

They are on the input of a low pass filter <b>4</b> out at its output this in <b>Fig.</b> 1 shown separately completely reconstructed sinusoidal signal Desired current values ​​appear. This smooth sinusoidal signal is applied to a power amplifier<b>5</b> out that the sinusoidal motor current for the first phase of from two-phase stepping motor <b>6</b> generated.

For the second phase of the stepping motor <b>6</b> the motor current is also to the method according to the invention produced in the manner already described. For this purpose, the CPU reads <b>2</b> selected digital desired current values ​​for the motor current the second phase of the step motor <b>6</b> from EPROM <b>1</b> out. The digital Current values ​​are to phase by ninety degrees from de NEN for the first phase of the stepping motor <b>6</b> read from the sine table. The read-out digital desired current values ​​are applied to a D / A converter <b>7</b> out, stepped analog desired current values ​​appear at the output, the ninety degrees phase shifted from the after the D / A converter <b>3</b> are appearing.

The output signals of the D / A converter <b>7</b> be the input of a low -pass filter <b>8th</b> performed at the output of a fully reconstructed sine signal appears. This is phase-shifted by ninety degrees with respect to which, after the low pass filter <b>4</b> appearing. The output of the low -pass filter <b>8th</b> is on the input of a power amplifier <b>9</b> out that the sinusförmi gen motor current for the second phase of the step motor <b>6</b> generated.

As long as both phases of the stepper motor <b>6</b> with the generated sinusoidal, phase-shifted motor currents are supplied, rotate the stepper motor <b>6</b> with a constant speed. The final position of the stepping motor<b>6</b> becomes determined by the two last, phase-shifted by ninety degrees from the CPU <b>2</b> read digital desired current values ​​in the sine table, waiting to be Motor currents for the two phases are processed. This rotates the motor to its final position.

As long as the stepping motor <b>6</b> no sinusoidal, phase-shifted motor currents are supplied, so long as the CPU <b>2</b> No further digital Desired current values ​​are read from the sine table, it stops. Only by reading further digital, phase-shifted by ninety degrees Current values ​​are again sinusoidal, phase-shifted motor currents generated and the stepper motor <b>6</b> again set in motion.

In <b>Fig.</b> 2 are compared with the <b>Fig.</b> 1, the D / A converter <b>3</b>.<b>7</b> by the pulse widths modulators <b>10</b>. <b>11</b> replaced. This embodiment is particularly advantageous because the pulse width modulators in many commercially available CPU Blocks are included. This saves cost and circuit complexity. A CPU <b>2</b> reads out the in an EPROM <b>1</b> stored sine table a constant sampling of selected digital desired current values. Again must to fulfill the sampling theorem, the condition must be fulfilled that per sine period of at least three selected digital desired current values ​​from the Sine table are read. These are on the input of the Pulswei width modulator <b>10</b> out.

This sets the digital desired current values ​​in high frequency pulse trains with kon stant PWM frequency in order. The analog average of the pulse train at the output the pulse width modulator <b>10</b> respectively the height of its digital input signal proportional. That is, for example, that a digital input signal zero at the output of pulse width modulator <b>10</b> generates a pulse sequence, their average due to the pulse width of the individual pulses and zero is. For a digital input signal is greater than zero widen the positive Shares of the pulses of the output signal such that the average of the new pulse sequence is greater than zero and proportional to the level of the input signal. For a digital input signal is less than zero to narrow the positive On parts of the pulses of the output signal such that the mean value of the new pulse sequence is less than zero and is proportional to the level of the input signal. The at the output of pulse width modulator <b>10</b> appearing pulse sequence at very large step sizes is separately in <b>Fig.</b> 2 is shown schematically.

The from the pulse width modulator <b>10</b> be generated high-frequency pulse trains to the input of a low pass filter <b>4</b> out. The PWM frequency of the Pulse width modulator <b>10</b> Pulse sequences generated at least equal to the Ab sampling frequency of the CPU <b>2</b> or greater and more than twice as high as the Cutoff frequency of the low pass filter <b>4</b>, The low pass filter<b>4</b> forms to enter into any the high-frequency pulse sequence the analog mean and filtered out both the PWM frequency of the pulse width modulator <b>10</b> and the sampling frequency the CPU <b>2</b> out. This means that all frequencies that are not to the reproduzie -generating sinusoidal signal corresponding filtered out. This provides the low-pass filter<b>4</b> a complete smooth sinusoidal signal of the current values ​​at its output. The low-pass filter at the output of <b>4</b> appearing, completely reconstructed, smooth sinusoidal signal is separately in <b>Fig.</b> 2 shown. This smooth sinusoidal signal is to the input of an output stage <b>5</b> out therefrom the motor current for the first phase of a stepping motor <b>6</b> generated.

According to the motor current for the second phase of the step motor <b>6</b> generated. These are from the CPU<b>2</b> at a constant sampling frequency of EPROM <b>1</b> read out selected digital desired current values ​​and on the Input of the pulse width modulator <b>11</b> out. The digital desired current values be to phase by ninety degrees relative to those for the first Phase of the stepping motor <b>6</b> read from the sine table.

For each incoming digital desired current value provides the Pulsweitenmodula gate <b>11</b> at its output a high-frequency pulse train with the A output signal proportional analog average. The from the pulse width modulator<b>11</b> Pulse sequences are generated at the input of a low pass filter <b>8th</b> out. This reconstructed, as already above the low-pass filter <b>4</b> described from the incoming pulse trains a complete smooth sine wave electricity Setpoints. The low-pass filter at the output of<b>8th</b> appearing reconstructed Sinusoidal signal is phase shifted by 90 degrees relative to the in <b>Fig.</b> 2 separately illustrated reconstructed sinusoidal signal after low-pass filter <b>4</b>, The Si nussignal at the output of low pass filter <b>8th</b> is on the input of a power amplifier <b>9</b> out that the sinusoidal motor current for the second phase of the resulting stepper motor <b>6</b> generated.

Also in this example, the stepping motor moves <b>6</b>As long as its two Pha sen are supplied with sinusoidal, phase-shifted motor currents. The approached end position of the stepping motor <b>6</b> is also the case the last read out from the sine table, digital desired current values ​​be agrees to be implemented in the motor currents.

LIST OF REFERENCE NUMBERS

<b>1</b>

EPROM

<b>2</b>

CPU

<b>3</b>

D / A converter

<b>4</b>

low pass filter

<b>5</b>

output stage

<b>6</b>

biphasic stepping motor

<b>7</b>

D / A converter

<b>8th</b>

low pass filter

<b>9</b>

output stage

<b>10</b>

Pulse width modulator

<b>11</b>

Pulse width modulator