Using Stepper Motion By: Chuck Raskin P.E. CMCS

Transcription

1 Using tepper Motion By: Chuck Raskin P.E. CMC What s It All About? I ve noticed that in most motion control write-ups, servo systems seem to be favored over stepper systems. teppers appear to gain attention only when there are major breakthroughs like Micro-stepping or an0-tepping, for example. A possible reason that servo systems appear to be more popular in industry could be because servos motors are relatively easy to analyze and are more forgiving as far as overload and following error are concerned. tepper system capabilities are generally misunderstood concerning their capabilities. ervo systems require at a minimum, a position loop, phasing capability, and some form of gain structure (lead, lag, PID, PD, or other). In addition, servos need to be operated in some form of consistent time interval (update period) in order to maintain smooth motion and high speed mathematics. But most engineers don t realize that even though the tepper does not require direct Gain control, their use is more complex due to the open-loop (zero feedback) nature of their operation. Therefore, steppers require more preliminary analysis. It is generally because of this that servos wind up in many applications otherwise ideally suited for steppers. Load changes, acceleration rates, top velocity, high velocity torque reduction, zero following error, and other factors all need to be taken into account to successfully apply a stepper application. Most designers do not realize that a functionally comparable stepper system generally costs less than a servo based one. uccessful stepper application requires at least a basic understanding of the internals of stepper motor operation, the various stepper motor amplifiers (translators) available, allowable stepper motor control methods, and a good understanding of the system in which the stepper is to be used. The objective of this paper, therefore, is to cover these stepper issues in as simple a manner as possible, to increase the understanding of stepper capability and use, and to close up any holes in the stepper puzzle. The better understanding you have of what the stepper is, and how it works, the easier it will be to incorporate it into your design. How Do tepper Motors Work? The first rudiment to understand is the basic difference between a servo motor (as it is called in the industry) and a stepper motor. For the purpose of clarity, all servo motors (DC, Brushless, DC/AC ervo, etc.) will be grouped into a motor class called DC Brush. The main difference between a servo motor and a stepper motor is that when a voltage is applied to a servo motor, it will develop both torque, and rotation. When that same voltage is applied to a stepper motor, however, it will develop only torque. This difference is based on

2 the fact that the DC Brush motor is able to self-commutate. Commutation Commutation is the principle by which the amplitude and/or direction of the current flowing in one or more of the electromagnetic coils within a motor is altered. Thus, the attracting and repelling magnetic force fields existing between the motor s stator and armature are altered allowing rotational torque to be developed. DC Brush motors have current altering capability built in which allows the armature of the motor to begin to rotate as soon as a DC voltage is applied to the motor. The stepper motor, however, is designed with no internal commutation capability, and until its internal currents are externally altered by some means, it will only develop static torque. As with all DC motors, the stepper motor operates within the basic principle of magnetic attraction and repulsion. Most individuals at some point in their lives have observed this magnetic property by playing with magnets to make them either push away from each other or attract one another. The following is a simple illustration showing the similarity and difference between the operation of the DC servo motor and the stepper motor. Figure shows a simple DC motor and the rotational force produced when the motor coils are energized. It must be remembered that the internal commutator and brushes of the DC Brush motor perform the current switching within the motor coils, thus producing armature rotation. As the brushes move along the commutator, they touch different commutator segments. This action applies current to certain armature coils, and dis- charges current from others, just like a switch. Figure shows a simple stepper motor and the relative detent positions of the armature. Torque is produced when the stepper motor coils (see Fig.) are energized. It must be remembered that it is external commutation of the stepper motor that performs the current switching within the motor coils thus producing the rotation of the armature. Attracts Repels Attracts Fig.a Fig.c Attracts Repels Repels Repels Attracts Changin g from to Fig.b Fig.d Changin g from to Attracts Repels Fig. Principle of DC Motor Operation. et up two magnets to attract each other (call them #, and #).. Position them just far enough apart so that they do not physically move toward each other.. Place a third magnet (#) of the same field strength as # next to magnet # in the identical orientation. You may need to force them together. ote that magnet # will now move towards magnets # and #. The magnetic field of magnet # is now strengthened by that of magnet #. This is known as magnetic field addition. In turn, if magnet # is turned around and its magnetic field is made stronger, it will cancel the field strength of magnet # and repel magnet #. This would be magnetic-field subtraction.

3 The stepper motor operates on this principle, but it uses an electromagnet in place of magnet #. The stepper motor is constructed with a series of magnets on both the stator and the armature, and electromagnets on the stator. 7 7 As you can from Figure, the magnets will come to rest in some natural magnetic orientation called its detent position. If you turn the shaft of the stepper motor by hand, you will feel the magnets try to maintain this detent position. However, when you apply enough rotational force to the shaft to overcome the magnetic attraction, you will feel the armature jump to the next detent position. Figure shows a stator with eight magnets and an armature with six magnets. This yields a total of detent positions. As the armature is turned counterclockwise from Fig. 7 Fig.a Fig.c Principle of tepper M otor Operation C A A' Fig.a Fig.c 7 Fig.b Fig.d the position in Figure A to the position in Figure B, then to positions C and D, different magnet sets will exhibit a stronger attraction and will pull the armature into new detent positions. In this example, there are three full-step positions between each of the stator magnets. To provide some manner of controlling the armature position, the principle of the electromagnet is used to generate the force required by the stepper motor to sequence from one detent position to the next. A series of coils surrounds each of the stator magnet poles, and by having current flow through these coils with certain amplitudes and in given directions, the magnetic attraction and repulsion of the magnetic fields will cause the armature to rotate to new positions (Fig ). By changing the coil currents, and/or current direction, at the proper speed (frequency), continuous rotation will result. Figure shows that rotational motion can be achieved in the stepper motor by combining static magnetic, and dynamic electromagnetic forces. In Figure A, the motor is oriented with armature magnets and aligned with stator magnets and 7 respectively. If a current is passed through electromagnetic coils C and C, the motor will exhibit only a holding torque. If the current is then removed from coils C and C and passed through coils B and B, the armature will then rotate and align armature magnets and with stator magnets and, respectively (Figure B). The effect of altering, or removing existing coil currents is the external commutation required by the stepper motor to produce rotation. C B B D' D' 7 C' C 7 C' D B' D B' B D 7 A A' D' B' C' Fig. Principle of tepper M otor Operation C B D 7 A A' Fig.b A A' Fig.d D' B' C'

4 If current commutation is halted, the stepper motor will stop rotating (stepping). At this point, an alignment of the armature and stator magnets will remain in the existing orientation, until coil current is once again altered. The typical stepper motor contains 00 detent positions in full-step mode (0 stator poles) yielding a resolution of. degrees-per-step. Operation of the stepper motor between these full-step positions (called half or micro-step operation) is accomplished by the stepper motor amplifier (translator). However, remember that a stepper functions within electromagnetically generated fields, and therefore, it still suffers from all of the normal losses associated with any other type of motor, including viscous damping, hysteresis, eddy currents, back-emf, resistive and inductive losses, and more. Many of these losses are only evident when the stepper is in motion, since it takes either varying current or rotational motion to generate them (eddy currents for example). Windage is a rotational loss more commonly referred to within a combined group term known as Viscus Damping. Basic tepper Motor Calculations tepper motors, like any other motor, contain winding Resistance and Inductance. You need to understand these two properties and their effect on the response of a stepper system. The DC static current in a stepper motor is limited only by its coil resistance. In turn, it behaves according to Ohms law: E = I x R For a stepper motor, the higher the coil resistance of the motor windings, the lower the winding current (for a given voltage). The lower current will allow higher motor speeds to be achieved, but with a lower output torque. Commutating the electromagnetic fields inside the stepper motor provides not only motor rotation, but also induced currents. To define inductance is a paper in itself, and there are many good resources. But suffice to say that inductance is a property of all coils, and the currents produced by inductive action tend to oppose the current flow which is creating it. In other words, if current is flowing in a wire, and the voltage that is producing that current is removed, the current produced by the inductive property of the wire will try to maintain that current by producing a generated voltage large enough to do so (this phenomena can shown by observing an arc produced when relay contacts remove power from an inductive load). In a motor, this voltage is know as its counter EMF, or Kv. What will be discussed now, is the effect of a motors inductance on the stepper motors performance. Current in a motor winding is given by the following relationship: Rt E I coil= (-e L ) R Where: I = Coil Current (Amps) E = Terminal Voltage (Volts) R = Coil Resistance (Ohms) L = Coil Inductance (Henry) t = Elapsed Time (seconds) e = atural Log (.7)

5 This equation relates the instantaneous current (I) through the coil as a function of time (t). The ratio of wire resistance to coil inductance determines the time-rate-of-change of current for a given applied voltage. The static DC current is limited by (E/R). As the coil DC resistance increases, the static DC motor current decreases (with constant E). As the coil resistance increases due to a rise in wire temperature, the time that it will take to reach the limiting static DC motor current decreases. otice that this reduction in current is not by design, but by temperature. This results in a motor power reduction as well. The coil inductance, on the other hand, only works on the time portion of the equation. Therefore, as the inductance goes up, so does the time it takes to reach T erm inal V oltage T erm inal V oltage the static DC current value, and viceversa (refer to Figure ). Charging Coil Current Discharging Coil Current Time comes into play during the motor coil switching operation ( commutation). As coil voltage is switched on and off, the current in the coil tries to charge up to the E/R value, or discharge to zero. The time required to do this is determined by the motors Time Constant (R/L), and is approximately five times this period. ince the motor coil resistance and inductance do not change once the motor is built, this time period is fixed except for the effects of temperature and possible motor saturation. Current --> Current --> Current --> Time --> Time --> tep Pulse Voltage Time --> tep Pulse Voltage Time --> Current --> Full Charge and Discharge of tepper Coils Higher Pulse Frequencies do not allow Inductors to fully Charge or Discharge each Pulse Cycle. This results in a DC bias of the tepper Coils lowering motor power (performance). To increase the stepping frequency of Fig. Performance vs. Frequency a stepper motor, the applied terminal voltage must be increased. To prevent current overload, the motor current must be monitored and limited if necessary. This becomes the task of the stepper motor amplifier (Translator). To produce high-speed/high-torque stepper operation, the translator may vary the motors terminal voltage, switching waveform, coil current(s), current direction, and other factors. tepper Motor Errors One characteristic of a servo motor is that a sudden change in motor loading will not cause adverse consequences in its operation except for a definite change in motor current, and a possible change in its following error (the difference between where the control says the motor should be, and where the motor actually is). The motor will perhaps slow down or speed up, but it will not stop (assuming the load is still within the torque range of the motor). A stepper motor, on the other hand, operates best with smooth load transitions (no drastic discontinuities). In addition, the stepper motor is a zero following error device. This means that the acceleration, deceleration, slew rate (running velocity), loading, and other operating conditions, must all remain within the stepper s immediate capability, or it will stall. A sudden change in the frequency or an excessive acceleration rate can also cause the stepper motor to fail.

6 Although a stepper system offers attractive benefits such as hardware simplicity, low cost, ease of operation, and open loop capability, it comes with a price. That price is a lengthy list of considerations that can each affect its accuracy, stability, and overall operation. When you are ready to incorporate a stepper system, you must not to take any of its operating requirements and attributes for granted. The one thing a stepper does best is stall. tepper Accuracy tepper accuracy is typically guaranteed to be within to percent of one Full tep under no-load conditions. This relates to ±0.0 angular degrees for a. degree-per-step (00 step per revolution) percent unit. Check the stepper motor specification for the accuracy of the motor you are using. Due to the mechanical placement of the stator and armature magnets and the stator electromagnetic coils, stepper accuracy is not quite as perfect as one might think. However, the repeatability of a given step position in a given rotation is much tighter, generally within five arc-seconds or ±.009 angular degrees. The effect of accuracy must be taken into account when working out your system resolution. If, for example, you require system resolution to be 0.00 ±0.0 inch, and a full step is 0.00 inch, a half step would be 0.00 ± inches (% x 0.00) or a percent error. This may not seem much, but it is outside the specification. As another example, what if you were micro-stepping at steps per full step, and had to achieve a resolution of ± step ( steps-per-step = 0.00/ = inch, the repeatability being ± inch). You could possibly accomplish this move by the repeatability of the stepper, but definitely not because of its accuracy. You would have to construct a step-angle compensation table to precisely position the motor shaft, similar to a lead screw error compensation table used in CC controls. A second alternative would be to use encoder feedback. To round out the problem, stepper accuracy can change due to motor loading, two-phase-on operation, one-phase-on operation, as well as by phase coil-current matching, magnet location, temperature, and other factors explained in the following section. tepper Errors Phase errors are due to non-linearity of the phase waveforms, amplitudes, or amplifier current generated when developing the stepping power signals. Quadrature, One-Phase-On, and Two-Phases-On are terms that describe specific phase errors due to either physical coil positions, and/or current imbalance through these coils. Bias errors are due to offsetting DC current levels within a phase coil. These offset currents will produce a variance in coil torques offsetting the actual armature position from the desired position. Hysteresis errors will generally show a dead-band area in which no motion will occur despite any micro-steps being generated. This could be due to unbalance of stepper coil currents, voltage waveforms, or the mechanics. A stepper should always move one step per full or half step generated.

7 ystem Mechanics including the load will tend to increase or decrease the friction force. It will also affect the inertial loading and offsets in system balance. As these changes occur, stopping positions will be disrupted (in either direction) causing a motion dead-band. In other words, the controller will step, but there will be no resulting motion. The accumulated effect of this problem will cause major positioning errors especially in micro-step applications. Reducing tepper Errors Buy more accurate motors ( or phase) when necessary. Check the error patterns of similar motors to verify consistency. Use the optimum motor current based on the manufacturer s recommendations. Use all of the coils in their optimum configuration to derive the best torque for a given power dissipation. Generally, a motor with low detent and harmonic torque proves to be the best choice. Definitely, this is the best choice for micro-stepping system applications. If positioning is paramount to the application, use a form of position feedback such as an encoder or a scale to insure the required accuracy. Buy a stepper amplifier and motor as a matched pair in which motor rotation has been mapped, and the amplifier automatically compensates for any rotational errors. tepper Annoyances Velocity Ripple When a position step is made, the stepper system inertia will possibly oscillate (or ring) about the new step position point. This can be attributed to the fact that the motor s magnetic field(s) may not allow the two mechanical masses (stator and armature) to come to rest in a perfect trajectory when moving laterally to each other. ince there is no position feedback control loop, there will generally always be some lateral position overshoot. The larger the step size, the more torque will be required to make that step, and the larger the mechanical oscillation will be. Even when running, the stepper motors velocity profile can show this oscillation. To minimize the oscillation reduce the step size (half step or micro-step), reduce the step current required to make the move, or mechanically dampen the system. Figure shows an oscillation occurring in the physical positioning of step. It is important to note, that the physical motion of the stepper Position --> Fig. tep Point A tep Ideal R esponse Time --> tep n tepper Motion Response

8 armature has reversed as it moves to point A. The degree of movement may not necessarily be large, but if a step pulse is applied to the system during period of reversed motion (just prior to point A), the motor can actually begin moving in the reverse direction even though it was commanded to move in the forward direction. The hard part is that you ll see it, but you won t know what just happened! Electrical Ringing The phase coils that create the motion also form a tuned circuit with the internal coil capacitance in addition to any other stray or other capacitance in the circuitry. This will create ringing on the control step waveforms. Damping circuitry is required to either reduce or eliminate this condition. evere electrical ringing can cause interference with other electronic equipment including the controlling computer. This can have dangerous side effects. Mechanical Resonance in certain Frequency Ranges There can be one or more natural resonant frequencies or harmonics associated with any stepper motor. When operating at these frequencies, a heavy oscillation will be heard and/or felt from the stepper motor. This can have detrimental consequences on the motor torque and the overall system. To eliminate this problem, either refrain from running at these frequencies, use micro-stepping techniques, or do low current stepping in these regions. Possible tep Misses Without tall ince all motor and load systems exhibit inertia, it is possible for a stepper motor to fall out of, and then back into, synchronization with its controller (translator). If this happens, the position of the system will be corrupted. To prevent this, reduce the ramping and/or velocity rates. If this reduction is done in steps, then it might be possible to find the problem area. Another solution is to use a sine wave amplifier. ever allow a stepper motor to miss steps since stalling is just around the corner! Motion Definitions Full tep This term is used to denote a stepper motor operating in a detent-to-detent stepping mode. If power is removed, the stepper motor will not advance or retract from its current position (no load, horizontal attitude) since it is presently located at a detent position (see ection on stepper motor principles) any time a step is made. A 00 step Full tep motor operation will rotate. degrees-per-step pulse. Half tep This term is used to denote a stepper motor operating in a mode that allows its magnets to be placed half way between physical detent positions. If power is removed, the stepper motor will advance, or retract from its current half step position to the nearest physical detent position (if it is not presently located at one - no load, horizontal attitude). A 00 Full tep motor will produce 00 Half teps at 0.9 degrees-per-step pulse. Micro-step This term is used to denote a stepper motor operating in a mode such that its magnetic fields allow the armature to be physically positioned anywhere between actual detent positions within

9 range of the number of micro-steps in use. If power is removed, the stepper motor will advance or retract from its current position to nearest physical detent position (if it is not presently located at one - no load, horizontal attitude). A typical 00 Full tep motor can be subdivided into specified even or odd divisions up to micro-steps per full step without the need for gearing. Five phase steppers will yield 0,000 steps per revolution (0 ) and will guaranty 0,000 equally spaced steps. E Current Flow Definitions BiPolar Bipolar operation simply means that phase coil current is allowed to flow in either direction within the coil. The coils are constructed so that by sequencing the coil current, armature movement will be generated. Figure illustrates an H bridge bipolar driver. When and are closed, and are open and vice-versa. This style driver allows current to flow through the phase winding in either direction as required. Figure 7 is another example of a bipolar drive. In Figure 7 (Bi-level drive) switch is closed at the beginning of a step, allowing the higher E voltage to drive the motor. is then opened after a prescribed period of time or when a current monitor indicates that the coil current is at the required value. Opening D D then places the lower voltage E in control of motor current. R L EMF Misc Mtr Losses D D Fig. H-Bridge Bipolar driver circuit - Resistance limited D D D E E R L EMF Misc Mtr Losses D D Fig.7 H -Bridge Bipolar driver circuit - Bilevel

10 UniPolar Unipolar operation simply means that coil current in any given coil will only be allowed to flow in one direction. The coils are constructed so that by sequencing the coil current, armature movement will be generated. Figure shows an example of a Unipolar driver. E R L Misc Mtr Losses EMF R L R Translator Types The following section lists the major classifications of stepper motor drivers. However, this list does not mean to imply that these are the only types available. If you intend to use D Fig. Unipolar Resistance Limited Driver stepper motion in your application, it would be good to understand the differences between each type so that you can select the proper type for your application. For example, a chopper drive always causes the stepper motor to vibrate slightly when not commanded to move. In the film handling industry, any stepper motor using a chopper style drive that is directly coupled to a film drive roller will cause the film to vibrate as well. This would be unacceptable when positioning film for developing or splicing. A Bi-level, current limited (non-chopper) or micro-stepper drive would be a better fit. If you current limit using an external resistor, you enjoy an added feature. This approach allows the power supply voltage to be increased, resulting in an improved speed-torque curve. However, the practical limitation of the power supply size, power dissipation in the external resistor(s), and motor driver constraints must be taken into account. Basic stepper amplifier styles: D BiPolar BiPolar Resistance Limited BiPolar Bi-level BiPolar V..I BiPolar Two Quadrant Chopper BiPolar Four Quadrant Chopper UniPolar UniPolar Resistance Limited UniPolar Bi-level UniPolar V..I. UniPolar Two Quadrant Chopper UniPolar Four Quadrant Chopper Microtep,,,,,,, and steps per full step are the most popular ranges available at reasonable cost. However, odd divisions are available as well. AIC & Processor Control Methods Open Loop verses Closed Loop tepper Control Considering open loop versus closed loop control is not a matter of which is better, it is a

11 matter of which is necessary, which method requires the least amount of effort, and which method is the most cost effective! AIC operation An AIC is a pre-programmed Application pecific IC. tepper control AIC s on the market today provide an invaluable tool for generating stepper motion. I read an article not to long ago that asked, Why use another processor to generate a move, when your computer already has one? Although that company is in the business of selling motion, it sells software not hardware. My response to that question would be: One axis possibly. But have you ever really done multi-axis, processor-based, coordinated motion control? PC systems must handle the operator, the I/O, data collection, motion, intermachine or factory communications, and possibly a host of other things. PC processors are capable of only so much, and even the CC Industry uses multiprocessing. However, if the AIC fits use it. It will only aid in helping you achieve your goal. A hardware solution of comparable or lower cost to a software solution is generally the best solution. Processor Based tand-alone Operation Processor-based stepper control does not necessarily imply a PC-based processor, but rather a separate processor such as the Dallas 0, DP, etc. I have used these types of processors many times to generate, and axes of stepper control and have had excellent results. The only limitation with a software controlled processor, is the real-time stepping frequency that can be generated. Twenty thousand (0,000) steps per second is quite comfortable, but depending on the required motion, thirty thousand (0,000) steps per second is a maximum. AIC s, on the other hand, can generate up to three million (,000,000) steps per second and not even get tired. The advantage of a stand-alone device, would be the ability to multi-process. That is, to allow the system to work from one computer processor, and the motion to work from another - the AIC. A good stand-alone unit can quite satisfactorily control up to eight or more axes of stepper motion. These types of controllers will not only relieve the host processor of the motion work, but reduce the amount of software writing necessary to do the job at hand. ince stand-alone units presently incorporate software and processors on board, they also have the ability to collect data and control I/O; freeing up more of the host computer. Developing a tepper Trapezoidal Algorithm The necessary considerations for developing a trapezoidal profile are: Type of move position or velocity Accel/decel rate pulses/sec^ Max move velocity (slew rate) pulses per sec Distance to move pulses (if a position move) Time to accel/decel from initial velocity to final velocity. Desired update time to milliseconds. With the previous information the formula pattern and the software algorithm can take shape. The trapezoidal requirement is for a pulse and direction type translator. First, you need to

12 figure out the minimum pulse width that can be accepted by the stepper translator (consult the translator specification). Here, our translator requires a 0-microsecond minimum positive (true) pulse signal. ext, knowing the maximum velocity required for any of the moves, figure out the worst case time that your software can spend in the pulse-generating interrupt handler. For example, my required maximum velocity will never exceed 9,000 steps-per-second. Therefore, I ll set the worst case update time to 00 microseconds (0,000 steps/sec). Knowing that other functions besides stepper pulsing must also be handled, I will restrict the maximum time spent in the interrupt handler to 7 microseconds. That will allow a minimum of percent of the software time to be spent in the main program. We need to now find out if real-time math can be used, or if move table must be developed. My only objection to using tables is that they need to be constructed before each move, which takes valuable time (the tables need to be preprocessed just prior to or just after issuing a start command). This takes time and requires memory. The processor I decided to use is an Intel 0, running at MHz, which provides a one microsecond average instruction cycle. The equation used to figure out the required stepping interrupt time during an update is: Interrupt_Time=( Velocity )(,000,000) usec Using Logarithms, the fastest this equation can be solved is around 00 microseconds. Obviously, this will not meet the update time criteria. A move table will be required. Or will it? Examining the move requirement closer, there is nothing that suggests that the acceleration must be done in one-step jumps. If we allow another indicator to control the number of steps per second the velocity will be incremented each time an acceleration step is completed, then we can force the time interval between increments to at least. times longer than the time required to calculate it. Another interesting point is that when we reach the slew velocity, no further lengthy calculations are required until reaching the deceleration point. The deceleration position can, therefore, be calculated just prior to the move and then counted out in the pulse generating interrupt to flag when the deceleration must begin. In my operation I did not allow changing profile parameters while in motion, although it will only take your imagination to do so. The accel/decel increment that will be used to change the velocity during the respective ramps will be calculated before the move, based on the accel/decel time, the indicated update time and the maximum velocity. When we are within one velocity increment of the slew velocity, we will calculate an interrupt value that will place the desired speed right on the number. Voltage to Frequency Operation By design, no motor is capable of changing its momentum faster than the systems electrical and mechanical time constants will allow. When operating a DC motor, if the command voltage changes faster than the motors ability to keep up, extra current is pumped into the motor to increase its torque. This means that a typical motor (DC, AC, brush, brushless, etc.) can actually LAG behind the command signal indefinitely. The worst thing that will happen is

13 that the motor will be late when arriving at the commanded position. This phenomenon is known as Following Error. Following error can be position, velocity, acceleration, or other error. That is, the difference between where the control wants the motor to be, and where the motor actually is, at any time. A stepper motor, however, is a ZERO following error device. This means that when a motion command is generated, the control expects the stepper motor to move. But, the stepper motor has electrical and mechanical time constants just like any other motor. Put these two facts together (zero following error, and inherent delays), and we have a requirement for actual motion profiling, or actual motion monitoring to ensure the stepper capability is not being exceeded. In other words, the motion command signal cannot change its trajectory (or profile) path faster than the stepper motor s ability to keep up. ince the stepper is controlled by a pulse stream, the pulse stream must be extremely stable in both frequency, and rate of frequency change (accel and decel profiles). If continuity is not maintained, the stepper motor will slip and possibly stall. There are many methods of pulse generation in use today for stepper motor control. Following are some ways in which stepper motor pulses are generated, and a discussion of their benefits and/or traps. Processor generated pulse stream Microprocessors have the ability to generate pulse streams. This method requires the use of an internal timer interrupt to ensure continuity of pulse stream generation. ince this is a realtime effort, the only problem is in determining the maximum interrupt frequency (MIF) that can be tolerated by both the stepper (its maximum speed in PP) and the processor (its interrupt handling capability). The MIF sets the top frequency that can be generated which, in turn, sets the top RPM of the stepper motor. When using a processor such as an Intel 0, the top frequency is limited to around 0Khz. When using the Dallas 0, the top frequency is limited to around 0Khz using the same control software. A DP might be capable of generating pulses up to Mhz for the same control software. The limitation using processor control, therefore, is the top frequency that the processor can generate while still doing other control work (I/O sampling, operator interface, etc.). AIC generated pulse stream AIC stands for: Application pecific Integrated Circuit. This means that an AIC is designed to perform specific tasks. For stepper motor operation the AIC is a programmable frequency generator. Ultra precise ramping and slew pulses are produced from the AIC ensuring the best possible continuity for stepper motor control. Many of the stepper AIC s allow the steppers velocity, and in some cases, the acceleration and deceleration ramps to be changed on the fly. The limitation associated when using an AIC to control a stepper motor is simply ensuring that you understand all of the mathematical requirements for loading the AIC s internal registers. Many AIC manufactures supply math handling routines which take the work out of this requirement, many don t. The math ritual, however, is always given in the AIC documents. Moves such as multiple axis interpolation can then be reduced to simple software driver routines provided a complete understanding of the AIC exists. Pulse generation up to Mhz are possible from the AIC for high speed, micro-stepping applications.

14 V-to-F conversion of an Analog ignal A V-to-F converter is a device that changes an analog voltage into a variable frequency. For example, it might be desirable to convert a voltage ranging from zero to ten volts into a frequency ranging from 00Hz to 00Khz. In theory, this seems like a good way of allowing an analog servo motor controller to control the actions of a stepper motor. There are, however, several major traps lurking in the woods. First is onlinearity. It is very difficult (and expensive) to convert a purely linear voltage into a purely linear frequency without digitizing. The result of simple analog conversion could yield % non-linearity right from the start, but remember that any sudden surge in frequency could cause a stepper to slip or stall. There is also an inability to produce a quality frequency below 0% and above 90% of the V-to-F limits. econd is the problem of maintaining a stable frequency anywhere within the V-to-F s frequency range. The usual (in-)stability is in excess of %. ince a stepper motor requires continuity in frequency, duty cycle, and ramping, the V-to-F converter is not applicable nor is it practical for position control. It is, however, a qualified device for simple velocity control. A third trap is that the typical servo algorithm requires position error in order to generate the analog command signal for V-to-F use. ince the stepper is a zero position error device, the implementation of velocity, and possibly acceleration feed-forward gain must be present in the servo s gain algorithm. If it is not, the particular servo controller cannot be used without trajectory control modification. In other words, it becomes a special device. A fourth trap occurs because the servo algorithm expects positional error to occur. If the stepper motor begins to slip, the servo gain structure will try to speed up the motor in order to maintain its following error. Any attempt to speed up a slipping stepper motor will force it to prematurely stall. When trying to implement the V-to-F approach, be certain you test it to insure it can do what you want. If not, you may find yourself replacing the entire system with a properly qualified stepper or servo device that should have been used from the start. The use of a PWM servo output signal Pulse width modulation (PWM) is another form of control used in the servo motor world. What PWM does is vary the On/Off duty cycle of a pulse to produce either an effective voltage to the servo amplifier (voltage averaging), or a window within which a time can be measured and current be generated. As the requirement for a motor to speed up or slow down is invoked, the duty cycle of the PWM control signal will change. In order to use a PWM signal for stepper motor control two things are required: First, the PWM signal must meet the minimum pulse width requirement of the stepper motor amplifier. econd, the PWM signal must meet the frequency stability requirement of the stepper operation. A typical PWM signal generates a pulse stream based on a division of the clock frequency used to produce the PWM timing. The top pulse frequency is typically in the 0 to 0Khz range. (This is not very high when you consider an AIC can do Mhz).

15 In order to maintain the PWM output within the stability requirements for good stepper control, the encoder resolution (remember, the PWM signal is being generated by a servo controller that requires position feedback) must match the step resolution of the stepper motor. In other words, one pulse out to the stepper must equal one pulse of feedback from the encoder. ince continuous feedback is not possible, the PWM pulse width (duty cycle) can vary causing the stepper to stall. In order to overcome this problem, a virtual encoder is developed within the servo controller to maintain perfect synchronization of motion and feedback. It works well, but problems can occur. The problems involve tight coordination requirements between the servo controller s trajectory generator and the stepper motor s ability to respond. These requirements include the PWM frequency, the clock frequency and the required stepper pulse width. In other words, if you are doing a simple velocity move without position qualification, it could work. But, if you need to be position accurate, this approach should not be used. Conclusion In the haste to make things simpler, cheaper, faster, smoother, etc., the requirement for any control device must be qualified. All systems must be carefully analyzed to insure that the control fits the mechanical requirements, not the other way around. Although steppers are a good cost effective way to obtain motion, do not cheapen it up to the point where the system fails to perform before it has even been put onto the application!