Spiralling energy costs, increasing productivity demands, and quality requirements that keep getting more unforgiving. If even one of these trends hits home, you’ve probably already decided your next design project will use servocontrol.
Take heart. Others have made this step before you; some successfully, some with disastrous results. In servosystems, the difference between success and failure is preparation, and that’s what this article is about.
Before you begin designing a servosystem, you need a clear picture of what you want to accomplish. Write it down: who, what, why, where, when, and how much. Gather all the application information you can because it will point you to the combination of servomotor, drive, and controller that will meet everyone’s expectations.
There are a number of things to consider when selecting the best servosystem for a given application. The best components for one application may be a poor choice for the next. In some high-end applications, the choice of servosystem components is driven almost exclusively by the ability to deliver the best motion characteristics. These include:
• Fast settling time
• High bandwidth
• Shortest time to achieve an index
• Ability to perform special motion (contouring, s-curves, etc.)
In other applications, the required motion performance can be achieved by one of many servosystem solutions, and the final selection is often determined by other considerations, such as:
• Initial cost constraints
• Focus on final installed cost
• Reliability and ease of maintenance
• Power levels involved
• Overall control system strategy (stand-alone controller, highly integrated multiaxis, etc.)
• Communication preferences (digital I/O, serial, CAN, MACRO, etc.)
• Ease of programming or start-up
• Mechanical form (panel, mounting, Eurocard)
• Control cabinet volume required
• Minimization of field writing required
Choosing a motor
When it comes to choosing a servomotor, one concern is inertia. If motor inertia is too small (less than 10% of the reflected load inertia), the control loop may become unstable. If it’s too large (greater than 10X the reflected inertia), the motor will require excessive current to accelerate and decelerate.
The prime mover can be anything from an induction motor to a stepper or servo. Induction motors are the least expensive, but also the least powerful for their size–and they don’t provide for a stiff system. Although induction motors can run at peak capacity for long periods of time, they usually have too much inertia to accelerate fast enough for the typical servo application.
Stepper motors, on the other hand, can accelerate quickly (with the right amplifier/controller) and they maintain a stable velocity. They’re also relatively inexpensive and smaller than induction motors, offering more control. On the down side, they have limited holding torque and too little inertia to precisely control large inertia loads. Accuracy varies depending on the coarseness of each step.
Servomotors, size for size, provide the highest torque and very fast acceleration (because of their low inertia). They are also the most accurate and repeatable of the lot, smoother than steppers. Performance isn’t cheap, however. Servomotors are the most expensive, mainly because of their amplifiers.
High-torque, low-inertia motors are designed for response. With less rotor inertia, more power goes to the load. In addition, because of low winding impedance, the winding current delay time is shorter and there’s less winding power loss.
Because of sensorless flux technology, it is also possible to use ac induction motors in some less demanding motion applications. The same technology is also used on brushless permanent magnet motors. Their lower inertia allows higher acceleration than ac induction motors without needing the feedback wiring of a traditional brushless servomotor solution.
Rotary brushless permanent magnet motors have evolved in several ways. The have more powerful and uniform magnets, more reliable encoder construction to better withstand heat and vibration, and reduced manufacturing costs.
Today’s motor options now include not only rotary choices, but also linear choices. Linear motors are proving useful in many applications because they replace some of the complexity and cost of rotary systems with a direct coupled linear solution. They also allow much stiffer and precise motion (no backlash) and enable faster motions because of low inertia.
Today’s automation systems typically have several linear motion axes. Belts and pulleys provide fairly inexpensive linear motion when coupled with rotary motors, but they can be difficult to control when driving large masses (belts are elastic). Belt drives also lose accuracy over time as belts stretch and pulleys wear.
Ballscrews, on the other hand, provide stable motion and are easy to control. However, they have a limited velocity, especially when longer motion is required. Backlash is also a concern and tends to change as the system ages. Ballscrews are fairly inexpensive if high accuracy isn’t required, but low-backlash ballscrews can be costly.
The latest feed-axis drives are, of course, linear motors, the most common types being iron core and ironless. Iron core motors provide more force per volume, but the amount of force varies as the coil passes the magnets. This type is also somewhat slowed by heavy guidance systems added to withstand the pre-load imposed by the attraction between the iron core and magnets.
Ironless motors, though not as powerful, are capable of much higher accelerations and velocities. That’s because they don’t carry as much mass; their guidance systems needn’t be as heavy because there is no pre-load from the magnetic attraction between the coil and magnet track.
At present, linear motors are limited by the physical constraints of their guidance systems. Although they are typically the most expensive solution, they are cost-competitive with precision ballscrews in shorter lengths, and untouched by belts and ballscrews when a job calls for accuracy and speed over long lengths.
There are many types of motor drives available today, ranging from fully capable smart drives to very basic current amplifiers. Criteria for selecting a drive include:
• Initial cost • Power requirements of the application
• Need to match the motor feedback type (resolver, encoder, etc.)
• Ability to do motion control
• Multiaxis configuration for lower cost/axis
• Fast loop update times for highend applications
• Ability to match the preferred communications protocol (MACRO, CAN, SERCOS, etc.)
• Ease of system start-up and tuning
One alternative is a basic, inexpensive PWM drive (essentially a current generator with some fault protection) that relies on the controller to provide pulse-width modulation. This type of drive is becoming quite common as high-end controllers take on more PLC and motion control functions.
Drive power electronics have been trending toward higher PWM frequencies, designs that facilitate sinusoidal commutation, and reduction of ripple on the output current. One thing to bear in mind is that all PWM amplifiers produce force variations as power switches turn on and off. The variations are usually minor, but they do exist. Most applications, however, don’t require extremely smooth driving force.
An analog drive used in conjunction with motion controllers is yet another option, one that is relatively inexpensive and offers easy set-up. Analog drives offer the highest system gain because they form an internal analog control loop, but they need to be set up individually and normally use trapezoidal commutation. For high-speed indexing applications, analog drives are a viable alternative as long as the motion doesn’t have to be extremely smooth. Analog amplifiers continue to be popular because of their generally lower cost, ease of set-up, and the familiarity that exists with many users.
Hybrid digital drives, a variation of analog drives, allow users to download set-up parameters but still maintain the high bandwidth of an analog controller. Some offer basic indexing and ratioing, which is sufficient for most applications. One drawback, however, is that most of these drives don’t have the input frequency necessary to maintain good control at high speeds.
Totally digital drives, the latest in servomotor control, are also widely available. Most offer fairly complex control capabilities as well as limited PLC capacity. Some approach the bandwidth of analog drives. Besides controlling commutation, digital drives provide basic motion control functions and limited I/O. Although they are the most expensive drives, they can actually reduce system cost by eliminating the controller.
Digital amplifiers have a number of advantages, one being that they are able to be configured to an application and duplicated on successive installations. This makes digital drives adaptable to various motor types, including linear motors, and they allow sinusoidal commutation for smooth movement. In addition, many digital amplifiers have some motion control “smarts.” In some applications, they even eliminate the need for a separate controller and the associated wiring and cabinet space, bringing down costs.
More specialized applications may require specialized drive-controller combinations. For extremely smooth, consistent motion, for example, many designers turn to linear amplifiers. Linear amplifiers are generally used in low-power applications where extremely low noise emissions and smoothness are critical. Here, there must be an amplifier for each phase or pole of the motor. Usually the drive manufacturer balances each amplifier separately to ensure each motor delivers smooth, uniform force.
Sensorless drives, on the other hand, use the back emf of the motor for commutation timing and, in some cases, positioning. These drives can control the basic velocity of nearly any motor. Higher-end versions will also perform simple indexing. They are not well suited for very low-speed applications, however.
Choosing a controller
The selection of a controller depends on the motion requirements and the motor-drive combination. Specific factors that drive the selection of a motion controller often include:
• Number of axes to be controlled
• Flexibility required, i.e. ease of changing motion parameters frequently
• Integration methodology (PC, PLC, network, standalone, etc.)
• Type of motion required (contouring, indexing, etc.)
• Ease of programming and level of software
Controllers come in three basic types: stand-alone, PC-based, and controller-drive combinations. Some are basic indexing controllers, while others are highly complex devices capable of contouring and other sophisticated motions.
As far as space limitations are concerned, smart drives take up the least amount of real estate since no stand-alone motion controller is required. Often times a PLC will coordinate the machine function and instruct the smart drive to begin its motion, and actual motion generation and control reside in the drive. PC-based controllers normally take up the most real estate. However, some of the smart human-machine interface (HMI) units incorporate a PC, and this will sometimes free up the panel space required by standalone units.
Controllers range from one to any number of axes. Most PC controllers are eight-axis, with the capability to add cards and axes as long as you have the required slots in your PC. Because of this limitation, some controllers are connected to a bus communication structure of some type. This can be Profibus, CANbus, Ethernet, RS-•85, MACRO, CAN, DeviceNet, or Profibus. The required speed of the field bus depends on the amount and type of data that must be transmitted.
A controller may also be implemented in the amplifier. This is attractive because it saves panel space. A centralized controller is usually preferred if two or more axes need to be coordinated, as is the case when creating contours. If each axis operates independently, even if the complete system includes many axes, then a distributed system may be preferable. If limited coordination is needed, it can be accomplished via I/O.
If multiple control channels or very complex motions are required, stand-alone and PC-based controllers are normally used. These provide multiple camming options, ratioing many axes to a main reference, as well as circular and linear interpolation and direct CNC control. Most PC-based controllers are capable of running in a stand-alone application. Another common solution is to have PC-based or even PLC-based motion cards controlling drives with analog commands. A trend is developing toward the use of digital commands on networks for noise immunity and lower-cost installation (such as MACRO).
Most controllers also have a menu-driven programming capability, or they may be programmed in a high-level language such as Visual Basic or C++. The menu-driven is by far the easiest to implement, but it is limited in functionality. Being able to program the controller in C++ (or Basic) offers greater flexibility. However, with this flexibility comes the need for having software programmers on staff and the associated costs.
Input/output requirements may also dictate the type of controller used. Most smart drives have very limited I/O capacity. They normally are not very versatile as far as mixing I/O voltage and current. This usually creates a requirement for external relay packages or a full PLC rack to meet the I/O needs.
Standalone and PC-based controllers are usually very versatile as far as I/O capacity is concerned. Both have provisions to add I/O banks to accommodate the highest I/O requirements. They are usually able to mix voltages with different I/O boards and can directly control external functions with the use of relays for most situations.
Dierk Beneckie is product manager for the MPFLX Drives in Horsham, Penn. Pat Berkner is senior application engineer manager, Paul Kiner and Allen Berg are application engineers, and Bob Shoultz is motor design engineer, all with MTS Automation, New Ulm, Minn.