Richard W. Armstrong, Jr.
Kollmorgen Motion Technologies Group
Motion-control system designers agree that motors must match loads both mechanically and electrically. And motor manufacturers offer guidelines and software packages to help engineers determine numerous parameters such as load friction and inertia, motor acceleration, RMS torques, speeds, and voltage and current requirements that plug into design equations. But sometimes designers must estimate or empirically find unspecified constraints, and the systems fall short of their goals. These constraints may include amplifier peak-current time constants, demands of multiple axes on the power-supply bus voltage, shunt regeneration, and unique load qualities.
Engineers frequently use time-consuming, manual calculations tempered with experience to characterize motion- control systems, convert units of measure, and accommodate diverse mechanisms and motion-profile simulations. But recent software developments help engineers account for all variables, including many not considered by older software packages. The newer crop helps solve motion-control equations more accurately, efficiently, and quickly.
The advanced software considers more than just basic parameters. It uses a systems approach to calculate power supply and shunt regeneration needs for multiple axes. Furthermore, a separate algorithm simulates the amplifier’s foldback circuitry to ensure that the combination of motor and drive electronics matches the duty cycle. Often, even when RMS and peak motor torque satisfy the requirements, the amplifier can lose control and go into current foldback under some duty cycles.
But even more important, the software factors in a wide variety of standard load mechanisms and their related motion profiles. Typical mechanisms include leadscrews, racks and pinions, conveyors, nip rolls, rotary devices, linear drives, and spindles. The motion profiles that go with them include triangle, trapezoidal, and custom-built profiles allowing almost unlimited shapes.
The software handles system designs two ways. First, mechanism load values are custom fit to well-defined motion profiles. The method calculates peak and RMS torques, while considering loads, acceleration factors, and rotor inertia. When motion profiles are not defined, the database searches through mechanism load values and assigns one assuming a 100% duty cycle.
The software contains motion-control system design procedures for a vast servo, brushless servo, brushless direct- drive linear, frameless rotary, and spindle motors. A convenient tool in the software groups together multiple axes of motion under a single project name. This approach lets designers calculate separate power supply bus power and shunt regeneration values for the combination.
Designers first create a project and select a mechanism for various axes. When a standard mechanism does not satisfy the motion requirements, they can enter new mechanism data directly by specifying inertia, torque, and speed requirements. Moreover, the number and combination of axes defined in one project has no limit.
Project options include renaming, deleting, and saving projects as another name. The Save As pick is a handy tool for machine designers trying out various designs that contain only minor differences. For example, they can save one basic design and replace only a few variables under Save As for different power supply and regeneration calculations among several motion axes.
The software also exports or imports projects. Export assigns a file name to a specific project and saves project databases individually. Import retrieves previously saved Export files from another computer or storage device. This lets multiple users such as design engineers and application engineers share and examine project data.
Examining the menus and functions in one package called Motioneering from Kollmorgen show the capability in the new technology.
Mechanisms: Users choose from several different mechanisms representing the loads typical for servomotors such as leadscrews, racks and pinions, conveyors, nip rolls, and direct-coupled rotary devices. Unique algorithms evaluate linear direct drives and spindles with pop-up forms, for example, to show various input parameters entered for a rack and pinion.
All fields need not hold data. An X appears in the field button when exiting a field without filling in the necessary data. On the other hand, when all critical data are entered, a check mark appears, and a summary of input and output results displays. This allows a quick check of the basic requirements and serves as a prompt when something is overlooked. For example, it is obvious when an alternative reduction ratio better matches typical motor speed and torque capabilities.
Reports specific to mechanism data, motion profiles, and motor/drive selections may be printed. Other options in the Mechanism screen include Size Motor and Motion Profile. Assuming users have entered the speed requirements and maximum and thrust loads, they may search the database without specifying a motion profile. This might be the case in applications having 100% duty requirements and no critical accelerations. Otherwise, users can define a motion profile specific to that axis.
Motion profiles: These simulate mechanism speed changes for calculating more accurate values of peak and RMS (thermal equivalent) torques. Peak torques are a function of acceleration torques (the product of inertia and acceleration rate) plus load torques such as: T = Ja + Ft, where T = torque, J = inertia, a = acceleration, and Ft = friction torque. The motion-profile calculations proceed segment by segment.
RMS torque comes from
Simplified manual calculations often minimize designers’ efforts, but the method may overestimate motor requirements. One often-used calculation sums rotor inertia and load inertia divided by reducer efficiency to find acceleration and deceleration torques. This equation finds the correct torque only in the first and third operating quadrants. With a relatively low mechanism efficiency, however, calculating in all quadrants with this equation produces an excessive RMS torque. When accelerating a load, the torque calculation must take into account the speed reducer efficiency, thereby requiring a greater acceleration torque while reducing the deceleration torque. Equations that account for efficiency are listed in the table, Acceleration Torques by Quadrant.
The motion-profile parameters entered depend on the type of profile selected. For example, triangle and trapezoidal moves accept distance as a single quantity. In contrast, custom moves contain data on each motion segment. To evaluate the effects of gravity, friction, and acceleration in each quadrant, the program redefines the data by motion segment as needed.
Sizing methods: Peak and RMS torque calculations are based on motor and load parameters. The software calculates the speed-torque load points for a specific motor considering maximum speed, RMS torque at the average speed, and peak torque at maximum speed. The software calculates a safety margin in percent positive or negative, depending on whether the motor capacity is greater than or less than the respective load requirement.
Peak-torque safety-margin calculations contain more detail because no particular segment of the profile assumes the most peak torque. Rather, the software calculates peak torquespeed points for all motion segments and individually compares them with the maximum envelope. The software indicates the segment with the least margin of safety. All torque calculations show for individual motion segments as part of the Reports output.
Amplifier foldback: Before system performance curves were available, motor performance was defined empirically, from optimum current and voltage. In some cases, motor performance was estimated, and often differed significantly from the final physical system. But now, more reliable, empirical system performance data are available to size motors. The data includes motor and drive limitations such as voltage and current limits, and the current’s form factor.
The software also calculates one new additional factor — the duration the amplifier provides peak current. Peak current limits of only a few seconds are deliberately designed into transistorized servoamplifiers to protect them from overheating and ensure long-term reliability. Various algorithms such as the product of current and time and I2t functions provide the servoamplifier protection. An algorithm that satisfies Kollmorgen’s amplifiers is included in this software’s database. It factors in amplifier time constants and RMS continuous current ratings.
The amplifier foldback test operates over the entire cycle, one or more times as needed. When the software simulates a foldback condition, a warning appears telling the user that the system does not meet specifications. This is in contrast to green highlighted systems when all conditions are satisfied.
To minimize processing time, the test runs under two conditions. First, the system must meet or exceed performance minimums. And then the time of the motion profile must exceed the longest possible cycle time for a foldback condition.
Supply bus sizing: Designers usually oversize power supplies because the procedure to determine precise power requirements considers motion profiles that normally take long to calculate. Engineers often generously estimate the value to get through the process quicker. The software, however, calculates power requirements on a per-axis basis. It lets designers combine data for multiple axes of motion to size the power supply and determine shunt regeneration values.
This procedure helps product designers where power supplies drive multiple servoamplifiers. In general, a power supply for multiple systems must satisfy the bus power, logic power, and shunt regeneration. It takes only one variable to overload an undersized power supply.
The software uses several methods to determine the final power requirement.
• Continuous Mechanism Load Power is a function of the sum of all continuous loads multiplied by the speed at thrust load. The sum displays as Bus Watts under the Power Requirements heading.
The cycle time divides into the sum of the positive energy in joules to calculate average watts with the equation
The bus capacitance and the nominal and maximum bus-voltage limits determine how much negative regenerative energy can be stored for later use. This potential energy is disregarded in bus-power calculations to provide a margin of safety. When a specific mechanism does not have a defining motion profile, the calculation for power defaults to a continuous mechanism load method.
• One option estimates the percent of power in each axis for a typical machine-tool duty cycle. Percentages are generally conservative for simultaneously operating axes and should serve only as a starting point.
• Another method assumes maximum continuous power for all systems operating simultaneously. This provides the maximum requirement for the system’s design limits.
• A custom option lets users define specific percentages for each highlighted system. Selecting various product combinations lets designers optimize load configurations for a particular power supply rating.
Shunt regeneration: This is another calculation that designers often overlook or conservatively estimate instead of taking the time to go through the detail needed for a complete analysis. The calculation becomes more complex when the system involves multiple axes. The software provides algorithms to calculate resistor values and their power ratings in watts for single or multiple axes using either peak decelerations or decelerations of motion profiles. For peak deceleration calculations, the energy in joules regenerated into the bus divided by the time between regeneration cycles determines the average power, where cycle time is a function of deceleration time and percent duty cycle. The dynamic energy of the system during deceleration minus the losses during regeneration finds the energy regenerated into the bus. The losses consist of I2Rm motor losses, and power loss from load friction. Drive losses are not significant. For example, the calculation for energy in joules for a brushless servomotor is
The energy in joules divided by the time between cycles for the chosen duty cycle solves for the average power:
A resistor dissipates the power generated and is determined by the peak instantaneous current. The software finds the resistor value by dividing the maximum dc bus voltage by the equivalent peak dc bus current as
R = Vmax/Im.
The same power and resistance calculations can be applied to a specific motion profile rather than using the peak deceleration method. The solution requires two calculations. One uses a single, maximum regeneration cycle expressed in joules to solve the resistance value. The other sums the regenerative energy for the entire motion cycle, likewise expressed in joules. The power in watts calculates from the total regenerative energy divided by the total cycle time. When more than one mechanism is in a system, the maximum regenerative energy is the sum total energy of all chosen mechanisms.
Units selection: Click on the Units caption in the menu bar of the opening screen to pull up the Set Default Units screen. Then select the units of measure preferred for the project. Individual parameters appear in both English and metric units, but any combination of units may be saved. In the illustration, Default Units Screen, available linear acceleration units appear. After saving the desired default units, data-entry fields automatically show the preferred units. Frequently, some component or input parameter will need to be added in an alternative measurement unit. The user chooses among available units for data entry through a drop-down combination box. The Screw Parameters illustration shows a typical leadscrew data-entry screen with this option used. The specific application saves data in the units entered.
Another convenient aspect of Units is the built-in converter. When the user wants to convert a parameter value into another unit, pressing the shift key and clicking the mouse simultaneously pulls up a menu of conversions. To save data in a different unit, the user selects the preferred unit and the data (unit and value) transfers.
Unit conversion also applies to certain output functions. Results and printed reports appear in the previously selected default units. Users can immediately change units in the display of the speed/torque curve. The speedtorque figure shows a typical curve with options for selecting alternative speed and torque units. The software automatically rescales and replots the curve. Also, users can test a particular speed to find the available continuous and peak torque values.