Guidelines for motion-system design

Jacob Paso Quality Assurance Engineer

Brad Smith Regional Applications Specialist Delta Computer Systems Inc. Vancouver, Wash.

During our normal workday providing technical support, we help customers recover from a full range of problems with their motion-control systems. Looking back on the issues we deal with over and over, it is clear that if motion-system designers, engineers, and technicians kept the following guidelines in mind, projects would run smoother and on schedule, and equipment would work right the first time. Here are seven steps to a successful project.

1. Get organized.
Many motion-system engineers seem to approach system design without an organized plan. Instead, it’s best to begin by compiling a list of requirements, including overall goals for the project such as required levels of productivity and reliability.

Also define as many motion-system specifications as possible. The load, position accuracy, velocities, and accelerations dictate the selection of suitable machine components.

On the programming side, put together an objective profile of the machine’s operation, including required motion sequences for all axes. Specify the methods of communication with the controlling PLC or PC. Finally, develop a plan for verifying system performance after connecting the hardware, and again after integrating the motion system with higher-level machine controls.

2. Do the homework.
Time spent up front learning proper design and start-up methods for a given motion controller generally streamlines system design and implementation. Vendors typically offer technical training ranging from classroom sessions to live or self-paced Webbased programs. Often, time spent here is recouped at system startup.

Also review technical documentation to save time and minimize frustration. For instance, users must first know what hardware the controller supports before selecting the right motion components. Proper machine commissioning requires knowing the controller wiring and setup steps. We frequently hear from customers with feedback or control problems that could have been easily prevented had they even glanced at the startup guide. The most common problems are miswiring ground (common) pins, and not saving project data on the computer and in the controller’s nonvolatile memory.

3. Simulate first.
The first moves on new motion-control systems occasionally yield unexpected results. Moving an actuator too fast, too far, or at the wrong time can cause extensive damage, or worse, human injury. Motion controllers with built-in simulators let designers work out basic machine-function issues without connecting the controller to the actuators. A simulator will imitate any motion sequence and generate plots showing how the machine responds to a set of instructions.

Good simulators save time and help determine whether an application will actually work. For example, in a customer’s motion stage that required four synchronized cylinders, simulation safely demonstrated how all the machine’s axes work together. Another case involved a cyclic-testing rig with shortstroke, high-speed hydraulic cylinders. We first entered model parameters to approximate the system, then simulated motion and response. The resulting data helped the customer correctly size components and also permitted preliminary tuning of the PID gains, providing a headstart on machine setup.

A classic example of a motioncontrol application that benefits from simulation is a flying-cutoff saw — for instance, cutting continuously laminated I-beams in a lumber plant. In such a system, one motion axis controls a conveyor that moves the I-beam at a constant speed. Another axis controls a saw carriage moving parallel to the Ibeam. The carriage must catch up to the I-beam, match the I-beam speed while the saw cuts, then return to the home position. This requires precise position control and synchronization. Simulation saves time by letting the system integrator set up and refine motion commands beforehand, without wasting valuable product and possibly damaging the saw.

4. Use the right components.
Sometimes technicians build motion systems using hardware that just happens to be readily available, rather than hardware specifically selected for the task. It’s no surprise that results often do not match expectations in these cases. No matter how good the motion controller, it cannot compensate for poorly selected components.

For example, to ensure smooth, precise control in a hydraulic system, avoid valves with overlapped spools. Such valves cause delays when changing directions, and some overlapped spools make actuators jerk when flow begins. Good motion controllers compensate for this to a certain extent, but the motion never equals that of a system with a zero-lapped spool valve. A high-quality valve easily pays for itself in reduced setup time alone.

Controlling both position and pressure with a single actuator is an example of a task that may be impossible using “traditional” components in applications for which they were not designed. We’ve seen several instances where designers attempted positionpressure control with their tried-and-true PLC, only to find out the scan rate was too slow and programming the transition between position and pressure control was too difficult. Motion controllers designed for the task provide a deterministic update rate and easy programming. In addition, performing time-consuming calculations in the motion controller may allow for a smaller and lessexpensive PLC for general machine control.

5. Keep it simple.
A good motion controller is easy to use and offers programming flexibility. Whenever possible, system designers should use instructions designed for specific applications. If all that is needed is to move an axis from point A to B, and the motion controller has a simple “move” command, then use it.

Conversely, many profiles are beyond the capabilities of basic commands. Complex or dynamic motion, such as camming operations (where the motion of one axis relates to that of another by a nonlinear function), requires the controller’s advanced commands. If it is not obvious which commands to use, check with the controller supplier’s technical support staff.

If the controller vendor supplies “canned” software examples, take advantage of them to save time and money. Some vendors provide sample projects for common PLCs and HMIs that can form the foundation for new ones. It may be possible to simply copy and paste code to get started. Some of our customers select HMIs based on the sample projects we provide, saving them hours of programming and setup time. And some vendors provide ActiveX control software modules for communication between the motion controller and a host computer that include complete code examples.

Once the decision to use a motion controller has been made, use closed-loop control all the way — not in a piecemeal fashion. For example, an injection-molding machine builder wanted to slowly venture into servocontrol and decided to use closed-loop control with a high-quality valve to extend the plastic injector, but retract using open-loop control and a different valve. This resulted in unnecessary hydraulic plumbing, and the machine had trouble switching modes. Closed-loop control of the entire system would have been simpler, less expensive, and easier to maintain.

6. Debuging and tuning.
Modern motion controllers provide a host of diagnostic tools. Plots of actual and target motion, for example, can verify that axes move as intended. Advanced plots visually correlate motion with other machine events, such as discrete inputs, commands from the host controller, faults, and so on. Some controllers also provide detailed status and event logs. These tools can show actuator status even when motion is too fast for the eye to follow.

An important part of a motioncontrol system’s startup procedure is tuning PID gains to make actual movements track the desired motion. In the past, tuning has generally been a trial-and-error process. Now, some controllers feature automated tuning tools. They mathematically model the motion system and generate optimal PID gains for the desired machine performance.

A controller’s built-in diagnostic tools can verify that components perform as intended. One way to do this is to move the system in open-loop control and generate a motion versus time plot. The motion should be smooth and easy. If not, the system may exhibit erratic motion during closed-loop control as well. Mechanical friction, undersized actuators, or poor valves typically cause jerkiness or oscillations.

After verifying open-loop motion, move the system in closedloop control with the same speeds, accelerations, and loads that will be used during machine operation. Again, use motion plots to verify the system meets specified requirements. Also check that the system moves as fast as intended. This verifies motor sizing in the case of electromechanical systems; or pump, cylinder, and accumulator sizing in the case of fluid-power systems.

7. Ask for help.
Designers can readily tap outside design talent if they need additional expertise. System integrators, component distributors, and motion-controller manufacturers are valuable experts to rely on. Used wisely, they can make a project successful, and save time and expense in the long run.