Smooth driving with linear motion

Design issues can be the source of headaches, especially when you're implementing linear motion. The technology can bring accuracy and precision to many applications, but selecting components requires a bit of forethought. Our panel of industry experts discusses some potholes to steer clear of.

What are some design problems you have seen that should be avoided when implementing linear motion?

Broc Grell, Nexen Group: Designers forget to add the mass of the motor, gearbox and linear system to the mass that needs to be moved by the linear system. Designers forget the cables will pull on the system, as well. When going into an existing system as a retrofit, make sure there is room for the new system and that it can be attached properly to the machine. Designers don’t understand how all the components—motor, gearbox—in the system add up to affect aspects such as friction, inertia mismatch and total required force to move the system.

When using a software calculator system for motor sizing, it is the same old saying with a hand calculator, where junk in equals junk out. Making sure the inertias are correct and being located correctly through the different ratios in the system and making sure drag is understood are very important. A designer should always look at the motor that the system recommends and make sure that it makes sense with the load and ratios in the system before just running with that motor. Sometimes mistakes are made in the calculator that can cost a lot of money if the motor is extremely oversized or undersized because of a mistype in one of these calculators.

Broc Grell is applications engineer at Nexen Group.

Matt Prellwitz, Beckhoff Automation: Before you start the implementation, you need to be sure of the exact linear-motion requirements in your application—distance, load, mass, inertia—in addition to the limitations of your components and controller. This is a major reason why we suggest a PC-based control system with the motion handled in software, because it provides the ability to more easily scale your controller as the needs of your application grow and change. In addition, choosing robust components, with the help of motion-configuration software, will go a long way to avoid future mechanical failures or underwhelming performance.

Matt Prellwitz is motion product specialist at Beckhoff Automation.

Gary Rosengren, Tolomatic: As it pertains to linear actuators, not considering the effects of resulting moments, or torques, the position and size of the load on the cylinder determines the resulting bending moments applied to the cylinder itself. Even if a load is located on and directly over the center of the load-carrying device, it will still be subjected to bending moments on acceleration. It is important to determine if the cylinder is capable of handling the resulting moments. For off-center or side loads, determine the distance from the center of mass of the load being carried to the center of the cylinder’s load-carrying device and calculate the resulting bending moment (Figure 1).

Moments

Figure 1: Mx Moment (roll) is created by loads applied at a distance from the x axis. My Moment (pitch) is created by loads applied at a distance from the y axis. Mz Moment (yaw) is created by loads applied at a distance from the z axis.

(Source: Tolomatic)

Don’t overlook the effects of dynamic moment loading. Unlike rod-style actuators, many rodless actuators must support the load during acceleration and deceleration at each end of stroke. When there are side or overhung loads, the dynamic moments must be calculated to determine which rodless actuator is best equipped to handle the resulting forces.

Remember to account for forces which occur during motion—dynamic forces—particularity those which occur at the point of acceleration or deceleration. These forces acting upon the bearing system may overcome the capability of the bearing system and shorten life.

When an actuator is mounted vertically in an application, additional force and load air considerations must be addressed. An actuator mounted vertically needs to overcome the force of gravity first before it can accelerate a load upward. A vertically mounted cylinder will need to produce more force than a horizontally oriented cylinder to achieve this.

Oversizing actuators is a bad habit left over from fluid power applications where oversizing was considered inexpensive insurance against not having enough power. With fluid power cylinders, the additional cost of a slightly larger actuator than necessary was minor compared to the extra engineering time that might be involved in sizing it correctly. It was common for engineers to build in a 2:1 safety factor on fluid-power applications for a variety of reasons. These included erring on the conservative side to compensate for imprecise knowledge of the loads, fluctuations in available air pressure and oversizing in anticipation of higher loads in the future due to production growth or application changes. Electric actuators can cost significantly more up front, so over-sizing is a more costly mistake.

The environment in which the actuator will be operating can have a profound effect on performance, durability and maintenance. High temperatures can affect seals, lubrication, bearings and motor life. Extremely low temperatures can also affect performance, lubrication and wear. Contamination with oil, water or abrasive grit can destroy seals unless the actuator has an appropriate IP rating. Since IP ratings only address static conditions, dynamic conditions such as vibration, heat, cold or movement also have to be considered.

Rod-style actuators, characterized by the piston rod or actuator rod extending and retracting with each cycle, typically offer numerous mounting options. Mounting options such as drilled and tapped holes in the device, mounting feet, spherical rod joints, alignment couplers, clevises or trunnions are commonly offered by most suppliers of rod-style actuators. When employed with a guided mechanism, care must be exercised to assure each subsystem, actuator and guide assembly is capable of unimpeded, smooth motion. Systems that attempt to rigidly couple the drive element to the driven element may exhibit inconsistent performance as these two elements try to move in separate planes with one or both of the subsystems loaded beyond its capability.

A rod-style actuator in such a system is best employed with some compliance member between the drive member—actuator—and the driven—guide system. For example, a spherical rod end mounted to the actuator rod allows the mounting point to swivel about the spherical joint. This type of connection at the guide is best used in conjunction with a trunnion or clevis at the opposite end of the actuator where it attaches to the machinery frame element. Such a mounting scheme allows compliance in the connection without adding undue stress to either the drive—actuator—or the driven—guide system.

Rodless style actuators, characterized by their strokes being contained within their overall lengths, may also contain a guide system built into the actuator. Rodless actuators, when used in conjunction with a separate guide system, as I mentioned, will also need to include a compliant member in the connection between the drive and driven members. Most actuator suppliers offer a variety of mounts intended for this type of installation, such as floating brackets.

Rodless actuators that include a guide system can perform the task of guiding and supporting the equipment while taking the place of a separate guide system. This feature can be particularly useful and many times saves the machinery builder time and money in the process. Rodless actuators with integral guides can be built into the machinery in combinations to meet a wide variety of motion needs. Multi-axis configurations such as x-y or x-y-z along with gantry configurations are all possible with proper sizing. In the installation of rodless actuators with integral guides, alignment is equally important.

When mounting an actuator or a series of actuators to a structure you must consider parallelism and perpendicularity. Parallel misalignment can apply an unfavorable Mx-axis bending moment on the bearing system. Carriages of rodless actuators must be mounted at the same height. They also need to be mounted at a consistent distance apart from each other from one end to the other to prevent an unfavorable Fy-axis side load on the bearing system which can cause binding. In addition, they need to be mounted level to each other to prevent an unfavorable bending moment in the My-axis on the bearing system.

Perpendicular misalignment in an x-y-z system in the x plane applies an unfavorable y-axis bending moment on the actuator’s bearing system. In a gantry system where two actuators are in the x-axis or y-axis, they need to move simultaneously. Misalignment or inadequate servo performance will apply an undesirable bending moment in the Mz-axis to the bearing system.

Actual tolerances related to alignment recommendations and mounting vary from actuator manufacturer to actuator manufacturer, as well as from bearing type to bearing type. However, a general rule of thumb is to consider the bearing system type. High-performance bearing types such as profile rail systems tend to be quite rigid, and alignment is more critical. Medium-performance systems using rollers or wheels often have clearance which offers some forgiveness in alignment. Plain bearing or sliding systems often have greater clearance and may be even more forgiving. When installing linear actuator mounting systems, there are a number of measurement tools ranging from gauges to laser systems. Whatever tools are used, always create one axis as a reference for the x-y and z planes and mount the other devices with respect to the reference axis. Doing so will help to get the maximum performance and longest life from your actuator system.

Gary Rosengren is director of engineering at Tolomatic.

Brian Zlotorzycki, Heidenhain: Make sure the motor being used is properly sized to avoid any overheating due to the motor working close to its capacity. To help alleviate any unwanted heat generation, users can dissipate the heat through an external cooling method, such as a heat sink, or increase its size. Also, the tolerances of the air gap need to be maintained. The standard gap is typically 0.9 mm, and, if it increases, then the performance of the motor suffers.

Brian Zlotorzycki is Etel motors product specialist at Heidenhain.

Clint Hayes, Bosch Rexroth: Not accounting for full loading conditions—mass load, tooling or process forces—results in undersized components for the application. Not achieving the required degree of precision results typically from a poor understanding of the difference between travel accuracy, repeatability and positional accuracy. Oversizing of bearings—build to perception—can result in missing the required price points for the market. And there’s not accounting for environmental impact on the life of the bearings—humidity, washdown requirements and contamination.

Rexroth uses the acronym LOSTPED as a way to remember all critical factors in designing linear motion systems: load, orientation, speed, travel, precision, environment, and duty cycle.

Clint Hayes is product sales manager, linear motion technology at Bosch Rexroth.

Derrick Stacey, B&R Industrial Automation: Alignment of the load bearing elements is crucial. Most linear motors, including all electrical, as well as ball screw and belt type, are designed to provide higher thrust force than the load they are physically able to carry, so they are often coupled with load-bearing elements. This can be recirculated ball guides or roller bearings among others. Ensuring that you have enough overhead to handle the load is important, but in the end we need to be able to smoothly and controllably move from A to B.

Designers need to take into account that the higher the precision and load capabilities of a linear bearing usually mean a loss of flexibility and misalignment tolerance. This lack of flexibility often leads to control/tuning issues that can yield delays or performance degradation of the overall system. By keeping this in mind from the start, these downstream issues can be minimized or even avoided.

Derrick Stacey is solutions engineer at B&R Industrial Automation.

Chris Bullock, Bishop-Wisecarver Group: Not including proper lubrication systems is, by far, the most common. The next is trying to design a machine without having adequately determined the system requirements. Never purchase until you figure out exactly what needs to be done by the machine and how.

Chris Bullock is applications engineer I at Bishop-Wisecarver Group.

Josh Teslow, Curtiss-Wright: One age-old issue is machine-tool programmers guessing the user units during initial programming and crashing servo mechanisms the first time they push the start button. It is so important to double-check all parameters/user units before test-running expensive equipment.

There tends to be a strong focus on the mechanical portion when sizing a system and not enough focus on the controls. The controls, especially the programming, are just as important as the mechanics. The mechanical portion will only do what it’s told.

Different types of misalignment can occur during installation or operation. The load must be properly guided, whether that’s with the actuator or some kind of external guide. It must be properly guided when long life or high force is required.

Josh Teslow is applications engineer at Curtiss-Wright.

Jay Johnson, Sick: Linear motion is a wide scope, so I’ll focus this response to a problem that may arise with the addition of a linear feedback device to a rotary servo motor driven axis.

Ball screw or belt driven linear axes typically have compliance and/or backlash, and it increases with use as the mechanical components wear. In addition, there will be some degree of non-linearity in the manufacturing and assembly of the drive components. Therefore, in applications requiring high positioning accuracy and repeatability, it may be necessary to add a linear encoder to close the position loop while the motor mounted encoder is used for speed control and commutation. With this addition, the system is now positioning to the linear encoder, eliminating the variability of the ball screw grind accuracy or linear bearing alignment.

In this scenario a new problem can be introduced depending on the control scheme and how much compliance is in the mechanical system. Consider that when an axis is enabled and holding position, there may be external forces such as a cutting tool or gravity in a z-oriented axis. The axis will move slightly from the pressure, and so will the linear encoder. The encoder signals this movement, and the controller sends a corrective signal to the drive amplifier. Yes, this is normal PID loop behavior, but the problem comes if the lost motion is large enough that the corrective move doesn’t complete quickly. For instance, if there is .003 inch of backlash in a ball screw and the axis is loaded to one side, the position feedback will be delayed while the slack is being taken up. With typical gain settings, the control will quickly react with a stronger corrective signal because the feedback isn’t changing. Now the axis is traveling faster and it moves slightly past the target position. The control immediately reverses the signal to correct, but again the lost motion must be taken up. A hunting effect can begin resulting in an axis continuously reversing in an attempt to close the position loop. This problem is magnified in systems with high stiction such as with large machine tools or heavy moveable structures.

One solution is to lower the system gain so that the controller doesn’t respond too harshly, but this may not be practical for the application as it softens the overall performance. Another option is to increase the feedback resolution, but this will only help but not eliminate the problem. Some motion controllers are able to effectively deal with this issue, especially in the CNC industry. Methods include different gain and acc/dec settings for commanded moves versus holding position, or the use of mechanical clamps during idle/holding operations. Obviously, the best synopsis is to reduce the backlash to a minimum if possible.

Jay Johnson is national product manager at Sick.