When using servos or steppers, what applications present the most challenges in terms of machine productivity and why?

SCOTT/XTREME: If there are challenges to productivity, I would have to say they occur when there is little or limited knowledge about applications and the motor requirements associated with them.
BOB/IMS: Machine productivity equates to product throughput. In high-speed applications, milliseconds count. This quickness challenges the internal program execution time, which is a function of having a succinct, well-written program. In addition, we have I/O response and recognition times, motor move times, and motor settling time.
CHRIS/KOLLMORGEN:
Applications that create special challenges are those requiring high speed, high throughput, and high accuracy. Some specific examples include electronics assembly (pick and place and PCB drilling), semiconductor manufacturing (die bonding, chemical mechanical planarization, metrology, and robotic arm movement), high-speed packaging applications (cartoning, case packing, and form, fill and seal machine control), and converting applications (registration printing and web/film control).
DARYL/AGILE:
Converting from a centralized control solution to a distributed servo solution requires the user to understand and design his system with additional performance parameters. Adding a network introduces latency into the system.

There are methods available to the user in writing his host application to minimize the impact of the network latency. Understanding not only the network latency, but also where the motion calculations occur and how to avoid inserting these at the start of every motion become critical in the system design to optimize performance.

What are the worst cases of improper design and implementation you’ve seen? Describe what went wrong and how these and similar problems can be prevented.

DARYL/AGILE: Tricky applications typically involve switching from a centralized solution with cabinet-mounted amplifiers and plenty of wires distributed throughout the machine to a distributed solution that involves re-engineering the mechanical layout of the machine and robot.

As both the electrical and mechanical systems are greatly affected by this change, it is necessary for both engineering groups to work closer together on the distributed servo solution.

Full implementation of a distributed servo system also includes I/O and how this can improve the system performance. I/O sometimes is an afterthought in the software design, but distributing the I/O controls to reduce system delays can greatly impact the system performance. CHRIS/KOLLMORGEN: Machines with very compliant structures frequently cause control issues. Too often, high-performance motion components are coupled to the load with very compliant couplings, ordinary gearboxes or even belt arrangements. These effectively defeat the benefits of highperformance motion components and create difficult-to-resolve control issues. If a system cannot be constructed as a direct drive (rotary or linear), which dramatically reduces or eliminates the compliance issues, then careful design using stiffer structures and often highercost couplings or gearbox types are needed.

Failure to use manufacturer-supplied or equivalent high-quality cables (motor and feedback device) between drive and motor. This results in intermittent electrical noise problems, and is difficult to diagnose. Highquality cables are more than just wires connecting two points; they are not inexpensive afterthoughts. Use manufacturer-supplied or recommended cables. If this is not possible for some reason, obtain the cable requirements from the drive or motor manufacturer. Plan the cable design as part of the total system. Budget up-front for proper cables.

Drive not matched to motor resulting in overheating or underperformance. If the drive is not pre-selected as part of a matched system, then great care should be taken to ensure that the drive will provide for the full capability of the motor and the application.

Inadequate feedback device. Common feedback devices such as resolvers, hall sensors, and incremental encoders can meet the needs of many, but not all, applications. For systems needing high accuracy, high response, or very precise resolution, it may be necessary to utilize higher-performance feedback devices such as absolute single or multi-turn encoders. When the wrong feedback device is utilized, the system may not be stable or may be incapable of operating at its design limits.
SCOTT/XTREME:
When specifying a motor it’s best to heed the advice of design engineers experienced with different motors and their unique properties.
MARK/DYNASERVO:
Once a customer simply replaced a stepper with a servo on a system with a compliant belt-drive and a wide range of variable load — they did this without using gearhead — to save money. It turned out that the servo gain could not be raised because of resonant vibration caused by compliance, and system performance changes when payload changed dramatically.

Servos are suited for high speed and precise positioning; it is not recommended (and it does not make sense) to use servos on very compliant systems.
BOB/IMS:
Here are some difficulties we’ve witnessed:

• Undersized motors that can’t move the load in the required time frame.
• Systems that are not defined well enough or that are defined with data that is in error.
• Oversized motors that are more expensive than they should have been.
• Environmental and mechanical issues that cause premature motor failure.
• Highmoisture content or corrosive atmospheres that cause motors to rust and fail in a matter of weeks.
• Motors operating in a vacuum so little or no heat transfer takes place.
• Motors operating on a bench, with no mounting heat sink during the prototype stage.
• Improper shaft coupling to the load, causing premature bearing failure.

Electrical mistakes can also cause premature drive failure. Examples include miswiring, operating at too high an input voltage, disconnecting or connecting motors with power on, point-to-point wiring, power lines that run with low-voltage signal lines, wire sizes that are too small for the appropriate current level, and improper grounding. Poor ventilation, and high noise susceptibility are other situations to avoid.

What are “best practices” when designing with servos and steppers?

CHRIS/KOLLMORGEN: Take the time to adequately define the machine structure and motor performance based on mass, force, and motion profiles. Realize that for existing designs the machine (especially its rigidity) may not be able to handle new and improved motor and system performance.

Many manufacturers offer software that helps in the selection process, both in solving the motion system requirements and identifying system components that will meet those requirements. Good manufacturers have large application engineer staffs that assist in the use of the software and in addressing special considerations that may not be covered by the software.

Select the right tool for the job. Don’t let a supplier that sells only one type of solution convince you to meet your motion system requirements with something that is not optimized for the job just because this is all it can offer. A servo solution can be overkill for an application that is readily done with a stepper. Cobbling together a variety of mechanical parts to do the job that a direct drive system is better suited for is trading a lower first-cost for a higher life-cost, and invites system control problems.

Use proper shielding and grounding techniques. As noted, cables must be part of the system design.

When possible use standard catalog products. This may require the designer to make adjustments to accommodate the standard, but it will keep costs down and makes support easier. If a standard product will not meet the requirements, see if the supplier offers modifications to the standard. This will keep costs in line, leverage the supplier’s installed base, and get you the product in a time frame that is usually only slightly longer than the pure catalog standard. When only a custom product will meet your needs, be sure to pick a supplier that is set up to do customs and has the experience to deliver what you will need. Ideally, you want a supplier that can offer you the solution that best fits your needs — whether a catalog standard, modified standard, or a complete custom.

Gain an understanding of all system components’ new technology and how to use it. Manufacturers invest considerable time in producing detailed installation and start-up materials. Review the materials both before you purchase and before you start the installation. Know what you are buying.
BOB/ITM:
• Motor sizing issues
• Confirm and reconfirm what parameters are assumptions and what aren’t. Point these out as the sizing progresses and try to get a confidence level from your customer as to how solid these assumptions are.
• Cross-check the given data with calculated results. For example, a 2-inch-diameter aluminum pulley that’s 1/2-inch thick should weigh less than 0.125 pounds. If the customer says it’s around 0.5 pound, maybe it’s made of steel or the dimensions are different. These errors have an impact on the load inertia and the size motor that is required to move it.
• Getting accurate information up-front helps you generate a costeffective system solution in a timely manner.
• Environmental conditions.
• Ask questions not only about the motion side of the application but also about the environment that it’s going to operate in.
• Emphasize that good wiring practices are important to the successful completion of the project. Run down the list of important practices that are needed. It’s easier to do it right the first time than to have to pull out a wiring harness and do it over again.
SCOTT/XTREME:
I am not sure this really applies to us.
MARK/DYNASERVO:
Best experience: built a three-axis gantry pick-and-place robot in three hours. The system consists of three servos with a gearhead on each axis, and one Fics-TPC1 touch-screen motion controller. From troubleshooting the pre-wired control panel, to system configuration, servo tuning, and motion programming, it took only about three hours to have the entire system move as desired.
DARYL/AGILE:
Fortunately the decision to switch from a centralized to a distributed servo system involves many functions within an organization, resulting in a team environment from the start. System-level planning is critical in a distributed servo application and having all specialties (mechanical, controls, software, drives) involved is important.

Continue on Page 2

What can servo/stepper manufacturers do to improve productivity?

DARYL/AGILE: Unifying standards for distributed motion controls is an area that would improve on the OEM engineering time. The number of different solutions that are presented to customers and how every implementation is different is confusing, and it’s difficult to understand the pros and cons of each. The OEM has to chose from different networks and different levels of distribution. Are they amplifiers that are connected on a network to a centralized motion controller, or is it a distributed motion controller with traditional analog interface to amplifiers, is it an amplifier, motion controller in one package?
CHRIS/KOLLMORGEN:
The primary contribution from manufacturers comes from increasing three factors:
1) Ease of use
— making the products easier to select, apply, install, operate, and maintain. So you get the right product, get it installed and running in the shortest amount of time, and can easily maintain it.
2) Breadth of offering — making more products, in more sizes, using more alternative technologies available to let users exactly match their needs. The right product for the job.
3) Educating users
— so users can understand motion products and their application. This encompasses the broad range of information and training that manufacturers make available to users.
MARK/DYNASERVO:
One of the challenges for servo manufactures is to produce next-generation servos that require “no tuning” and that adapt to any mechanical system without performance degradation. Others include integrating motion controller, drive, and motor in one unit to reduce wiring and cost. Motion and logic programming should also be integrated to improve productivity.
BOB/IMS:
Bring new innovative designs to the market place. Some designs simplify wiring and reduce panel space.
SCOTT/XTREME:
Allow for the extra design time, and maybe costs, to properly integrate a motion control system.

What can designers do to improve productivity?

MARK/DYNASERVO: Servos are quite different from steppers. Traditional steppers are open-loop (though there are also closed-loop steppers now) while servos are closed-loop. Servos feature high speed and high accuracy, but servo gains have to be carefully tuned to achieve good response and maintain closedloop stability.

Servos usually do not work properly for big inertia mismatch. Gearheads might be required to reduce inertia mismatch and to increase torque, especially for belt-driven systems and for variable load systems.
SCOTT/XTREME:
Know the precise motion control needs of the design. If designers know their precise needs and not a best guess, a motor can be designed and built to match the requirements — taking less time and costing less money to get the job done.
BOB/IMS: Designers would do well to define realistic product throughput goals for their applications. Sizing and component selection may meet defined needs, but if the production target moves, it may be extremely difficult to hit the new targets with the existing choices.
CHRIS/KOLLMORGEN:
Take advantage of the information, tools, and training that the manufacturers provide. Use manufacturers’ application engineering resources. Select the right product for the application.

What should end users do to maximize life and productivity?

BOB/IMS: Operate within the system’s specifications.
CHRIS/KOLLMORGEN:
First, and easiest, is to follow the manufacturer’s maintenance recommendations for the products they already have.

Second, end users can be proactive by demanding systems that put life cost as the primary consideration rather than what is cheaper to purchase at the onset of a project. As a case in point, direct-drive brushless systems (both rotary and linear) will generally cost more to purchase than systems that employ gearboxes, belt drives, or mechanical translation components such as ballscrews; but these direct-drive systems require little or no maintenance for the life of the system, and, as a result, maximize uptime while minimizing the maintenance costs.
MARK/DYNASERVO:
Maintain payload within the specified range. Do not increase speed without consulting your machine builder. Increasing speed usually leads to an increased torque requirement, and in turn generates more heat and possibly damages machines.

How do choices made involving servos and steppers affect other areas of the machine or system?

DARYL/AGILE: The choice to switch to a distributed servo system involves changes to the entire system. By putting the amplifier and controls next to the motors, there is a significant wiring reduction. While space needs to be available by the motors, a distributed system saves overall space.
CHRIS/KOLLMORGEN:
The type of linear actuator selected can drastically affect machine performance and size. Trade-offs arise relative to productivity and performance, and cost expectations and limits. Conventional mechanical translation devices that create linear motion (ballscrews and leadscews, for example) keep first costs down but can limit ultimate performance in a system that has other axes that will use higher-performance technologies such as direct drive. Once again, the key becomes selecting the right product for each axis based on a careful analysis of the total machine requirements. Designers should consider the optimal solution for each axis and then evaluate the resulting system performance.

It may be necessary to upgrade one or more axes to ensure the total system performance. In these situations, designers must obtain or have expertise in a wide variety of motion technologies.

BOB/IMS: Prime mover choices are driven by price and performance (dollars vs. speedtorque), followed by size, ease of use, and implementation cost. The sum of these dollars impacts what can be spent on other areas of the system. A solution with an inexpensive purchase price might be more expensive if it’s difficult to implement. A slight decrease in throughput or performance might result in a large cost savings. For example, a small motor with a lowbacklash gearbox might be replaced with a larger motor and no gearbox or with the same size small motor and a less expensive, slightly higher backlash gearbox.