You can't keep a good system down: Reliable machines perform their intended functions with few hiccups and no catastrophes, year after year. Electrically and mechanically simple arrangements have the best shot at consistent operation, but with planning and consideration, more complex systems can return unfailing performance as well. And with today's higher throughputs, “users see reliability as a given, and are unforgiving when they experience a lack of it,” says Andy Stanfield, manager of the repair and warranty group at Omron Electronics, LLC, Schaumburg, Ill.
Keeping drives alive
From an electrical perspective, the first step in reliable operation is to make sure controls are producing the intended signals. In variable-frequency drives (VFDs) voltage ripples pose a major challenge. The solution according to Ron Koehler, director of product engineering at Yaskawa Electric America Inc., Waukegan, Ill., is to sense the amplitude of ripple on the dc bus and stop the system if it exceeds the normal value indicative of single-phase input supply. “This goes a long way in extending lifetime and reliability of a three-phase VFD system,” he adds.
Another way to deal with ripple is to switch to non-electrolytic capacitors for dc bus application. These new film capacitors occupy less space in control cabinets and have higher ripple-handling capability than conventional electrolytic types.
Dc-link inductors are also on the rise; they shrink dc bus capacitor size for improved reliability. When built into VFDs, they also reduce dc bus ripple significantly for improved capacitor life. “The typical lifetime for dc bus capacitors is 5 to 7 years. But right now, drives with these inductors have a projected life of about 10 to 11 years,” says Koehler. Some designs include dc link inductors in both the positive and negative drive bus for even greater reliability.
“Automation and reliability engineering are two separate fields only now coming together — in closely integrated systems. These greatly simplify planning and installation of complex automation solutions,” says Sam Bandy, Engineering Supervisor at the Schneider Electric Motion Competency Center, Palatine, Ill. One example is a new breed of electric motor controllers — matrix converters — that provide direct ac-to-ac conversion without an intermediate dc bus.
Most industrial and commercial VFDs are designed for a three-phase ac power source. Increasingly, however, users are employing VFDs to power three-phase ac motors with a single-phase ac supply. This deteriorates the performance of capacitors and reduces their life dramatically. “Matrix converters are inherently more reliable than standard VSI-type converters,” says Koehler.
Another way to boost the reliability of electrical signals is to protect them from extraneous influences. For starters, proper shielding and grounding allow filters to drain high-voltage transients to ground. In fact, the bondings of all interfaces should be electrically connected, including the earth ground for the dc supply voltage. “Module interfaces with electrical isolation such as Profibus are the only exceptions. The voltage drop on dc power-supply lines must be kept as low as possible — certainly less than 1V — because at larger frame-potential differences between drives, communications and control signals may be affected,” says Bandy.
When system components are spread out, say over a plant, decentralized power units (closer to the drives) are better at supplying voltage. However, these individual power supplies must be grounded with the largest conductor possible and their internal 24-V signal power supplies shouldn't be connected in parallel.
“If the master controller (for example, the PLC or IPC) does not have electrically isolated outputs for each drive, it's up to designers to make sure that the current for the dc power supply has no path back to the power supply via the master controller,” says Bandy. How? “The master controller earth should only connect to the dc power supply earth at one point. This is generally the case in the switching cabinet,” he explains. For this reason, earth contacts for various signal connectors in compact drives are unconnected. There is already a connection via the dc power supply earth.
Suppose a controller has an electrically isolated RS485 interface for communication with the drives. In this case, the electrically isolated earth of this interface should be connected with the corresponding signal earth of the first drive. To prevent earth loops, this earth should only be connected to a drive. The same applies for electrically isolated CAN connections.
Electromagnetic interference can affect signals from control lines and adversely affect system reliability, so before operation, a system's electromagnetic compatibility must be verified. “Input and output filtering improves reliability in electrically noisy environments,” adds Koehler.
Drive systems must actually conform to the requirements of EC directives on EMC immunity under DIN EN 61800-3: 2001-02 for the second environment — where the grounding is established during installation. Drives must also be grounded to maintain the limit values for the EMC interference resistance and radiation — whether by a motor flange or electronics housing. Drives are commonly earthed by bolting the motor to an electrically conductive and earthed machine component.
“Here, make sure the electronics case is electrically connected to the motor. Earth the drive through the motor flange. If this is not possible, ground with a wire connected to a plug coverlid or with a cable clip to a flange,” says Bandy. Note, however, that the drive will not be earthed when the cover is removed.
Some additional advice from Bandy: “Earth shields on digital signal lines, over a wide area, at both ends or via conductive plug housing, and connect large surface areas of cable shields with cable clamps and tapes. Other measures can be taken to reduce EMC. Fieldbus, dc supply voltage, and 24-V signal-interface cables can be shielded. Also, make cable as short as possible, don't form any ground loops during wiring, and prevent capacitive and inductive fault interference.”
When it's hot, hot, hot
Heat is another major issue associated with lifetime and reliability. Heat generated outside and from within a system is conducted quickly through connected parts. “But motor technology that runs cooler than traditional brushless motors extends life, with less temperature degradation of motor components, especially lubrication,” says John Walker of Exlar Corp., Minneapolis.
Most components used in a motion system specify an ambient operating temperature, the maximum temperature permitted during operation. Dependant upon the distance from heat-generating parts and the unit's own output, this value is usually around 120° F or so. “End users should operate all machinery within this specified temperature range,” Koehler advises.
Transport and storage temperature limits are another factor. “There are always specific ranges within which a product can be stored without degrading life span. This is typically about -10 to 160°F,” says Bandy.
Water can help or hinder system reliability. For relative humidity, the allowed range is usually wide — 15 to 85%. A dehumidifier should be used if humidity is excessive. But systems cooled with liquid in general are increasingly common, because any improvement in the cooling system — including a tighter, closed-loop approach — improves life expectancy. One recent trend: Power semiconductor chips built into water-cooled heat sinks. “This integration reduces inverter size and improves cooling-system efficiency,” says Koehler.
Circulating lubrication is another way to cool systems down. Ironically though: “Besides waste, safety, and environmental considerations, overlubrication can contribute to material buildup and overheating that may actually damage machinery,” says Kurt Rommelfaenger of Oil-Rite Corp., Manitowoc, Wis. “The result is apparent in constant failures and repairs.” In fact, the solution is not always as easy as cranking it up a notch. The lubrication may not be reaching the right point or may not be applied at the proper intervals. That's why some lubrication systems include self-monitoring capabilities.
Similarly, sensor circuits that detect when their fan has stopped working are increasingly common; these shut off VFDs immediately to prevent damage. Placing power-switching banks in an airflow path and protecting them from atmospheric contamination helps extend their life. “Because heat management is perhaps the most important part of a reliable design, cooling systems should be designed with the possibility of a total fan failure in mind,” says Koehler. Another design feature that extends life: “Thermal switches on heat sinks (to serve as a backup thermal protection system) that prevent thermal runaway of semiconductors,” he adds.
In VFD systems, reliability (and hence, life) very much depends on the choice of components and ambient operating temperature — so component selection and heat management are key factors that influence their reliability. “Even a good component able to withstand very high temperatures will not function well if subject to continually excessive heat. That's where thermally efficient heat sinks, high flow fans, and strategic placement of heat ducts and cooling systems really extend system life,” Koehler concludes.
Another step end users can take: Ensure that cooling vents are not blocked and the fans are in operation. This should be part of daily or weekly maintenance checks.
The cost trade-off
There's a reason Bose headphones cost more than the pairs at the corner drugstore. “Typically, those systems with higher reliability and lifetime come with higher cost,” says Walker. “That cost difference can induce system designers to choose lower-cost product.”
System designers typically provide details of operating conditions, performance expectations, and equipment life expectancy. “If there is a cost tradeoff as lifetime requirements increase, if the cost is excessive, this is when designers can add provisions to offer various design concepts, or create designs to reduce downtime in case of a failure,” says Koehler.
Alternative designs are one solution. “Components like roller screw linear actuators that perform the same function as more common components — in this case, ballscrews — sometimes offer reliability as well as cost savings,” adds Walker.
System design, construction, installation, and integration with other manufacturing processes can represent a very large investment of capital, which may be recovered only over extended periods. “But even though the acceptable lifetime for a total system may be years or even decades, individual components may have much shorter life spans,” says Rommelfaenger.
It is easy to choose components with the lowest price, but they may or may not be the ones with the lowest cost of ownership. “If long life and reliability are critical, take care to select the proper components during the design stage,” says Charles Manning, operations manager of Omron's Service Plus drive centers.
“Sometimes a product that seems more expensive at initial purchase is actually a better investment because it lasts longer over time,” says Joseph Ciringione, Energy Chain business unit manager of Igus Inc., East Providence, R.I. He adds that designers must consider every angle when determining which cable carrier is right.
“We have systems that have been running for more than twenty years, but in general, designers want components that last about as long as the average machine — in our experience, about five to seven years,” he adds. But if a component only needs to run for a short time, then lower-cost, shorter-life componentry is the way to go. “There is no one way to determine what's best; that's why the application and its parameters must be considered in their entirety. But any time a product is selected based on price and not how it will be implemented, reliability can be compromised.”
Save the drama: Specs and maintenance
All electromechanical components have a given calculated theoretical life under different conditions — a life that either meets or doesn't meet specification requirements. “With mechanisms, this estimated life is usually well defined by calculations. Electrical components like motors, on the other hand, typically last longer if not used at their absolute performance limits. Some overhead to the application specifications is typically a good thing to consider in product selection,” says Walker.
Adhering to parameters for extended machine life is fine and dandy — but what if a machine's function changes? What if the parameters shift, and suddenly more speed, more throughput, or more load carrying is required?
“Lifetime is not only the length of time a machine operates without major overhaul, but also its ability to adapt as new technology becomes available,” says Stanfield. “Open frame control systems address this issue,” he says. In contrast, single-purpose control systems can have short life systems, and often need a complete reworking to adapt to new technology.
So what determines the acceptable lifetime for a system from this perspective? “Length of time between product or process revision,” Stanfield adds. In fact, manufacturers are increasingly creating communication schemes that operate easily and efficiently with other manufacturer's equipment. “Not all, but some, are making their software forward and backward compatible, so when newer technology is released, it's easily adapted to systems and processes. Then the life of the equipment is extended and upgraded in steps as need, budget, or future planning dictates — without necessitating starting from scratch,” says Stanfield.
A large detractor to product life and reliability: Failure to adhere to recommended maintenance schedules. Much like changing the oil in a car, maintaining proper lubrication in mechanical products is a key to maintaining their life and reliability.
“The same is true for seals and other protective system components. Maintaining them can be a simple and inexpensive way to ensure long, reliable use of a system,” says Walker. “For mechanical products, lubricants continue to improve, as do sealing technologies. Taking advantage of these new products offers a manufacturer the opportunity to extend the life of standard products. Operating products at their limit usually leads to more maintenance requirements,” he adds. “In the case of a mechanical product, where life is directly related to the load (torque and force) it's required to produce, operating it at maximum load shortens calculated theoretical lifetime.”
“Lifetime and reliability are probably the two biggest factors besides performance that are key to consider when designing a motion system,” says Ciringione. “If a component lasts as long as the machine itself, it eliminates maintenance and replacement fees. Indeed, downtime can be extremely costly when production numbers are affected in a matter of minutes.”
In situations where outside influences are harsh, this is particularly true. “A system's lifetime is strongly affected by operating conditions, so they're designed with electrical, mechanical, and chemical environmental conditions in mind,” says Koehler. Regardless of operation or control method, a machine is designed so all its components operate at original specifications. “But when wear and variance from these original specifications begins, the effect is exponential,” warns Stanfield. In the end, an engineer's familiarity with his or her design goes far in boosting life.
One basic tenant is readily accepted by engineers and maintenance people alike: Proper lubrication extends the life of components and reduces the cost of ownership for machinery. “Yet methods often are insufficiently addressed in the design phase, and not properly implemented in the field. We routinely correct lubrication methods for plants that have, by their own admission, limped along for years,” Rommelfaenger explains.
Of course, a great approach to extending machine life is regular preventive maintenance. But another approach is predictive maintenance. “With predictive maintenance, you can establish operating parameters or quality standards that indicate normal manufacturing quality,” says Manning. Continual maintenance can be made easier with system monitoring. “Also, software can be upgradable to include fan sensors, heat sink temperature sensors, single-phase operation, and other protection functions for improved reliability,” says Koehler.
One caveat: A designer must know which parameters should be observed for changes or irregularities. But once established, if parameters or standards are violated, then the component or machine can send a flag indicating that maintenance is required.
“I think we'll see a trend toward more and more remote monitoring and operation as the proliferation of wireless technology continues,” Ciringione adds. “For example, if an engineer can see where a cable carrier, for example, is positioned at any time from a remote site, he can ensure continuous performance and anticipate any problems.” But Ciringione says periodic inspection is still the best way to ensure service life. “Make sure components are operating properly and are free from debris or obstruction. Early detection of a problem can save a lot of time, money and headaches caused by a machine that breaks down unexpectedly.”
Motor and drive diagnostics are increasingly important to reliability as well. Learning algorithms embedded in modern software can monitor current signatures and predict abnormal operation. “Such software is already available in the European market and is fast becoming a desired feature worldwide,” adds Koehler.
The geometry of a system directly affects its chances for long, trouble-free operation. For starters, it must be strong enough — or flexible enough to withstand vibration and shock loading. Component strength during oscillation stress corresponds to EN 50178 Section 126.96.36.199 and EN 61131 Section 188.8.131.52. The electronics must be protected from the harmful effects of vibration-induced transients and harmonic contents.
Also, shields should be connected at both ends for fault protection. Potential differences can result in excessive currents on the shield and must be prevented by equipotential bonding conductor cables. If lines over 100 m are approved, the following applies: A cable cross section of 16 mm2 is sufficient for up to 200 m length; for longer cables, a cross section of 20 mm2 is required.
Installation height is another consideration. “This should normally be less than 1,000 m above sea level. A factor of duration is needed if the altitude is too high,” says Bandy.
The mechanical tonic
It can't be helped that chains, gears, and bearings wear with time. But lubrication is a tonic that cures many ills. “In mechanical design, a great deal of time is spent identifying the weakest link — the part most likely to fail from wear through friction, stress, or impact,” says Rommelfaenger. Parts that include a bit of sliding during their operation are suspect. But a comparatively small effort is directed towards designing a lubrication method — even though proper lubrication can be essential to maintaining the integrity of a component during prolonged operation. “That's why we often step in to correct, improve, or sometimes even establish a lubrication method for some of the world's finest engineered machinery,” adds Rommelfaenger.
“A lubrication system optimized for particular conditions can reduce meantime between mechanical failures and reduce the cost of ownership of expensive machines, and a machine with a good track record will typically outlast the machine that is always down,” says Rommelfaenger. In fact, the acceptable life of a system may ultimately depend as much on a maintenance philosophy as it does its functional design. Dispensing lubricant is itself a mechanical undertaking, and thus subject to its own lifetime and reliability considerations. “But well-designed lubrication pumps survive over 360 million cycles without appreciable wear,” he says.