Riding Those Sags and Surges

Jan. 14, 2007
Keep automation and profits safe from ups and downs in power quality.
Blank computer screens greeted the operators of a new assembly line at Delphi Corp.’s Saginaw, Mich., plant as they reported for work one Monday morning. Nine of the thin panel computers controlling each of the line’s 22 stations had died some time over the weekend. The operators had experienced similar welcomes when they returned from weekends before, but the problem was occurring with greater frequency now that the beta test had progressed to the point where the giant automotive supplier was ramping up production of the new steering assembly being made on the line. Of course, Delphi sent the computers back to their manufacturer. But the repairs wouldn’t be covered under warranty, because the problem was due to power quality, not the computers themselves. Delphi had become one of a growing number of companies learning the hard way that sags and surges in electrical power can ruin the sophisticated, but sensitive electronics in automated equipment and computer networks.“Because many of the nearby plants are down over the weekend, the voltage in the area rises a little bit,” explains Andy Hernandez, the staff electrical engineer sent to solve the problem. “The slightly higher voltage propagates switching transients and other anomalies in the power quality a little more strongly.” In Delphi’s case, the boost was enough to propagate short, but devastating power sags. During these events, the equipment tries to continue drawing the power that it needs according to a fundamental law of physics, the power equation (P = I2R, where I is current and R is resistance). “When the voltage goes down, the current rises to keep the power constant and can overheat the device,” says Hernandez. The tiny, low-power chips in today’s advanced electronics are much more sensitive to fluctuations in power than the relays and starters in conventional automation.Because the sags were less than two seconds, Hernandez found that he could compensate for them with a kind of dynamic sag corrector called DySC, developed by SoftSwitching Technologies Inc., of Middleton, Wis. Hernandez installed two of them, one for each of the two circuit breakers protecting the 22 panel computers connected to the 120-Volt lighting panels. During a short voltage sag in three-phased power, the correctors use internal inverter and cross-coupler technology to borrow power from the unaffected phases to boost voltage in the affected phase. They also can draw from capacitors whenever a sag affects all three phases and consumes less than 1 percent of the energy in the application and uses no batteries.Because the sags at Delphi’s Saginaw plant typically lasted between five and 15 cycles, the conventional solutions to the problems were overkill. The constant-voltage transformer (CVT) that Hernandez tested rode through sags well, but the cost was high. “Both it and SoftSwitching’s device held the voltage constant up to 120 cycles at a 50 percent sag,” he says. “Because we had 22 stations with these panel computers, we would have needed to oversize CVTs dramatically.”The other conventional solution—batteries or uninterruptible power supplies (UPSs)—also was overkill. A small UPS offered 11 minutes of backup, but cost more than twice the price tag for SoftSwitching’s dynamic sag corrector. “We didn’t need a UPS because we’re not looking to ride out an interruption for a long period of time,” says Hernandez.A Mix of TechnologyDespite its advantages, the sag corrector does not necessarily preclude the use of other solutions. For Hynix Semiconductor Manufacturing America Inc., of Eugene, Ore., for example, the unit is part of a mix of technologies in its 1 million square foot facility making computer memory chips. The company maintains several megawatts worth of back-up power through generators and UPSs. Not only are both mandatory in the fire code to give workers enough time to evacuate a building safely during an emergency, but the UPSs also buy enough time for the computers to shut down correctly.Hynix added DySC to its mix of power-management technologies to prevent the huge losses that it was experiencing from power sags, which had been causing more problems than power interruptions had over the last few years. Although the company has not had an interruption in its power during this time, it has experienced a couple of dozen sags a year. In about half of them, the voltage dropped below 75 percent of nominal or lasted more than four or five cycles. “The Northwest has a fairly stiff grid and multiple interconnected transmission lines,” explains Terry Prottsman, the facilities electrical team leader at Hynix. “It improves the reliability, but increases the transfer of incidents throughout the system.”These sags can be very disruptive and expensive because they stop some of the equipment in its tracks. The loss of power in some units, for example, releases a vacuum holding a gas in suspension in a chamber, causing the gas to splatter in the chamber and compromise its purity. A sudden stoppage also damages the silicon wafers in various stages of production, sometimes irreparably and sometimes not. “The work on the wafer is microscopic so it takes time to sort out the state of the wafers,” says Prottsman. Roughly two days are necessary on average to find the cause of the stoppage, determine what work-in-process must be scrapped or repaired, clean and re-qualify contaminated equipment, and restart the process in the correct sequence. As an around-the-clock operation, the company often must call in off-shift workers to assist in resolving the problem. So 12 events a year adds up fast to serious money. As well as the generators and UPSs worked as backups for power interruptions, they were not as effective for sags that last less than a few seconds. “The generators are too slow to react to a sag,” says Prottsman. “The UPS systems are 90 percent to 92 percent efficient and require a lot of care and feeding.” For this reason, he, like Hernandez at Delphi, thinks that they are too expensive when the need is for less than a few seconds. So he also decided to try SoftSwitching’s DySC correctors.To prove that the technology would work without having to invest much time and money, Prottsman started with a small unit, one for an air handler that needs only a few amps worth of protection but is extremely sensitive to the smallest of sags. Connecting the corrector without interrupting the handler was a challenge, but doable. With each success, Prottsman increased the size of the applications until he graduated to a 10-MVa DySC unit. The technology is scalable, meaning that SoftSwitching offers models both for small applications—such as contactors, relays, compressors, and programmable logic controllers (PLCs)—and for large ones—such as bus-level protection, computer numerical control (CNC) lines and aseptic packaging.Stopping Surges on the InsideAlthough power sags tend to cause more problems than surges in industrial settings these days, surges can be devastating too. A surge along an Industrial Ethernet data line, for example, is usually quite expensive to repair, especially when you consider the cost of downtime that it causes.“Replacement cost of the damaged equipment might be insignificant to its value in the revenue stream,” says Ed Doherty, product manager for Trabtech surge protectors at Phoenix Contact Inc., of Harrisburg, Pa. “The value might be $1,000 per hour for a water well or more than $100,000 per hour for an energy pipeline application.”He thinks that many companies have left themselves vulnerable to surges by installing inadequate surge protection—that is, on the AC power side only. A surge can travel along any wire, whether it carries power or data. By not protecting all connections to your equipment, including phone lines and cable modems, you leave a back door open to destructive forces. “Although the energy on phone and cable lines is less than on power lines, they enter sensitive areas of the PC and are more likely to cause a catastrophic failure,” notes Doherty.Although home computers have sensitized most people to the need to protect electronics from surges from external sources such as lightening, many people have not thought about protecting their equipment from surges generated inside their plants independently of the weather. Just the normal cycling of equipment transmits electromagnetic impulses into the air, much like lightening does. Any device with an antenna can receive those disturbances, like radios do during storms. What most people don’t consider is that any length of wire, including a power line or Ethernet cable, is an antenna.Although the twisting of the wires inside the Ethernet cable eliminates nearly all of the static burst, or noise, generated by the impulse, it does not eliminate the capacitive effect. “So a damaging voltage can be created and travel on an Ethernet cable to the equipment,” says Doherty. “The voltage need not be very large to cause damage, because microprocessor voltage is usually in the five to 12-volt DC range.”Industrial plants are not like offices, where the magnitude of the power and noise is much less, and you can get away without putting surge protection on Ethernet cables. “The only way to protect equipment in the plant is to identify power and communication paths to a piece of equipment and place protection on each one,” says Doherty. Taking this precaution, and finding an appropriate way to fill sags, will help to smooth out your power curve and make it much easier for your automation to ride those sags and surges.To see the accompanying sidebar to this story  - "Betting on Better Batteries", please visit www.automationworld.com/view-2851For more information, search keywords “power quality” at www.automationworld.com.