Triplen harmonics lead to poor laser scanner performance and power circuit damage.
The consolidation of equipment was supposed to make things easier for a manufacturer of printed circuit boards. But when it lead to an increase in service calls, the manufacturer sought help from its utility.
The manufacturer used laser technology scanning devices in its quality control process to microscopically scan newly etched boards for short circuits, voids, and other imperfections. About six months after consolidating equipment, one of the scanners began logging a dramatic increase in "no-trouble-found" service calls from the scanner's service personnel, who tried everything to improve performance but were unsuccessful. Performance continued to degrade until the machine eventually failed completely. A quick diagnosis revealed that the input power plug to the machine had burned and melted.
This same problem cycle began to appear on the other machines over the next few months. At this point, the scanner service personnel connected a power line monitor at one of the failed machines in a further effort to diagnose the problem; it indicated that the voltage supplied to the scanners was fluctuating beyond the equipment specifications. The service company then proposed that the customer purchase a voltage regulating, tap switching transformer at a cost of several thousand dollars to correct the "utility voltage problem."
Utility-conducted power quality investigation
Before deciding to buy this special transformer, the manufacturer contacted its utility, Northern States Power Co. (NSP), believing there was a problem with the power being supplied to its facility. NSP referred the company to its Ultra Power Service, a power quality consulting service assisting customers with power quality issues. After discussing the problem, the customer asked Ultra Power to conduct a power quality investigation at its manufacturing facility.
The electrical distribution system to the scanner room includes a 75kVA, 480V delta-208Vwye/120V transformer that served only the six scanners. Each scanner had two dedicated 3-pole, 20A, 5-wire circuits running in a 3/4-in. conduit: One supplied an electronic control cabinet; the other supplied the laser scanning section. Circuits were wired with No. 12 AWG wire and terminated at a NEMA L21-30R type receptacle.
Inspection of the scanners revealed that they did not actually use 3-phase power, but rather contained several individual single-phase, 120V switch-mode power supplies to power the various components of the machine.
The investigation began with an examination of samples of the failed power plugs. Ultra Power noticed that the neutral pin of the NEMAL21-30P type plug was the most severely burned. (See photo on page 90.) The heat had been so severe that the plug had to be sawed off to remove it from the receptacle. This evidence lead to the immediate suspicion that a high level of harmonic currents was the cause of the problem symptoms as well as the failed connectors.
Measurements provide the answer
Ultra Power took measurements with a harmonic analyzer of current, voltage, and harmonics at the secondary of the 75kVA transformer, and at circuits supplying one of the scanners. Fig. 1, on page 90, and Fig. 2, on page 91, show some of the results. The measurements indicated a classic example of the additive effects of triplen harmonic currents on a shared neutral conductor. As the waveforms at both locations show, the neutral currents are approximately 1.5 times higher than individual phase currents and present a near perfect sine wave at 180 Hz.
Neutral current measurements made with true-rms and average responding type digital meters were virtually identical. This can be attributed to the nearly perfect sinusoidal wave shape of the neutral current.
Based on the visual inspection of the plug and receptacle, and on the measurements, the conclusion was that the continuous heating effect of the high percentage of harmonic current on the neutral conductor was the culprit. Due to the pressure-fit nature of the plug-and-receptacle connection, this point of the circuit had the potential to be a location of slightly higher resistance and, therefore, even more susceptible to the harmonic heating effects. Eventually, this connection deteriorated, neutral-to-ground voltages increased, and line-to-neutral voltages decreased. Results were the voltage variations observed by scanner service personnel. As this continued, the scanner's performance degraded until the connection eventually failed.
To solve the problem, the circuit board manufacturer chose to change the input wiring to the machines from No. 12 AWG to No. 10 AWG. Also, the plug-and-receptacle connections were eliminated, with the scanners being hardwired.
Craig S. Lofdahl is a Consultant for Ultra Power Service Div., Northern States Power Co., Minneapolis, Minn.