A few months ago, we answered a customer’s question about what equipment might be affected by a particular type of power quality phenomenon. The answer was based on Table 4.2 from IEEE Std 1159 Recommended Practice on Monitoring Electric Power Quality, and it seemed to be a logical way to address what can be a fairly complex process.
The customer wondered, “Why can’t I just use the CBEMA curve?” This hit one of my hot buttons, as there is significant misuse of the CBEMA (sa-BEEM-ah) susceptibility curve, which came from the Computer Business Equipment Manufacturers Association back in the 1960s and was publicized in the “Guidelines on Electrical Power for Automatic Data Processing Installations” (Federal Information Processing Standards publication DU294). The susceptibilities outlined by the curve are based on those of computers of the time, namely large mainframes, which aren’t something that this customer or many others have in their facilities.
The CBEMA curve actually was replaced before the turn of this century by the ITIC (“I tick”) curve, which comes from the Information Technology Industry Council. It uses straight lines, rather than curved lines, to represent the so-called susceptibility limits. An example is found in the figure (right). The idea is, if the power quality event’s magnitude and duration fall within the two lines, your equipment most likely is not affected. If outside the lines, you probably had a problem. Again, this susceptibility curve is based on a specific type of load, namely single-phase computers (such as a PC) and peripherals. Herein lies the problem with taking the easy route.
Most electrical loads are three-phase, and the majority are motors, often with adjustable speed drives (ASDs). Their susceptibility to power quality variations is not the same as that of a PC. Which of the three voltages that are affected by the disturbance, for how long, and at what phase angle the disturbance starts have a significant impact on whether the motor rides through the disturbance, along with what the loading on the motor was at the time. It may be the undervoltage or phase imbalance relay that is much faster acting and will trip off with a less severe disturbance than would affect the motor. What if it is a motor on an ASD? Some ASDs are susceptible to tripping offline on an overcurrent condition when the power inputs are subjected to an oscillatory transient, which can result from a power factor capacitor switching in to the distribution system, which doesn’t even show up on the ITIC curve.
In most businesses where computers play a critical role in daily operations, the power sources of the PCs are supplied from an uninterruptible power supply (UPS), and though not immune to power quality phenomena, these PCs aren’t likely to be affected by events that show up outside the lines on the ITIC curve. Laptops have their own mini-UPS built into them, as they run off the internal batteries that are charged by the AC adapters. Of course, these batteries, and all UPS battery systems, must be properly maintained and replaced per the manufacturer’s recommendations.
Getting back to transients—though the ITIC (and CBEMA) goes down in duration to 0 seconds, that would imply very fast power quality phenomena, such as transients, would be covered. Several factors make that a problem. First, the standards-making groups have not addressed how one accurately measures the amplitude and duration of a transient, as they have with rms variations, such as sags and swells. Transients often are measured by their peak voltage, not the rms voltage as on the chart. Duration can be measured from the 10 to 90 percent of peak or the 50 percent rising to 50 percent falling portion of the peak of an impulsive transient or some other method. And what about negative impulsive transients, those that subtract energy from the voltage sine wave, such as notching events? These can be destructive to some equipment yet are not represented on the ITIC curve.
This doesn’t mean the ITIC curve is fundamentally wrong. The work was done by knowledgeable people with an effort to improve the operation of the equipment and facilities related to those interests that they represent. Rather, it is the misapplication of the curve to try to get an easy answer that ticks me off. Like any tool, it should be used for the application for which it was designed; otherwise, the results that you get are misleading or flat-out wrong.
BINGHAM, a contributing editor for power quality, can be reached at 732.287.3680.