An update on flicker

Electrical Apparatus, Mar 2007 by Nailen, Richard L

Standards for measuring the annoyance of dimming lights, difficult as it is to quantify, are now going international

TO MOST OF US, "POWER QUALITY" MEANS harmonics or transient voltage spikes, which disrupt processes, sabotage controls, or damage equipment. More than a century before those concerns arose, however, engineers began talcing notice of a different power quality problem that bothered electric utility customers and still does. That wa& flicker-the repeated dimming and brightening of electric lighting caused by voltage disturbances. Because of its electrical origin, this is sometimes improperly described as voltage flicker.

Definitions of the phenomenon don't involve an observer's level of annoyance-only whether or not the fluctuations are perceptible. Here, for example, is what IEEE Standard 519 has to say: "Flicker is the impression of unsteadiness of visual sensation induced by a light stimulus whose luminance or spectral distribution fluctuates with time."

How is an annoyance level determined? By asking the opinions of large numbers of observers. A deep sag is tolerable if it doesn't happen often. The more rapid and persistent the fluctuations, the higher the annoyance level. Conversely, if the voltage sags only slightly, the individual disturbance attracts little attention and tends to be ignored even at a higher repetition rate.

In pan, those relationships derive from the nature of the light source. An incandescent bulb emits light from a heated filament. Because the heat takes time to dissipate upon a drop in voltage, the bulb will remain lighted throughout a brief disturbance, but with diminished light output. For different reasons, the same is true for some newer types of lighting equipment.

Early studies of flicker

During the years when simple incandescent lamps constituted most electric utility system load in urban areas, switching other loads on and off produced voltage dips, causing noticeable variations in light output. Motors and electric welders were especially troublesome. Utilities began making tests to determine how readily such variations could be perceived, and what level of voltage disturbance resulted in customer objection to those variations. One of the earliest studies was published in 1914. Another appeared in 1926.

Reaction to flicker varies with not only the frequency and magnitude of the voltage disturbances, but with their waveshape. Most tests were taken using square wave voltage changes-an abrupt dip in voltage to a constant reduced value, followed by an instant rise to the original level (Figure Ia). Little attention was paid to the effects of voltage dips having triangular, sinusoidal, or other wave shapes, as in Figure Ib.

Most early research involved subjecting groups of individuals to lighting environments supplied with controllable voltage fluctuations at various repetition rates. Each study produced somewhat different levels of both perceptibility and annoyance. Curves were plotted relating those levels to the magnitude and frequency of voltage changes. Understandably, such so-called "flicker curves" varied widely (see, for example, "Voltage flicker" in EA August 1996). Such curves deal only with voltage fluctuations occurring at a regular frequency. The disturbances may be aperiodic (at random intervals) and be equally annoying.

What is "tolerable," "tolerable but annoying to some degree," or "intolerable" was (and remains) entirely subjective. No two individuals are likely to experience the same degree of annoyance (if any) when exposed to specific flicker phenomena.

Other researchers have tried to quantify toleration, however. Data reported by 21 groups of observers analyzing data from thousands of tests resulted in a 1937 curve still in use 40 years later by some utilities. However, the most widely accepted plot of both perception and tolerance levels of voltage disturbance magnitude as a function of repetition rate was one developed in 1930 by the General Electric Co. and a group of utilities. Still known today as the "GE Curve," it appears in some IEEE standards (see Figure 2) such as No. 141, the Red Book.

Much of the literature on the subject (including an important 1970 paper by C. Rashbass titled "The Visibility of Transient Changes of Luminance," published in the Journal of Physiology) deals only with limits of perception. Figure 2 shows a single line as the "borderline of irritation." The 1937 report did include one plot showing the degree to which observers considered flicker "objectionable." The range was extremely wide (typically 2:1 depending upon frequency). The plotted data, however, went no higher in frequency than 60 disturbances per minute (one per second).

The curve in the 1986 edition of IEEE 141 differs markedly from that in the latest (1993) publication. For example, at a disturbance rate of 60 per minute or 1 per second, the 1986 curve shows a 3.5% dip in voltage as the "threshold of objection." For the same repetition rate, the 1993 curve gives a "borderline of irritation" at less than 1% voltage dip. No explanation for the discrepancy has been given.