Considering the type and quantity of loads placed into most facilities, one fact becomes clear: The lighting system is one of the most pervasive of the electrical loads. Lighting can create localized electrical field problems and affect susceptible electronic equipment.
Let's discuss some of the issues associated with electrical noise (both conducted and radiated) that lighting systems can generate.
Typical lighting scenarios
Field installation guides for computer systems frequently contain warnings that advise users to avoid powering fluorescent lighting systems from the same panelboards that supply power to computer systems. In large facilities, it's unusual to find lighting systems powered from the same electrical panel that serves computer equipment, since lighting systems are often powered from 480/277V, 3-phase, 4-wire systems. However, you can't say the same for small office environments in which 208/120V, 3-phase, 4-wire is the sole power source configuration. Even if the 480/277V and 208/120V systems are separate, installation problems may enable coupling between the power systems.
To get the desired light coverage (lumens per square foot), you typically need a significant quantity of lighting fixtures. For instance, lighting coverage for an older commercial facility is six lighting fixtures per 400 square feet of floor area. Each fixture contains four lamps rated at 32W or 40W, which amounts to 768W to 960W for the area. Modern facilities equipped with T8 lighting systems might use 3-lamp fixtures for the same area. With each lamp rated at 32W, the energy usage drops to 576W for the same 400-square-foot space.
From the above comparison, it's easy to understand why power for lighting systems is one of the most dominant uses of power in a commercial building. Indeed, a large modern office building may have thousands of fluorescent lighting fixtures.
While energy usage is certainly important from a cost perspective, the power signature of lighting systems is equally important, but from an electromagnetic compatibility standpoint. Older fluorescent lighting systems used ferroresonant ballasts, which were not extremely efficient. Also, the power factor of their current draw was often quite low. Figure 1 shows the current draw of a ferroresonant lighting system, where the displacement power factor is very low (roughly 0.4). On the other hand, the current distortion is only 19% THD.
In contrast, Fig. 2 and Fig. 3 show the current draw of modern lighting systems. Figure 2 shows the current draw of an electronic discharge lighting system (fluorescent lamps with electronic ballasts). The displacement power factor is nearly unity, but the current distortion has increased dramatically (49% THD). Figure 3 shows the current distortion caused by the small, screw-in compact fluorescent lamp, which is sold throughout the United States in hardware and home improvement stores as a replacement for the standard tungsten (incandescent) lamp. With these compact fluorescent lamps, the current distortion reaches 120% THD, and there is actually a small amount of leading displacement power factor.
From a facility standpoint, the impact of a single lamp or fixture is quite small. However, the cumulative effect arising from widespread usage can be significant. For example, we recently measured the signal shown in Figure 2 in a facility in which we were conducting a power quality survey.
Electronic discharge lighting fixtures were used throughout this facility. As you can see, the effects of the signals upon susceptible electronic equipment were very negative. However, distortion and interference effects need not reach epidemic proportions. Specifications for a fluorescent lamp should include good displacement power factor as well as low current distortion. Some ballast manufacturers work hard to control power factor and emissions. You can also take steps at your facility during construction to isolate lighting systems and localize interference effects.
Whereas the old ferroresonant ballast relied on a saturating transformer for regulation of lamp current and voltage, today's modern electronic ballast rectifies the applied AC voltage and then supplies the lamp with high-frequency voltage from pulse-width modulation (PWM) inverters. As a result, a high-frequency electrical field, which matches the applied voltage from the PWM inverters, develops at the fluorescent lamps.
On the surface, you might view this development without major concern. After all, the old magnetic ballasts and fluorescent lamps also created magnetic fields. Well, the difference revolves around the voltage supplied to the fluorescent lamp.
The supply voltage of older style magnetic ballasts was essentially sinusoidal and 60 Hz, so that electric field surrounding the lamps was more or less sinusoidal and 60 Hz as well, with limited high-frequency content. However, PWM circuitry inside today's electronic ballasts generates the voltage that's supplied to the lamp. If the PWM circuitry operates at 25 kHz, then the electric field will reflect the 25 kHz signal.
Ballasts for 2-lamp configurations will normally operate at a different frequency than ballasts for 4-lamp configurations. Compound this condition with the fact that fixtures may be wired master/slave, and conditions really get interesting. In a master/slave configuration for T8 lamps, two ballasts in one fixture will supply the voltage to that fixture and to an adjacent fixture. One ballast will power the center lamp in each fixture (23 kHz), and the other ballast will supply the outer two lamps in each fixture (25 kHz). The slave fixture will have no ballasts.
Because the two ballasts usually operate at different frequencies, the resulting electric field signal will reflect the combination of the two ballasts (e.g., A, B, A+B, A-B). Figure 4 shows the electric field measured near a master/slave fixture. Figure 5 shows the electric field measured near a fixture with a ferroresonant ballast.
Effects on susceptible electronic equipment
As previously discussed, electric fields develop around fluorescent lighting fixtures, and today's electronic ballasts create high-frequency fields. The impact of these fields depends upon the sensitivity of equipment within those fields and the distance from the fixtures to the equipment.
Most commercial equipment today is efficiently shielded to withstand the fields; however, these fields can adversely affect very sensitive electronic equipment, such as scanning electron microscopes, electronic beam etching equipment in semiconductor fabrication facilities, and receivers for partial discharge testing in industrial applications. In fact, any parts-per-million, high-impedance measurement system that does not have sufficient shielding can potentially be affected.
Distance is certainly a factor as well. At the fixture, electric fields can typically be about 80V/m, and the field strength decays rapidly with distance, decaying as the square of the distance. But it's not just super-sensitive scientific devices that are affected. Hearing aids, for instance, may interact with the electric field signals and then generate an unwanted hum, hiss, or whine. For some individuals, simply walking into certain buildings can become a problem.
One step ahead
You should consider adverse interference effects when making your selection of lighting fixtures. Our advice is to control the end results in the beginning rather than spend countless dollars and hours combating problems after the fact.
To minimize interference effects, take the following steps:
Install all wiring for lighting in separate conduits and power lighting circuits from dedicated-use isolation transformers.
Never combine lighting circuit wiring with computer system or other susceptible electronic system wiring in the same raceway.
In areas with susceptible equipment, use lighting systems without high-frequency electric fields (e.g., DC) or lighting fixtures with electromagnetic interference screens to control the electric fields.
Also, there are a few simple precautions you should take for installing lighting systems near computers and data cables. They aren't new, and you can find them in numerous site preparation manuals. Nevertheless, they warrant repeating here:
Keep lighting circuits out of computer power panels.
Do not install data cables so that they run on top of fluorescent lighting fixtures because the magnetic fields from electronic ballasts may inductively couple into the data cable.
Make sure that you do not install poorly shielded data cables in the electric fields below lighting fixtures.
Shaughnessy is a power quality consultant and vice president of PowerCET Corp. located in Santa Clara, Calif.