Guide to Mechanical / electrical equipment for buildings: PART III: ILLUMINATION: Light Sources

DESIGNING THE LUMINOUS ENVIRONMENT today involves a complex balance of considerations, such as illuminance, luminance, view, visual com fort, shading, and glazing types-and associated concerns such as thermal comfort and energy efficiency. Historically, human activities and tasks were relegated to daylight hours and often within proximity of tall windows. Candles and oil lamps were expensive and a fire hazard, in addition to providing poor illumination for certain tasks. Electrical lighting began around 1870 with the development of commercially usable arc lamps and was given greater impetus nine years later by Edison's first practical incandescent lamp. Development of the fluorescent lamp (and other electric discharge lamps) has revolutionized the workplace. Buildings such as offices, shopping centers, and factories can operate during evening hours. The many new technological developments in the lighting industry offer the designer a variety of energy-efficient and environmentally responsible sources and controls to fully integrate daylight and electric light into the design process (FIG. 1).

1. BASIC CHARACTERISTICS OF LIGHT SOURCES

FIG. 1 (a) A daylit space; sunlight streams through a window at the Santa Anna Monastery in Santa Anna, Italy. (b) An electrically lit space-a wall sconce at the Westin Peachtree Plaza hotel in Atlanta, GA. Daylight and electric light sources are discussed in this Section. Daylight sources may be categorized as direct (direct sunlight or diffuse skylight) or indirect (light reflected or modified from its primary source). Electric light sources today fall generally into two generic classifications: incandescent lamps (including tungsten-halogen types); and gaseous discharge lamps (including fluorescent, mercury vapor, metal-halide, high-pressure and low-pres sure sodium lamps, and the induction lamp).

Efficacy is a basic characteristic common to both daylight and electric light sources-measured in lumens per watt (lm/W). Efficacy is the ratio of lumens provided to watts of heat produced by a light source. TABLE 1 lists efficacies of common light sources. Due to its high efficacy, daylight introduces less heat per lumen than electric sources, making use of daylight an attractive strategy for reducing cooling loads in buildings caused by lighting (assuming effective, balanced distribution and utilization of illumination).

TABLE 1 Efficacy of Various Light Sources

The efficiency of a standard incandescent lamp in converting electrical energy to light is approximately 7%; the other 93% is released as heat. Fluorescent lamps are approximately 22% efficient, and although they are a great improvement over incandescents, the generally low efficacy of lighting in buildings accounts for a large proportion of building energy use.

2. SELECTING AN APPROPRIATE LIGHT SOURCE

Choosing light sources for buildings-whether day light or electric light (or, more likely, a combination of both)-involves simultaneous lighting and thermal considerations. Because electric lighting in American nonresidential buildings consumes 25% to 60% of the electric energy utilized, any attempt to reduce this quantity must necessarily include integration of the cheapest (insofar as energy is concerned), most abundant, and, in many ways, most desirable form of lighting available-daylight.

In selecting appropriate light sources for buildings, understanding the characteristics of the light sources will allow a designer to use them appropriately for energy efficiency and to provide visual and thermal comfort. For resource efficiency, a designer should first optimize daylight sources through building geometries and material finishes, and then design the electric lighting system to supplement and enhance illumination and effect. This Section describes the characteristics of daylight and electric light sources to assist a designer in understanding the limits and capabilities of each source.

DAYLIGHT SOURCES

3. CHARACTERISTICS OF DAYLIGHT

The most prominent characteristic of daylight is its variability. The source of all daylight is the sun.

Exterior illumination, at a particular place and time, depends upon (1) solar position, which can be determined if the latitude, date, and time of day are given, (2) weather conditions (e.g., cloud cover, smog), and (3) effects of local terrain (natural and built obstructions and reflections). The position of the sun in the sky is expressed in terms of its altitude above the horizon and its azimuth angle. For all latitudes in the northern hemisphere, the sun's altitude is highest in summer, lowest in winter, and in between in spring and fall. Azimuth angle is defined as the sun's horizontal position angle, measured from the south. Solar position is absolutely predict able for any given time and location.

Cloud cover, unlike solar position, is only statistically predictable, on the basis of extensive U.S. Weather Service observations at numerous weather stations throughout the United States.

At locations other than those for which recorded data are available, an educated guess is necessary.

Outside the United States, a designer must rely on locally available data, which are often difficult to obtain.

The third factor-local terrain and construction conditions that either reduce illumination by shadowing or increase it by reflection-can be considered only on a case-by-case and site-specific basis.

For manual calculation procedures, it is sufficient to establish four basic sky conditions. These are:

1. Solid overcast sky
2. Clear sky without sun (in the field of view)
3. Clear sky with sun
4. Partly cloudy sky

4. STANDARD OVERCAST SKY

FIG. 2 (a) The completely overcast sky has a zenith luminance LZ, which is three times the horizon luminance. With such a sky, illuminance on unobstructed exterior horizontal surfaces (EH ) is about 2½ times that on similar vertical surfaces (EV ). (b) The clear sky has the area of brightest luminance around the sun. The area opposite the sun is darkest and can be considered as essentially uniform at approximately 3500 cd/ m^2 (1000 fL).

This condition, which occurs for much of the year in northerly climates such as England, Scandina via, and the Pacific Northwest, is called the CIE sky, because it was adopted by the Commission Internationale de l'Eclairage (CIE) as the standard design sky for daylighting calculations (CIE, 1970). This sky, as defined by the CIE, has a nonuniform brightness distribution, increasing from horizon to zenith in approximately a 1:3 ratio. Sky luminance at any altitude angle above the horizon is defined in Equation 1 as:

(eqn. 1)

where ... LA = luminance at Aº above the horizon (in any direction)

LZ = luminance at the zenith

Thus at the horizon, where A = 0º,

The illuminance (density of light in lux) on unobstructed exterior horizontal and vertical surfaces produced by this luminance distribution has an approximate ratio of 2.5:1 (FIG. 2). This results from an integration of Equation 1 over the whole sky.

There is agreement among all sources that with an overcast sky, exterior horizontal illuminance varies directly with the sun's altitude, irrespective of azimuth. Various formulations for this relationship have been put forward. One formulation that gives good agreement with observations is by Krochman (1963), shown in Equation 2:

EH = 300 + 21,000 sin A (eqn. 2)

where EH is exterior horizontal illuminance (lux) and A is the solar altitude, in degrees. Solar altitude (and azimuth) for various times of day can be obtained from Table D.1. Figure 3 is a plot of year-round averages for both vertical and horizontal illuminance from an overcast sky as a function of solar altitude, based on U.S. Weather Service observations.

It is interesting to compare the exterior horizontal illuminance obtained from the two sources given: Krochman's formula (Equation 2) and the observation-based data of FIG. 3 for a few typical conditions. Solar altitude is obtained from Table D.1.

Latitude: 38º Solar Time: 10:00 A.M.

Dates: Dec. 21, March/Sept. 21, June 21

The degree of agreement is generally satisfactory, and either source will yield suitable results.

FIG. 3 Curves giving unobstructed exterior surface illuminance directly from an overcast sky.

One of the most convenient ways of expressing the quantity of daylight illuminance during the schematic design of buildings is the concept of day light factor (primarily intended for overcast skies). Daylight factor is the ratio of indoor illuminance to available outdoor illuminance. Daylight factor is discussed in Section 14 as a means of setting criteria for, and determining the effectiveness of, a day lighting design.

5. CLEAR SKY

(a) Horizontal Illuminance

Exterior horizontal illuminance on a cloudless day consists of two source components: diffuse illumination from the entire sky plus the much larger component of direct sunlight. As with overcast sky, various empirical formulas for both components have been proposed, and here too, all sources agree that the total illumination, diffuse plus direct, varies directly with solar altitude.

FIG. 4 Components of the exterior horizontal illuminance on an unobstructed surface, from a clear sky, as a function of solar altitude. Total illuminance EH is the sum of the two components.

Figure 4 gives values for both components of exterior horizontal illuminance based upon observations. The sky only values are used to determine shaded skylight illuminance or daylong ground illuminance outside a shaded window- that is, a north-facing window, or an east/west window when the sun is on the opposite side of the building. In determining ground illuminance, the values given in FIG. 4 must be reduced some what, as they represent unobstructed horizontal illuminance, whereas the area outside a building window is partially obstructed from sky light by the building itself. If a building is so large that the ground outside the shaded window effectively receives diffuse radiation only from the half of the sky away from the sun, an average figure for EH of 1000 fc (~10,000 lux) can be used. This is because the luminance of the half of the sky away from the sun varies from a minimum of approximately 300 fL (1031 cd/ m^2) for the deep-blue patch directly opposite the sun to about 2000 fL (6874 cd/ m^2) at the sides, giving an average half-sky luminance of about 1000 fL (~3400 cd/ m^2). This, in turn, gives a horizontal illuminance EH, diffuse, of about 1000 fc (~10,000 lux) (see FIG. 2).

Figure 4 also gives horizontal illuminance from the sun only, as a function of solar altitude.

This value, when combined with the proper portion of diffuse illuminance, as discussed previously, is useful in determining ground illuminance out side a sunny building exposure or illuminance on an unshaded skylight. The light incident on an external reflector or light-shelf at a window can also be determined from these figures.

(b) Vertical Surface Illuminance

Inasmuch as most daylighting is accomplished via vertical fenestration, vertical surface illumination is the major component of interior daylight. It is also important for determining the daylight contribution of vertical elements in skylights. There is no simple relationship between horizontal and vertical illuminance from a clear sky, as there is for an overcast sky, because the illumination on a vertical surface depends upon solar azimuth as well as altitude. More specifically, it depends upon the bearing angle (FIG. 5), which is defined as the horizontal angle between a vertical plane containing the sun and a plane perpendicular to the vertical surface in question. A bearing angle of 0º indicates that the sun plane is perpendicular to the vertical surface. Like EH, EV (vertical illuminance) is divided into two components: sky only and direct sun only, which are plotted in FIG. 6 as a function of solar altitude and bearing angle. The sky only component is effectively for the half-sky because a vertical surface can be exposed to a maximum of only half of the full sky. Solar radiation data may be translated into illuminance by using average "efficiency" figures for solar energy in units of lumens per watt of received radiation (FIG. 6).

FIG. 5 The bearing angle of a vertical surface-the angle between a hypothetical vertical plane perpendicular to the surface (say, a window) and a hypothetical vertical plane containing the sun. (Other sources refer to this angle as the window-to-sun azimuth angle or surface azimuth.)

6. PARTLY CLOUDY SKY

The luminance of a partly cloudy sky cannot be expressed mathematically because of its infinite variability of conditions. However, statistical data on cloud cover are available from observations at many weather stations, and these data should be used in computer-calculated, hour-by-hour energy analysis programs. For the purpose of lighting design, it is important to note that the illumination from a partly cloudy sky is higher than that from a clear sky by 10% to 15% because of additional reflected sunlight from cloud edges. Several attempts have been made to account for this type of sky in terms of the effect on the daylight factor within a room, but none has received general acceptance.

FIG. 6 (a) Vertical surface illuminance, year-long average, sun only, no sky contribution. (b) Vertical surface illuminance, clear summer sky, no sky contribution. (c) Vertical surface illuminance, clear sky during various seasons, no sky contribution. (Libby-Owens-Ford.)

FIG. 7 Construction of a standard incandescent lamp.

ELECTRIC LIGHT SOURCES

Incandescent Lamps

7. THE INCANDESCENT FILAMENT LAMP

(a) Construction

The standard incandescent lamp consists simply of a tungsten filament inside a gas-filled, sealed glass envelope (FIG. 7). Current passing through the high-resistance filament heats it to incandescence, producing light. Gradual evaporation of the filament causes the familiar blackening of the bulb and eventual filament rupture and lamp failure. Incan descent lamps are available in many bulb and base types, with special designs for particular applications (FIG. 8 and Fig. J.1 in Appendix J). In order to dif fuse the light output, most bulbs are coated inside with white silica, providing almost complete light diffusion at a cost of approximately 2% to 3% of the light output. Colored light is also available from either coated bulbs or bulbs of colored glass.

The incandescent lamp base is the means by which a connection is made to the socket and thereby to the source of electric current. Most lamps are made with screw bases of various sizes, the most common being the medium screw base. General service lamps of 300 W and larger use the mogul screw base. When lamps are placed in precise reflectors or in lens systems where exact positioning of the filament is important, one of the special bases illustrated in FIG. 8 is used.

FIG. 8 Common incandescent lamp bulb and base types with nomenclature. The complete bulb nomenclature indicates type and size. The letter is an abbreviation of the shape, and the number is the diameter in eighths of an inch. A PS-52 is a pear-shaped bulb 52/8 (6½) in. (165 mm) in diameter. A PAR-38 is a parabolic reflector lamp 38/8 (4¾) in. (121 mm) in diameter.

(b) Operating Characteristics

Critically dependent upon the supplied voltage, the life, output, and efficiency of a lamp can be markedly altered by even a small change in operating voltage, as illustrated in Fig. J.2. For example, operating a 120-V lamp at 125 V or 115 V (TABLE 2) affects lumen output and, in particular, lamp life.

In installations where lamp replacement is difficult and/or expensive, and use of an incandescent lamp is indicated, lamps may be operated slightly under voltage to prolong life, thereby decreasing the frequency of replacement. Because luminous efficacy (the number of lumens emitted for each watt of electricity used) is decreased by this procedure, and recognizing that energy cost is normally a major consideration over the life of any lighting installation, a detailed life-cycle cost analysis should be made by the design professional involved.

TABLE 2 Comparison of Operating Characteristics

Conversely, where lamps are replaced before burn out using a group replacement system, and initial installation cost per lux and/or energy costs are high, lamps may be operated overvoltage, thereby increasing their output and efficacy but shortening their life. This procedure is normal in sports-lighting installations because of the high installation cost of tower-mounted floodlights, making it mandatory to extract the maximum light from each unit. In stadium installations with yearly lamp operation schedules averaging less than 200 hours, 10% overvoltage operation doubles the light output but still allows a once-a-year, off-season re-lamping and is therefore a highly economical procedure. Generally, however, it is advisable to operate incandescent lamps at the rated voltage, accepting balanced efficacy, output, and life.

(c) Other Characteristics

1. Lumen maintenance. Light output decreases slowly with lamp life as an incandescent bulb blackens. Lamp position (vertical or horizontal) during operation and the resulting bulb temperatures affect this characteristic.

2. Color. Incandescent light has a large yellow red component and is therefore highly flattering to the skin. The spectral content of the light produced by a heated source depends upon its temperature: high-wattage lamps are bluer, low-wattage lamps are yellower. Dimmed lamps give yellow-red light.

3. Surroundings. Generally, incandescent lamps are impervious to surrounding heat, cold, or humidity. Starting is completely unaffected by ambient conditions. Bulbs, however, must be appropriately selected if exposure to water is expected.

4. Luminous efficacy. Incandescent lamps produce light as a by-product of heat; as a result, they are inherently inefficient. Luminous efficacy increases with wattage. Thus, a 60-W general-service lamp produces 890 initial lumens, or 14.8 lm/W, whereas an A-21 100-W lamp produces only slightly less output than two 60-W lamps, but the higher wattage results in an 18% energy savings.

(d) Summary

The principal advantages of incandescent lamps are low cost; instant start and restart; simple, inexpensive dimming; simple, compact installation requiring no accessories; cheap fixtures; focusability as a point source; high power factor; lamp life independent of the number of starts; and skin-flattering, full-spectrum color. From a human factor perspective, the full-spectrum quality of light, with higher amounts of light in the red wavelengths, is best for rendering skin tones.

The principal disadvantages are low efficacy (see TABLE 1), short lamp life, and critical voltage sensitivity. Low efficacy means more fixtures and larger heat gain than more efficient alternatives. Short lamp life results in high lamp replacement labor costs. Voltage sensitivity may require careful and expensive circuit design. Also, light concentration at the filament (point source) requires careful fixture design in order to avoid glare and, if undesirable, sharp shadows. Because of its poor energy characteristics, incandescent lamp use should be limited to the following applications:

• Where use is infrequent

• Where there is frequent short-duration use

• Where low-cost dimming is required

• Where the point source characteristic of the lamp is important, as in focusing fixtures

• Where minimum initial cost is essential

• Where its characteristically good color rendering is desired

A brief list of conventional incandescent lamps and their general physical and operating characteristics is provided in Table J.1. Specific lamp data for use during design development should be taken from current manufacturers' literature.

8. SPECIAL INCANDESCENT LAMPS

Beyond the tungsten-halogen lamp, which is discussed separately, numerous special types of incandescent lamps are available. Some of the more important types are covered briefly in the following pages.

Rough service and vibration lamps are built to withstand rough handling and continuous vibration, respectively. Both conditions are extremely hard on general-service lamp filaments. Neither of these types is intended for general use, and both have lower luminous efficacy than a general service lamp (see Table J.1).

Extended-service lamps are designed for 2500 hour life. They are useful in locations where maintenance is irregular and/or relamping is difficult. Such lamps are really designed for slightly higher voltage than that which is applied, and therefore efficacy is reduced (see Table J.1 and Fig. J.2). So-called long life lamps, which are guaranteed to burn for 2, 3, or 5 years, are actually just lamps designed for higher voltages than those listed. As they normally sell at a high cost and are very inefficient, their use is seldom advisable. Before using such lamps, a life-cycle cost comparison, including the cost of lamps, energy, and relamping, should be made (see Appendix I).

(a) Reflector Lamps

These are made in "R," "BR," "ER," and "PAR" shapes (see FIG. 8). They contain a reflective coating on the inside of the glass envelope that gives the entire lamp accurate light-beam control.

Many reflector lamp types are available in narrow or wide beam design, commonly called spot and flood, respectively. R lamps are generally made in soft glass envelopes for indoor use, whereas PAR lamps are hard glass, suitable for exterior application. When using R and PAR lamps, the fixture acts principally as a lampholder since beam control is built into the lamp. These lamps have an improved reflector design that increases their efficiency.

The Energy Policy Act (EPACT) of 1992 made a number of incandescent reflector lamps obsolete because of the act's minimum efficacy requirements. These requirements state that R and PAR lamps rated 115 V to 130 V with a medium screw base, a bulb diameter greater than 2.75 in., and nominal wattages between 40 and 205 shall have minimum efficacies, as shown in TABLE 3. All major manufacturers produce incandescent reflector lamps that meet EPACT requirements. Among these are elliptical reflector (ER) and bulge reflector (BR) lamps that use a more efficient reflector design. These lamps are normally catalog listed as "energy-saving" lamps. (The Energy Policy Act of 2005 appears to have no provisions that substantially change the 1992 requirements.)

(b) Energy-Saving Lamps

Major national energy legislation including the National Appliance Energy Conservation Act of 1987, the Energy Policy Act of 1992 (and its 2005 update), and the comprehensive energy legislation passed by the U.S. House of Representatives and U.S. Senate in 2003 have created requirements to conserve lighting energy. The American Council for an Energy-Efficient Economy (ACEEE), the U.S. Department of Energy (DOE), and the Environ mental Protection Agency (EPA) have developed and support lighting energy efficiency programs, new research, and initiatives. For example, the Energy Star® program (energystar.gov), established by the EPA in 1992 to improve and provide energy-efficient products (appliances, lighting, and heating and cooling equipment) and practices, is aimed at reducing emissions from power plants, avoiding the need for new power plants, and reducing energy bills. Energy Star encourages every U.S. household to change the five fixtures used most often at home (or the lamps in them) to Energy Star-qualified lighting to save more than $60 every year in energy costs. If every household did this, it would keep more than 1 trillion pounds of greenhouse gases out of our air-a $6 billion energy savings equivalent to the annual output of more than 21 power plants.

TABLE 3 Minimum Required Efficacy of R and PAR Lamps

Every major manufacturer is producing a line of energy-saving lamps. Energy-efficient lamps are frequently known by trademarked names. They are rated at a wattage lower than that of the standard lamps they are intended to replace and are generally more efficient. Any additional first cost should be analyzed by a life-cycle cost analysis and the pay back period calculated. The designer will find that a proper control system is often more economically attractive than low-wattage lamps and that energy saving incandescent lamps are primarily useful in retrofit work.

FIG. 10 The self-regenerative halogen cycle slows the evaporation of the tungsten filament, and consequently lowers light depreciation and lengthens the lamp life compared to a standard incandescent lamp.

FIG. 9 Common tungsten-halogen lamps are available in a variety of designs.

9. TUNGSTEN-HALOGEN (QUARTZ-IODINE) LAMPS

This lamp type, illustrated in FIG. 9, is similar to the standard incandescent lamp in that it produces light by heating a filament. It differs in that a small amount of halogen gas (iodine or bromine) is added to the inert gas mixture that fills a small capsule constructed of quartz glass that surrounds the filament within the bulb of the lamp. This addition results in retardation of filament evaporation, which is the usual cause of incandescent lamp failure, and thereby extends lamp life (FIG. 10).

Although the tungsten-halogen lamp has only slightly higher efficacy than an equivalent standard incandescent lamp, it has the advantages of longer life, lower lumen depreciation (98% output at 90% life), and a smaller envelope for a given wattage (see FIG. 9). The last characteristic is due to the high temperature required by the halogen cycle, which in turn requires a compact, high-temperature filament. As a result, the lamp is effectively a point source, making it ideal for use in precision reflectors. Indeed, it is in this area, as discussed later, that most of the recent developments in tungsten-halogen lamp technology have occurred.

Due to the lamp's high filament temperature, the bulb envelope is generally made of quartz or a special high-temperature glass, which can with stand high temperatures better than glass; this, in turn, gave rise to the alternative name-quartz iodine-that is sometimes applied to this lamp type.

Another result of high filament temperature is that the gas pressure inside the quartz envelope is elevated, and the lamps have been known to rupture violently, spraying hot quartz fragments over a wide area. As a consequence, all manufacturers now provide a cautionary notice with their lamps.

The wording varies but essentially states that due to the possibility of rupture, lamps should be handled carefully, guarded against abrasion and overvoltage operation, and, most importantly, adequately shielded or screened. The shield can be a reflector cover, a fixture lens, a screen, or other devices that will contain hot flying fragments in the event that a lamp shatters or high temperatures cause a fire hazard.

A number of standards organizations and professional societies have adopted and published cautionary notices, including the IESNA Lighting Handbook, ANSI Standard C78.1451-2002, the Canadian Standards Association, and the International Electrotechnical Commission (IEC). Exceptions to these precautions exist when a lamp is protected by encapsulation inside a sealed envelope. Such encapsulated construction is now common in R, PAR, MR, and modified A-lamp shapes. These encapsulated lamps are intended for direct replacement of standard incandescent lamps of the same bulb shape.

Other halogen lamp characteristics are similar in all respects to those of the standard incandescent lamp. Color temperature ranges between 2000 K and 3400 K; spectral energy distribution is typical of blackbody radiation, and dimming characteristics are similar to those of standard incandescent lamps.

FIG. 11 Tungsten-halogen lamps can be mounted in a variety of ways inside an enclosing glass envelope. (a) Lamp is used either horizontally or vertically to the reflector or (b) used without a reflector in a protective glass envelope and an Edison medium screw base.

10. TUNGSTEN-HALOGEN LAMP TYPES

The basic lamp is a small gas-filled quartz tube, as shown in FIG. 9. Because the lamp must be used with some sort of reflector, it is manufactured with different terminations to suit the fixture reflector or secondary lamp envelope in which it is placed. All the lamp types shown in FIG. 9 can be used where lamp-only replacement is intended, such as in floodlights or reflector fixtures, by using appropriate bases-slide contacts for double-ended lamps, screw bases for screw-base lamps, and special ceramic pinhole bases for pin-type lamps.

(a) Encapsulated Lamps

These lamps (Figs. 11 and 12) are sealed units intended for direct replacement of either a corresponding distribution-type incandescent lamp in the case of reflector units or general-service incandescent lamps for reflectorless units. As a sealed unit, the halogen lamp is not replaceable, and the entire unit is discarded on burnout. Reflector units are available in a wide variety of beam patterns detailed in manufacturers' catalogs. Typical data are given in FIG. 13.

FIG. 12 PAR halogen lamp with standard medium screw bases in sizes PAR 16, 20, 30, and 38 with a short or elongated lamp neck. Lamp wattages range from 45 to 90 W. (Philips Lighting Co.)

FIG. 13 Typical PAR halogen lamp data. The beam of the PAR lamp is conical in shape. Each type of PAR lamp has a distinct illumination pattern (a-f) that varies in size and light intensity, depending on the angle at which the lamp is aimed and its distance from the area illuminated. (g) The round lighting pattern changes to oval or elliptical when the lamp is aimed at an angle, making illuminance calculations much more complex. (Philips Lighting Co.)

Reflector lamps are also available with a variety of filters: so-called cool lamps that direct much of the heat out through the back of the lamp; high-efficiency units that reflect and concentrate the heat back on the lamp filament; lamps with ultraviolet (UV) filters for use in displays of UV-sensitive objects; and others. For complete design information, access to current manufacturers' catalogs is a necessity.

FIG. 14 (a) Photo of a particular design of an MR-16-type lamp. (b) Typical illumination cones for the lamp shown in (a). Abbreviations: NSP, narrow spot; NFL, narrow flood; FL, flood; VWFL, very wide flood. (Osram-Sylvania Products, Inc.)

(b) MR-16 Precision Reflector Units

Miniature single-ended 12-V lamps of 2-in. diameter (and smaller), with multifaceted dichroic (heat-ejecting) reflectors and a bipin base, have found very wide acceptance in all types of display and accent lighting applications. These reflector units, illustrated in FIG. 14, essentially comprise an entire lighting fixture, like R and PAR lamps, requiring only a base for electrification.

The lamps are known by the generic name MR 16, after an early 2-in.-diameter model, although each major manufacturer utilizes its own trade name. A multi-mirror-faceted dichroic reflector produces a "cool" precision light beam by ejecting approximately two-thirds of the lamp heat (long wave radiation) through the back of the reflector. Luminaires must provide adequate means to dissipate this heat to avoid early lamp failure or creation of a fire hazard. The lamps are rated from 20 to 75 W. Beam characteristics of a few MR-16 lamps are given in FIG. 14.

Gaseous Discharge Lamps

Lamps in this category include fluorescent and high-intensity discharge (HID) lamps (mercury vapor, metal-halide, high-pressure sodium), which function by producing an ionized gas in a glass tube or container rather than heating a filament.

Discharge lamps are known for their long life and high efficacy. This section describes the function of ballasts and the various types of gaseous discharge lamps.

11. BALLASTS

All gaseous discharge lamps require a ballast to trigger the lamp with a high ignition voltage and to control the amount of electric current for proper operation. Ballasts discussed in this section primarily apply to fluorescent systems. Refer to manufacturers' information for details regarding HID ballasts. Matching of ballast to lamp is critical to successful lamp operation.

The function of a ballast is threefold:

• To supply controlled voltage to heat the lamp filaments in preheat and rapid-start circuits

• To supply sufficient voltage to start the lamp by striking an arc through the tube

• To limit the lamp current once the lamp is started Organizations involved with ballast standards and testing include:

ANSI-American National Standards Institute: originates standards on a national level CBM-Certified Ballast Manufacturers: a group of fluorescent ballast manufacturers who produce ballasts that conform to certain ANSI specifications ETL-Electrical Testing Laboratories, Inc.: a private, independent organization and recognized authority in measurement and testing of lamps and lighting equipment UL-Underwriters Laboratories, Inc.: an independent, nonprofit organization that certifies electrical products to ensure public safety from fire In the United States, ballasts should be UL labeled and CBM/ETL certified (for a limited number of fluorescent ballasts). The UL label ensures intrinsic safety. CBM establishes high-quality design criteria, and ETL tests ballasts to determine that design standards have been met.

(a) Ballast Characteristics

Because of the considerable energy that is lost in inefficient ballasts, manufacturers and standards organizations have established criteria by which ballast energy efficiency can be judged. These characteristics allow comparisons of lighting sys tem operation and performance parameters.

Ballast Factor. Ballast factor is the measured ability of a ballast to produce light from a connected lamp. It is the ratio of the light output of a lamp when operated on a tested ballast to the light output of the same lamp operated by a standard laboratory reference ballast (using ANSI test procedures). Ballasts with extremely high or low ballast factors can reduce lamp life because of inconsistencies in lamp current. ANSI Standard C82.11 prescribes a minimum ballast factor for CBM certification for a certain number of ballast types. A ballast may have different ballast factors for different lamps-for example, one ballast factor for operating standard lamps and another for operating energy-saving lamps. A lamp with a low ballast factor uses less energy, but light output is also less. A lamp with a high ballast factor uses more energy and provides more light output. Energy savings with high ballast factors may be achieved by using lower-wattage lamps and fewer fixtures.

Ballast factor is not a measure of energy efficiency.

Although a lower ballast factor reduces lamp lumen output, it also consumes proportionately less input power. Therefore, careful selection of a lamp-ballast system with a specific ballast factor will allow a designer to better minimize energy use by "tuning" the lighting levels in a space. For example, in new construction, high ballast factors are generally best, since fewer luminaires will be required to meet the light-level requirements. In retrofit applications or in areas with less critical visual tasks, such as aisles and hallways, lower ballast factors may be more appropriate.

Ballast Efficacy Factor. Ballast efficacy factor is the ratio of ballast factor (as a percentage) to power (in watts). The ballast efficacy factor is an expression of lumens per watt for a given lamp ballast combination. Comparisons using the bal last efficacy factor are valid only when comparing ballasts for equivalent systems in terms of lamp type and number of lamps.

Because ballast factor is an indication of the amount of light produced by a ballast-lamp combination and input watts is an indication of power consumed, the ballast efficacy factor is an expression of lumens per watt for a given lighting system. This measurement is generally used to compare the efficiency of various lighting systems. For example, a ballast with a ballast factor of 0.88 using 60 watts of input power has a bal last efficacy factor of 1.466 (where, 0.88 × 100 ÷ 60 = 1.466). Another ballast utilizing the same input power with a ballast factor of 0.82 has a ballast efficacy factor of 1.366. The first ballast therefore offers greater efficacy because it has a higher ballast efficacy factor (1.466 vs. 1.366).

Power Factor. Ballast power factor is a mea sure of how effectively a ballast converts the volt age and current supplied by a power source into watts of usable power delivered to the lamp. In general, the power factor is determined from the ballast design and is considered high (if above 0.90), low (below 0.79), or "corrected" (0.80 to 0.90). Power factor addresses the effective use of power supplied to a ballast, it does not relate to the luminous output of a ballast-lamp combination.

High power factor ballasts are more expensive, but the additional cost is readily repaid by lower line losses, smaller circuit conductors in long runs, and a larger number of fixtures per circuit. Energy conservation and economic considerations dictate the use of power factor-corrected and high power factor ballasts.

(b) Ballast Types

There are three basic types of ballasts: magnetic, hybrid, and electronic. Ballast technology has greatly changed in the past 10 years because of energy policy changes developed by the U.S. DOE, state energy offices, the ACEEE, the Alliance to Save Energy, the Natural Resources Defense Council, and lighting manufacturers.

Magnetic. Magnetic ballasts (core-and-coil) contain a magnetic core of several laminated steel plates wrapped with copper windings, and they operate at line frequency (60 Hz). These ballasts (FIG. 15) have become essentially obsolete, although they are found in existing buildings.

Hybrid. Also called cathode-disconnect bal lasts, hybrid ballasts use a magnetic core-and coil transformer and an electronic switch for the electrode-heating unit. Like magnetic ballasts, they operate at 60 Hz. The ballast disconnects the electrode-heating unit after starting the lamp.

Electronic. Solid-state electronic ballasts operate lamps at 20 to 60 kHz and have less power loss than magnetic ballasts. Lamp efficacy increases by approximately 10% to 15% compared to operation at 60 Hz. Electronic ballasts are lighter, more energy-efficient, generate less heat, and are virtually silent. They are also available as dimming ballasts, which allow light output to be controlled between 1% and 100%.

FIG. 15 Ballasts for fluorescent lamps have traditionally been of the electromagnetic type, operating at a frequency of 60 Hz.

Electronic ballasts operate at frequencies of 20,000-60,000 Hz and cause lighting systems to convert electric power to light more efficiently than systems run by electromagnetic ballasts.

Special. Lamps operating at other than 430 mA, including low-current and high-current units, require matching ballasts to supply the required current, waveform, and circuitry. Use of one manufacturer's low-current lamp with another's low-current ballast is not suggested without prior testing or a specific manufacturer's recommendation. The principal varieties of nonstandard ballasts are as follows:

1. Low-current ballasts are intended to match specific low-current lamps, including T8 tri-phosphor units, slimline lamps, and others.

2. High-current ballasts are intended to be used with high-output lamps. The purpose of this combination is either to increase output in an existing installation or to reduce the number of fixtures in a new installation.

3. Energy-saving ballasts are designed to reduce the total wattage of the lamp-ballast combination. Part of this power reduction is produced by more efficient design of the ballast itself.

Another part is due to a lower current rating for the lower wattage of the lamp itself. A third portion is frequently a switching arrangement in the ballast that disconnects the lamp filaments after an arc is struck (after the lamp ignites). This technique can save 4 to 8 W per lamp-ballast pair.

4. Multilevel ballasts are useful when it is desired to change lighting levels evenly. The usual unit is two-level-that is, full output and 50%-but three-level units are available for full, two thirds, and one-third output.

In addition to these ballast types, there are special units for low or high ambient temperature, weatherproof units, and low leakage-to-ground units for hospital applications.

(c) Ballast Performance Heat. Ballast heat is usually transferred to the luminaire body by direct metal-to-metal contact (which must be unimpeded) and is then dissipated by radiation and convection from the fixture. The location and method of fixture installation affect the heat transfer from the fixture and, consequently, the ballast temperature. Operating temperature directly affects ballast life. At normal operating temperature, a ballast life of 12 to 15 years can be expected. Generally, ballast life is halved for every 50ºF above the 194ºF (27.8ºC above 90ºC) operating temperature and, conversely, is doubled for every 50ºF (27.8ºC) reduction in operating temperature below 194ºF (90ºC). Electronic ballasts will usually start a lamp at 50ºF (24ºC) minimum.

A special ballast is required for starting at temperatures down to 0ºF (-18ºC). The cooler operation of electronic ballasts reduces air-conditioning costs.

Not only do the ballasts operate cooler, but lamps operated by electronic ballasts produce the same light output with lower losses. Therefore, overall energy costs for an electronic ballast installation are reduced because the fixtures use less energy and produce less heat for the same light output.

Noise. All electromagnetic and many electronic ballasts make a humming sound that originates from the inherent magnetic action causing vibrations in the steel laminations of the core-and coil assembly. Because electronic ballasts have a small (or no) core-and-coil assembly, they have the lowest noise output. Most electronic ballasts make almost no sound. Ballast noise, if any, may become amplified because (1) of the method of mounting the ballast in the fixture, (2) of loose parts in the fix ture, or (3) ceilings, walls, floors, and hard furniture reflect the noise. Ballasts are sound-rated by a letter, A through F, which indicates not actual sound developed, but performance in a space. A rating of A designates the quietest ballast. Selection should be made on the basis of the ballast sound rating and the requirements of the installation.

Flicker. Flicker is caused by extinguishment and re-ignition of the arc within a fluorescent tube and is visible only when a lamp is operated at a relatively low frequency (such as the 60 Hz typical of North American electrical systems) and where long-persistence phosphors are thin or entirely absent (i.e., at the extreme ends of a lamp). A magnetic ballast does not alter the incoming line frequency. Thus, the lamp voltage crosses zero 120 times each second, resulting in 120 light output oscillations per second. This flicker is typically not noticeable to most viewers-but there is evidence that flicker of this magnitude can cause adverse effects, such as eyestrain and headache. Most electronic ballasts operate at higher-than-line frequency, which reduces lamp flicker to an essentially imperceptible level. Manufacturers can specify the flicker percentage of a particular ballast working in conjunction with a given lamp type and phosphor composition. For a standard phosphor lamp operated at 60 Hz (magnetic ballast), the flicker percentage is about 30%; with an electronic ballast (high frequency) the flicker percentage is nil. Dimming Control. Dimming of electronically ballasted lamps is accomplished within the ballast itself. The dimming process uses energy that should be accounted for in lighting system energy-use calculations. Electronic ballasts alter the output power to the lamps by a low-voltage signal into the output circuit. High-power switching devices to condition the input power are not required. This arrangement allows control of one or more ballasts independent of the electrical distribution system.

Radio Noise. Occasionally, a defective bal last will cause radio noise (commonly referred to as radio frequency interference [RFI]). In general, however, RFI is not produced by a ballast, but by the arc discharge in a fluorescent tube. To minimize RFI, ballasts are available with integral RF noise suppressors. In extreme cases, additional suppression can be obtained by installation of RF noise attenuators in a lighting fixture.

Fluorescent Lamps

The second major category of electric light sources is gaseous discharge lamps, of which the fluorescent lamp is the best-known and most widely used type. Since their introduction in 1937, fluorescent lamps have almost completely supplanted incandescent lamps in all fields except specialty lighting and residential use. The typical linear fluorescent lamp comprises a cylindrical glass tube sealed at both ends and containing a mixture of an inert gas, generally argon, and low-pressure mercury vapor.

Built into each end of the tube is a cathode that supplies the electrons to start and maintain an electric arc, or gaseous discharge. Short-wave UV radiation, which is produced by the mercury arc, is absorbed by phosphors coating the inside of the tube, causing a reaction that emits visible radiation (light). The particular mixture of phosphors used governs the quantity and spectral quality of the light out put. Light from fluorescent sources radiates from a larger lamp surface area than is the case with incandescent sources. The light is diffuse, which is suitable for illuminating or washing large areas such as ceiling planes.

12. FLUORESCENT LAMP CONSTRUCTION

Rapid-start and instant-start fluorescent lamps are commonly used today. Preheat lamps are a legacy type. Lamp families include linear and compact.

Linear lamps are tubular in shape, with the most popular versions being T8 and T5 (26-mm and 16-mm) in standard and high output (HO) and the legacy T12 (38-mm) lamp. Compact fluorescents include dedicated socket versions of single-tube, double-tube, and triple-tube lamps. The descriptions that follow cover standard lamps and circuits.

Special lamps, accessories, and circuits, including low-wattage lamps, triphosphor lamps, and special shape lamps, are discussed separately.

(a) Preheat Lamps

Older fluorescent fixtures use a preheat technology that heats the gas in order to start the lamp and use a mechanism called a starter. Preheat fixtures either have an automatic starter or require a manual starting action. The original T12 (38-mm) fluorescent lamp was a preheat design.

Construction of a typical hot cathode lamp (used with both preheat and rapid-start types) is shown in FIG. 16. All preheat lamps have bi-pin bases (see FIG. 16).

This lamp circuit utilizes a separate starter, a small cylindrical device that plugs into a preheat fixture. When the lamp circuit is closed, the starter energizes the cathodes; after a 2- to 5-second delay, it initiates a high-voltage arc across the lamp, causing it to start. Most starters are automatic, although in desk lamps preheating is accomplished by depressing the start button for a few seconds and then releasing it. This closes the circuit and allows the heating current to flow; releasing the button causes the arc to strike. Preheat lamps are no longer the industry standard but are included here as a point of comparison.

FIG. 16 Details of typical fluorescent lamps and associated lampholders. (a) Construction of preheat-rapid-start bi-pin base lamp.

This type of lamp has type (b) base and is held in type (c) lampholder. High-output HO and VHO rapid start lamps use a recessed double contact base (d) and lampholders (e). Instant-start lamps are similar in construction to (a) except with cathode construction (f), have a single-pin base (g), and use single-pin lampholders (h), which are different at each end.

FIG. 17 Rapid-start fluorescent lamps have two pins that slide against two contact points in an electrical circuit. The ballast constantly channels current through both electrodes, creating a charge difference between the two electrodes and establishing a voltage across the tube.

(b) Rapid-Start Lamps

Today, the most popular fluorescent lamp design is the rapid-start lamp, shown in FIG. 17. This design functions similarly to the traditional pre heat lamp, but without a starter switch. Instead, the lamp's ballast constantly channels current through both electrodes, eliminating the delay inherent in a preheat circuit. This current flow is configured so that there is a charge difference between the two electrodes, establishing a volt age across the tube. Most fluorescent fixtures with two or more lamps are known as rapid start. When the lamp circuit is energized, the arc is struck immediately. No external starter is required.

Because of the similarity of operation, rapid-start lamps will operate satisfactorily in a preheat circuit. The reverse is not true, because the preheat lamp requires more current to heat the cathode than the rapid-start ballast provides.

Most rapid-start T12 lamps operate at 430 mA. If the current is increased, the output of the lamp also increases. Two generic types of higher output rapid-start lamps are available. One operates at 800 mA and is called simply high output (HO). The second, which operates at 1500 mA (1.5 A), is called (by different manufacturers) very high output (VHO), super-high output, or simply the 1500-mA, rapid-start lamp. There is also a 1500 mA lamp that uses what looks like a dented or grooved glass tube. This lamp has a somewhat higher output than a standard VHO tube. All HO lamps use double-contact bases and special bal lasts (see FIG. 16). HO lamps are used in applications such as outdoor sign lighting, street lighting, and merchandise displays where high output is required from a limited-size source. Because of the serious heat problems involved, VHO lamps are frequently operated without enclosing fixtures. Conversely, HO and VHO lamps are frequently used in cold environments that prevent proper operation of standard output 430-mA lamps. Most HO and VHO lamps have slightly lower luminous efficacy than a standard 430-mA, rapid-start lamp and have a considerably shorter life. It should be noted that only rapid-start lamps are to be used with motion sensors or in conjunction with sequential repetitive dimming of lamps. Use of instant-start lamps in this application will overload the cathodes, and lamp life will be reduced.

(c) Instant-Start Fluorescent Lamps

Instant-start fluorescent lamps use a high-voltage transformer to apply a very high initial voltage to the cathodes. An excess of electrons on the cathode surface forces some electrons into the fill gas, which ionizes the gas. This creates an instant voltage difference between the cathodes, establishing an electric arc. These lamps have only a single pin at each end that also acts as a switch to break the ballast circuit when the lamp is removed, thus lessening the shock hazard (see Figs. 13 and 16). The lamps are generally operated in two-lamp circuits at various currents; normal currents are 200 and 430 mA. The high voltage starting characteristic of instant-start circuits lowers the lamp life to about half that of a corresponding rapid-start lamp. Instant-starts have the advantage of being able to start at much lower ambient temperatures (below 50ºF [10ºC]) than rapid-start circuits. This starting characteristic makes the instant-start lamp and circuit particularly applicable to outdoor use.

13. FLUORESCENT LAMP LABELS

Standard fluorescent lamp labels are printed on the end of a lamp and are identified by several letters and numbers, as shown in TABLE 4.

The typical labeling is in the form FSWWCCC TDD (each manufacturer has variations on this format). Depending upon the type of fluorescent lamp, designations for color rendering index and color temperatures are also included on the label.

TABLE 4 Fluorescent Lamp Label Designations

TABLE 5 Comparative Characteristics of Tubular Fluorescent Lamps

14. FLUORESCENT LAMP TYPES

T8 Fluorescent Lamps. Over the past 20 years, T8 lamps have afforded designers (and their clients)

cost-effective and energy-efficient lighting systems that are visually comfortable and have a high degree of flexibility in their application. T8 lamps are 1 in. (25 mm) in diameter; they are available in wattages of 17, 25, 32, and 40 W at 2-ft, 3-ft, 4-ft, and 5 ft (600-mm, 900-mm, 1200-mm, and 1500-mm)

lengths, respectively. There are two color rendering categories of T8 lamps: 700 (75 CRI) and 800 (85 CRI) series, which relate to the color rendering properties of the tri-phosphor coatings used. Lamp manufacturers have developed a standard designation to indicate the color temperature of a lamp. For example, "30" indicates a 3000 K lamp. T8 lamps are available in 3000 K, 3500 K, and 4100 K color temperatures, designated "30," "35," and "41," respectively. TABLE 5 shows a comparison of lamp technologies as fluorescent lamp manufacturers have increased efficacy and color rendering while decreasing diameter and wattage.

T5 Fluorescent Lamps. A new line of T5 lamp technology was developed in Europe and introduced in North America in 1996. Although it was still expensive, the introduction of a T5 HO line in 1998 offered about twice the lumen output in the same length as its T8 counterpart, with an efficacy that is attractive in meeting project energy goals. The T5 is the first "metric" lamp introduced in the United States, yet it is commonly called T5 because of industry nomenclature. Standard and HO T5 lamps are available in 22-in., 34-in., 46-in., and 58-in. (560-mm, 864-mm, 1163-mm, and 1473-mm) lengths. The standard T5 and the T5 HO lamps are the same diameter and width. The 46-in. T5 (nominal 4 ft) is rated at 2900 lumens, similar to the lumen per watt output of a T8 lamp (2950 lumens). By contrast, the 46-in. T5 HO lamp is rated as high as 5000 lumens, offering twice the maintained light output of a T8 lamp. Because of its smaller 58-in. (15-mm) diameter construction, significantly less glass, mercury, and high quality phosphors are needed for its construction.

T5 lamps also allow a designer to use fewer lamps (and fixtures), thus providing certain savings on installation and long-term maintenance. The narrow lamp diameter has provided an opportunity for the design of new fixtures and for use in low-profile, indirect luminaires. The color rendering quality of light from T5 lamps (CRI 85) is excellent, although the potential for glare problems exists, which can be addressed by sophisticated shielding techniques.

Utilizing T5 (and particularly T5 HO) lamps in direct delivery lighting installations requires special attention to glare control.

15. CHARACTERISTICS OF FLUORESCENT LAMP OPERATION

Five characteristics define the operation of fluorescent lamps:

• Efficacy-light output per unit of power input

• Lumen maintenance-the decreasing output of light as a lamp ages

• Lamp life-average (statistically defined) lamp life expectancy

• Temperature and humidity-how a lamp responds to extreme environmental operating conditions.

• Dimming-output reduction of a fluorescent lamp

(a) Efficacy

luminous efficacy lumens (light output) w =

a atts (power consumed including ballast loss ses)

= lumens per watt (lm/W)

The design efficacy (lumens per watt) of a fluorescent lamp depends upon the operating current and the phosphors utilized. Fluorescent lamp efficacy is further dependent upon the lamp length, ambient temperature, frequency of the electricity supply, and ballast operation. Wattage, in itself, is a meaningless quantity unless it is associated with a lumen output figure. Thus, energy-saving low-wattage or high-output lamps with their special ballasts are seldom the indicated choice in new design because their efficacy does not usually justify the cost premium. These special lamps are useful in retrofit work, in which case field measurements of illuminance and lamp temperature are required before selecting a replacement lamp- ballast combination.

FIG. 18 Lumen maintenance curve for fluorescent lamps.

Lumen maintenance is the ability of a lamp to retain its lumen output over time. Greater lumen maintenance means that a lamp will remain brighter longer. The opposite of lumen maintenance is lumen depreciation, which represents the reduction of lumen output over time. In the United States, mean lumens is a measure taken at 40% of the rated lamp life. In the United Kingdom, lighting levels are based upon maintained illumination, and it is necessary to determine the minimum lighting level in the installation when replacement of lamps is due, taking into account all possible reasons for deterioration.

(b) Lumen Maintenance

The lumen output of a fluorescent tube decreases rapidly during the first 100 hours of burning, and thereafter much more slowly. Phosphors deteriorate, typically blackening at the ends of a lamp, thereby blocking some light. Most product catalogs list "initial lumens," which is the lamp out put under laboratory conditions after 100 hours of burning, and "mean" or "design" lumens, which is lamp output at 40% of life. Lighting levels gradually drop as a system ages until, somewhere in the middle of the effective life of the lamps (FIG. 18), the intended "maintained" illuminance of a system is (temporarily) achieved.

(c) Lamp Life

The life of a standard fluorescent lamp is defined as the period of time an average lamp is expected to last, depending upon the burning hours per start. It is expressed as "rated average life" in hours of operation. The values listed in lamp catalogs for life are based upon a burning cycle of 3 hours per start (and 20 minutes of "off" status) and represent the aver age life of a group of lamps; that is, half of the lamps in any group will have burned out at that time. Typical lamp mortality curves are shown in FIG. 19, and the effect of burning hours per start is shown in FIG. 20. Average rated lamp life is not the same as the time at which lamps are typically replaced, which is usually well before 50% failures occur in a batch. Several factors affect fluorescent lamp life.

Longer burning hours per start will extend lamp life. Lamp life is shortened by improper lamp cur rent, improper voltage to the ballast, or improper cathode heating.

From an energy utilization viewpoint, if a lighted space is not utilized for 10 minutes or more, fluorescent lamps should be shut off. This takes into account both direct energy consumption and the resource energy required to replace a lamp as a result of shortening its life. From a cost viewpoint, the break-even point depends upon these factors: (1) lamp life reduction as a function of burning hours per cycle, (2) cost of energy, (3) cost of lamp and lamp replacement, (4) amount of time the lamp remains off when shut off, (5) cost of switching equipment (if any), and (6) life of the building.

With this number of variables, it is not possible to give general solutions, and an individual analysis is required. However, several analyses for ordinary office conditions, using lamp life data as given in FIG. 20 (a 20-year fixture life, $0.085/kWh energy cost escalating 3% annually, $1.25 lamp cost, 15-minute relamping time, and $8 per lamp to provide the necessary switch [one switch per two 2-lamp fixtures]), have shown that lamps should be switched off any time they are not in use for 5 to 8 minutes or more. (The spread is caused primarily by variation in local labor rates.) It is thus clearly an economic fallacy to leave lamps burning to achieve longer lamp life.

FIG. 19 Typical mortality curves of standard fluorescent lamps.

FIG. 20 Effect of burning hours on fluorescent lamp life. Note that at 3 burning hours per start, the average lamp life is 100% of the nominal catalog figure.

(d) Effect of Temperature and Humidity

Fluorescent lamps are affected by extremes in ambient temperature and by high humidity. Outside of the optimal operating temperature range, 41 to 77ºF (5 to 25ºC), there is a rapid drop in light output and difficulty in starting. High humidity causes electrical leakage along the lamp surface, lowering the starting voltage provided by the ballast. Lamps are pre-coated with silicone to break up moisture films and prevent such leakage.

The temperature of the coolest point on the lamp bulb wall determines a lamp's mercury-vapor pressure, which in turn determines the lamp lumen output, wattage, and color. Maximum output for standard lamps occurs at a bulb temperature of 104ºF (40ºC). The bulb wall temperature itself is affected by room ambient temperature, airflow over the lamp (as with air-return fixtures), and the temperature of adjacent surfaces such as a ballast enclosure. Thus, catalog data on lamp output and wattage, based upon laboratory tests of bare tubes at 77ºF (25ºC) ambient temperature in still air, may be very far from actual field performance.

(e) Dimming

Dimming of a fluorescent lamp system reduces energy consumption, can correct overlighting, can balance illumination through integration with daylighting, and allows flexibility when full lighting output is required. Unlike incandescents, which can be dimmed with just a wallbox device, fluorescent lamps require dimming ballasts. The dimming range differs greatly among ballasts. With most electronic dimming ballasts, output can vary between full and a minimum of about 10% of full output. However, electronic, full-range dimming ballasts are also available for some lamp types that operate lamps down to 1% of full lumen output.

A ballast can be configured so that it (1) receives a signal from a control device and subsequently (2) changes the current flowing through a lamp, thereby achieving a gradual, controlled reduction in lamp output. The characteristics of the ballast circuitry affect the duration and extent of the change in current and subsequent lamp output.

Electronic dimming ballasts for fluorescent lamps are designed to respond to either an analog or digital signal to achieve the dimming effect.

Because a dimming ballast must be able to communicate with connected controllers, the method becomes the basis for a protocol (common operating parameters adopted by all manufacturers of dimming bal lasts and controllers that use that method). This ensures interchangeability between a ballast made by a particular manufacturer and various controllers made by controls manufacturers. Typical applications for dimming include both new construction and retrofit installations: auditoriums and training areas, conference rooms and boardrooms, department and specialty stores, education and health care institutions, hotels, houses of worship, private and executive offices, and restaurants.

The primary dimming methods are:

• Analog: An analog electronic dimming bal last includes components that perform these functions: electromagnetic interference filtering, rectification, power factor correction, and ballast output to power a lamp. There are several analog methods, including 0-10VDC, two-wire phase control, three-wire phase-control, and wireless infrared, with 0-10VDC being most often used.

• Digital: The digital electronic dimming bal last includes components that perform these functions: electromagnetic interference filtering, rectification, power factor correction, ballast output to power a lamp, and control (as a micro controller). The micro-controller functions as a storage, receiver, and sender of digital information. The micro-controller can store the ballast address, receive control signals, and send status information.

• Wireless infrared: This method uses an infrared transmitter to control the signal and does not require additional wires. The dimmer is either contained in the ballast or provided as an additional component in the light fixture. Wireless infrared control is a good retrofit solution and allows for occupant fixture control. Wireless infrared control is ideally suited for spaces where individual control is desired without additional wiring, such as conference rooms, boardrooms, and open and private offices.

16. FEDERAL STANDARDS FOR FLUORESCENT LAMPS

Enacted in October 1992, the National Energy Policy Act (EPACT) was designed to reduce the U.S. energy bill by approximately $250 billion over a 15-year period. EPACT 1992 and its successor EPACT 2005 mandate minimum standards for lamps in terms of efficacy (lumens per watt) and color rendering index (CRI; the ability of a light source to render colors accurately). EPACT standards eliminated the manufacture and distribution of several major fluorescent lamp types and incandescent lamps that provided the least amount of light for the highest use of energy. The major lamps eliminated are 40 W F40T12 (CW and WW); 75 W F96T12 (CW and WW), and 110 W F96T12/HO (CW and WW). Lamps with very good CRI and special-service fluorescent lamps are excluded from the act.

17. SPECIAL FLUORESCENT LAMPS

(a) Low-Energy Lamps

The need for energy conservation, the discontinuance of certain lamp types due to EPACT regulations, and a desire to reduce lighting levels in existing over-lighted spaces have resulted in the development of a complete line of low-energy lamps.

Wattage ratings for these lamps are lower than those of standard lamps because they are intended primarily as lower-energy direct replacements for existing lamps. All such lamps are clearly marked by the manufacturer. They require special matching ballasts for maximum effectiveness and have an efficacy equal to, or somewhat higher than, that of standard lamps and ballasts. They have the disadvantages of higher cost, the need for special ballasts where maximum energy reduction is desired, generally shorter life, inability to be used in most dimming circuits, and problems with inventory and proper lamp replacement. Their use is indicated only where other light-output and wattage reduction schemes, such as circuit dimming or reduced-wattage ballasts, are inapplicable.

(b) U-Shaped Lamps

U-shaped lamps were developed to answer a need for a high-efficacy fluorescent source that could be utilized in a square fixture. The U lamp is basically a standard fluorescent tube bent into a U shape and available with 358- or 6-in. (92- or 152 mm-) leg spacing; the former can be accommodated three to a 2-ft × 2-ft (610 mm square) fixture. (The narrower T8 envelope of tri-phosphor lamps permits a tighter bend; these U lamps have a 158-in. [41 mm] leg-to-leg spacing.) U lamps operate on standard ballasts and have slightly lower output than a corresponding straight tube. In all other respects, a U lamp has the same characteristics as a straight lamp of similar type.

(c) Ecologically Friendly Lamps

ALTO® lamps were developed in 1995 by the Philips Lighting Company to support the reduction of mercury at the source and provide users with environmentally responsible methods for disposal. The ALTO family of lamps includes a broad selection of TCLP-compliant lamps: linear and compact fluorescents, high-pressure sodium lamps, metal-halide, U-bent fluorescent, and the lead-free MasterLine ALTO lamps, all of which can be recycled (always the preferred method) or disposed of convention ally. TCLP is the Toxicity Characteristic Leaching Procedure--a test developed by the EPA in 1990 to measure hazardous substances that might dissolve into the ecosystem and that is used by the federal government and by most states to determine whether old fluorescent lamps should be characterized as hazardous waste. All fluorescent lamps contain mercury; however, ALTO lamps have the lowest mercury doses available (on average, 70% less mercury than the 2001 industry average) on the market. This product development encouraged other manufacturers to reduce the mercury con tent in their products-as in the Osram Sylvania (Ecological ) and General Electric (Ecolux) lines.

(d) UV Lamps

UV lamps emit radiation in the UV spectrum, which includes all electromagnetic radiation with wave lengths in the range of 10 to 400 nanometers (nm).

• The UVA range includes wavelengths from 315 to 400 nm. Wavelengths from about 345 to 400 nm are used for "blacklight" effects (causing many fluorescent objects to glow) and are usually slightly visible if isolated from the more visible wavelengths. Shorter UVA wavelengths from 315 to 345 nm are used for sunning.

• UVB refers to wavelengths from 280 to 315 nm.

These wavelengths are more hazardous than UVA wavelengths and are largely responsible for sunburn.

• UVC refers to shorter UV wavelengths, usually from 200 to 280 nm. Wavelengths in this range, especially from the low 200s to about 275 nm, are especially damaging to exposed biological cells. Such short-wave UV radiation is often used for germ killing purposes.

Although UV lamps are not commonly used in architectural lighting, they are included in this section as specialty lamps because they address a wide range of applications in industrial, technological, laboratory, and medical settings. UV lamps include fluorescent black lights, fluorescent tanning and medical UV lamps, "RS" reflector ("floodlamp") sunlamps, and germicidal and EPROM (erasable programmable read-only memory) erasing lamps.

FIG. 21 Family portrait of compact fluorescent lamp designs. Pin-base lamps designed for use with a separable ballast are shown in the foreground, the same lamps mounted in their screw-base ballasts are in the left background, and the one-piece combined lamp-ballast design stands in the right background. Globe and reflector-type replacements for incandescent lamps are in the center of the photo. (Osram-Sylvania.)

TABLE 6 Equivalent Wattage of Common Incandescent Lamps and Compact Fluorescents

TABLE 7 Cost Comparison for Operation of an Incandescent Lamp and a Compact Fluorescent Lamp

18. COMPACT FLUORESCENT LAMPS

Compact fluorescent lamps (CFLs) offer a comparable (in brightness and color rendition), energy-efficient alternative to incandescent lamps. Unlike standard fluorescent lamps, they can directly replace standard incandescent bulbs.

CFLs are simply folded fluorescent tubes with both ends terminating in a common base. Some com pact fluorescent lamps have the tubes and ballast permanently connected with a screw-in medium base.

Others have separate tubes and ballasts, allowing the tubes to be replaced without changing the ballast. As a result, an exhausted lamp is simply replaced while retaining the existing ballast, resulting in considerable economy. A CFL produces a diffuse light, unlike single-point incandescent lamps. This is an important factor to consider when replacing incandescents with CFLs in high ceiling applications.

CFLs are manufactured in a variety of styles or shapes: two, four, or six tubes; circular or spiral tubes (FIG. 21). They are efficient at lower watt ages and can produce light output equivalent to that of higher-wattage incandescents (e.g., a typical 60-W incandescent lamp with a 900-lumen output could be replaced by a 15- to 19-W CFL). The total surface area of the tube(s) determines how much light is produced. The efficacy of lamp-ballast combinations ranges from 55 to 75 lm/W, assuming an electronic ballast. Lamps with magnetic ballasts are available but are not favored because of excessive heat, weight, flicker, and reduced efficiency. Lamp colors are similar to straight lamps (i.e., 3000 K, 3500 K, 4200 K, and 5000 K). All CFLs have a CRI of 80 or higher. Their life is 10,000 to 12,000 hours based on 3 hours per start. TABLE 6 compares the wattage of commonly available incandescent lamps to that of a CFL that provides similar light output.

A major advantage of using CFLs is saving money, as shown in TABLE 7. This table assumes that a lamp is on for 6 hours per day and that the electric rate is 8 cents per kilowatt-hour.

HIGH-INTENSITY DISCHARGE LAMPS

High-intensity discharge (HID) lamps (FIG. 22) produce light by discharging electricity through a high-pressure vapor. Lamps in this category include mercury-vapor (CRI range 15 to 55), metal-halide (CRI range 65-80), and high-pressure sodium (CRI range 22 to 75). These lamps are characterized by high efficacy, rapid warm-up time, rapid restrike time, and historically poor color rendering capabilities. Mercury-vapor lamps were the first commercially available HID lamps and originally produced a bluish-green light. Today they are available with a color-corrected whiter light, but because of inefficiency and potential hazards they are being replaced by the newer, more efficient metal-halide and high-pressure sodium lamps. Standard high pressure sodium lamps have the highest efficacy of all HID lamps, but they produce a yellowish light.

High-pressure sodium lamps that produce a whiter light are now available, but their efficiency is some what lower.

HID lamps are typically used when high illuminance is required over large areas and when energy efficiency and/or long life are desired. Typical applications include gymnasiums, large public areas, warehouses, outdoor activity areas, roadways, parking lots, and pathways.

FIG. 22 Bulb shapes for most HID lamps, with their maximum overall length (M.O.L.).

FIG. 23 Typical construction of a clear mercury-vapor lamp.

19. MERCURY-VAPOR LAMPS

The mercury-vapor lamp was the first HID lamp to be developed, and for many years it was the only HID lamp commercially available. It has been largely supplanted by the metal-halide lamp because of the latter's better color and efficacy.

Though many installations of mercury vapor exist, this technology is no longer specified for new buildings. For the same energy efficiency, metal-halide lamps have better color rendering properties and are now often preferred for indoor applications.

Mercury-vapor lamps, most often used to light streets, gymnasiums, and sports arenas, must be maintained properly to be safe.

The mercury-vapor lamp operates by passing an arc through high-pressure mercury vapor contained in a quartz arc tube (FIG. 23). This produces radiation in both the UV region (as in the low-pressure fluorescent lamp tube) and the visible region, principally in the blue-green band. This color is characteristic of a clear mercury lamp.

(a) UV Radiation

A considerable portion of a mercury lamp's energy spectrum is in the UV range. This does not normally constitute a hazard to persons exposed to even a clear lamp because the outer glass bulb absorbs most of the UV radiation while transmitting light.

However, if the outer bulb is broken, the quartz arc tube will continue to burn, and the UV radiation emitted constitutes a safety hazard, particularly to the skin and eyes of persons exposed to it. As a result, manufacturers include a warning to this effect with all mercury lamps sold. Users are also generally informed that a safety-type mercury lamp is readily available that will self-extinguish if the outer glass envelope is broken. An alternative is to use mercury-vapor lamps in an enclosing fixture designed to both prevent lamp breakage from external sources, such as vandalism, and protect users of the lamp in the (unlikely) event of a spontaneous lamp fracture. In the interest of safety, it is suggested that mercury lamps be shut off at least once a week for at least 30 minutes to allow them to cool completely.

In connection with protection from the injurious effects of UV radiation, it is well to note two facts:

1. The shorter the UV wavelength, the more potentially irritating it is to the skin and eyes.

Germicidal UV radiation is in the short-wave range (200 to 300 nm).

2. White plaster and polished metal are good reflectors of UV radiation. As a result, UV radiation reflected from such surfaces is almost as dangerous as that from line-of-sight exposure to a UV source.

TABLE 8 Typical Data for Mercury-Vapor Lamps

(b) Lamp Life

Lamp life is extremely long, averaging 24,000 hours or more based upon 10 burning hours per start. Mercury-vapor lamps are not suitable for applications that are subject to constant switching.

Their life is affected by ambient temperature, line voltage, and ballast design.

(c) Lumen Maintenance

This depends upon the specific type of lamp and its burning position. Manufacturers publish data for each of their lamp types. In general, clear lamps have the best lumen maintenance, followed by color-improved and phosphor-coated units.

(d) Color Correction and Efficacy

Color correction is normally required because the blue-green light from a clear lamp distorts almost all object colors. (Mercury-vapor lamps are frequently used to illuminate outdoor gardens because the blue-green light enhances the green of trees and vegetation.) Color correction is achieved by adding phosphor to the inside of the outer bulb. The phosphors convert UV radiation to light exactly as in fluorescent lamps, and the stain on the glass acts as a filter to some of the blue-green radiation. The phosphors reradiate generally in the red band, which is entirely absent in the basic lamp color. Depending upon the arc tube design and the phosphors used, the color of the emitted light can be corrected to make it acceptable for general indoor use. Lamps are available in clear, white, color-corrected, and deluxe white in ascending order of color improvement. Efficacy, including ballast loss, ranges from 25 lm/W for a 50-W lamp to a maximum of 55 lm/W for a 1000-W color corrected lamp. Note that, in general, efficacy is lower than for fluorescent lamps. CRI ranges from a low of 20 for a clear lamp to a high of only 50 for a deluxe white lamp. A short list of representative lamp data is given in TABLE 8.

(e) Ballasts and Lamp Starting

Ballasts are required, as with all arc discharge lamps, to start a mercury-vapor lamp and there after to control the arc. From 3 to 6 minutes are required for the lamp to reach full output because heat must be generated by electron flow to vaporize the mercury in the arc tube before the arc will strike. Once extinguished, the lamp must cool before restrike is possible. This restart delay amounts to 3 to 8 minutes, depending upon the ballast type, and is an important consideration in design, as a momentary power outage will extinguish all lamps, leaving an interior area in the dark. Mercury-vapor luminaires are available that utilize small halogen lamps to supply light during such outages. Alternately, some incandescent lighting can be utilized to maintain minimum illumination.

The principal mercury-vapor ballast types are reactor, regulating, and electronic. Magnetic mercury ballasts are large, heavy, and quite noisy.

Where this may be a problem, remote ballast mounting should be considered or lighter, quieter, and more expensive electronic ballasts used.

Because lamp-operating characteristics depend heavily upon the type of ballast and because the choice of an appropriate ballast involves highly technical electrical considerations, selection should be left to an electrical or lighting consultant.

(f) Self-Ballasted Lamps

These have been available for some years and consist of a screw-base color-corrected lamp with an internal resistive/reactive ballast. They have a CRI of 50, an efficacy of 20 to 25 lm/W, and a correlated color temperature (CCT) of 3500 K to 4000 K. Their great advantage is their long life, which can be used to advantage in applications involving burning periods of 8 to 10 hours mini mum, relative inaccessibility, limited space that precludes ballast installation, and indifferent color requirements.

(g) Application

Mercury-vapor lamps are applicable to indoor and outdoor use with proper attention to color and fixture luminance. The most common exterior application is for parking lots. Indoor application is generally limited to mounting heights of 10 ft (3.1 m) AFF (above finish floor) or higher to avoid direct glare and permit adequate area coverage.

Their use in industrial spaces and stores was once common, but today retrofitting with metal-halide lamps is typical. Warehouses and non-color sensitive industrial areas continue to use mercury vapor lamps.

20. METAL-HALIDE LAMPS

This lamp began its life in the early 1960s as a modified mercury-vapor lamp. Major advances in miniaturization, color rendering, color tempera ture, and consistency-by the addition of halides such as thallium, indium, and sodium to the arc tube-resulted in changes in the output, efficacy, color, and life of the lamp. Metal-halide lamps have excellent color characteristics and therefore almost unlimited applicability. The number of types and sizes is so large, and changes so rapidly, that any abbreviated tabulation would be inadequate at best and misleading at worst. As with all lamps, but more so with lamps undergoing intensive development, a current manufacturer's catalog should be consulted for accurate lamp data. Pulse-start metal-halide lamps utilize a glass arc tube to contain the arc. Pulse-start technology includes a new family of lamps (it has been used with high-pressure sodium lamps), and improves the start system, efficacy, and lumen maintenance and yields faster warmup and restrike.

Ceramic metal-halide lamps were introduced to the market a number of years ago and have become an industry standard, offering a high CRI of 80 to 90, a color temperature of 3000 K or 4100 K, improved lumen maintenance, and stable color consistency. Typical metal-halide lamp characteristics and types are discussed in the following subsections.

(a) Lamp Configurations

Construction details for a basic metal-halide lamp are shown in FIG. 24 and are similar to those of its "parent" mercury-vapor lamp, illustrated in FIG. 23. In addition to this design (in a BT shaped bulb envelope), metal-halide lamps are manufactured in elliptical bulbs, PAR reflector lamps, and single- and double-ended tubular shapes (Figs. 22 and 24 to 26).

FIG. 24 Construction details of a 400-W standard metal-halide lamp, which can be mounted either horizontally or vertically.

FIG. 25 Various configurations of metal-halide lamps. Clock wise from the bottom left: phosphor-coated and clear elliptical bulbs; PAR 30 and 38 reflector lamps; single-ended and double ended tubular lamps-all have ceramic arc tubes and a CRI greater than 80. (GE Lighting.)

FIG. 26 Construction details of a PAR enclosure for a metal-halide lamp. The lamp itself is tubular and constructed with a surrounding protective shield designed to contain arc tube fragments in the event of a violent rupture.

(b) Safety

Being essentially a modified-vapor-mercury lamp, the metal-halide lamp carries the same safety warning as mercury lamps. An additional warning, however, refers to the fact that metal-halide lamp arc tubes have a tendency to explode; therefore, the lamp must be used in an approved enclosing luminaire. All major manufacturers also make lamps with internal shields that will contain the flying pieces of a ruptured arc tube without damaging the outer bulb. Such lamps may be used in open lighting fixtures. One such lamp is illustrated in FIG. 26.

As with mercury-vapor lamps, manufacturers also produce a line of safety metal-halide lamps that self-extinguish within 15 minutes of an outer bulb fracture, thus limiting exposure to harmful UV radiation. Both of these safety designs are clearly noted on the lamps by trade name and sometimes by description.

(c) Designs, Shapes, and Ratings

The metal-halide lamp designs available as of this writing are:

• Standard lamps, available in ED, BT, and PAR shapes, in wattages from 50 to 1500 W, efficacies of 75 to 105 lumens per watt with magnetic ballasts, and slightly higher with electronic ballasts (efficacy also increases with wattage). Lamps are clear or phosphor-coated, with a CCT ranging from 3000 K to 4200 K and a CRI ranging from 65 to 85. Their life, at 10 burning hours per start, varies from 10,000 to 20,000 hours depending upon lamp type, size, and burning position.

• Safety-shielded lamps with integral shields to contain a ruptured arc tube.

• Self-extinguishing lamps designed to shut down automatically upon a break in the outer glass envelope.

• High-output lamps designed for a specific burning position (that is specified on each lamp). Output is 5% to 8% higher than that of standard lamps, but the color is somewhat poorer, with a CRI range of 65 to 70.

• Single-ended and double-ended tubular lamps.

These lamps are characterized by a very high CRI (80 to 93); a somewhat shorter life than standard lamps, particularly for the single-ended units (6000 to 10,000 hours); and slightly lower efficacy (70 to 85 lm/W). They are intended for applications requiring very high color rendering.

TABLE 9 Typical Data for Clear High-Pressure Sodium-Vapor Lamps

(d) Operating Characteristics

Like mercury-vapor lamps, metal-halide lamps are not instant-starting, requiring approximately 2 to 3 minutes on initial startup and 8 to 10 minutes for restrike. (Tubular lamps require only about half of these times.) As a result, when they are used for indoor installations, a secondary instant-start source must be available. A number of manufacturers produce special hot-restrike ballasts that provide immediate restrike of lamps on restoration of power after an outage. Lamp output on restrike is inversely proportional to the duration of a power outage.

It is important to note that the spectrum of light produced by a metal-halide lamp changes as a lamp ages. The change is gradual, definite, but usually unnoticed and depends upon the particular design of lamp. Where color rendering is important, or where the lamp is used with other light sources, a designer should choose metal-halide lamps that are specially made for color stability.

These lamps are designed not to vary in CCT more than 200 K over the lamp life. Finally, dimming or reduced output operation of metal-halide lamps is not normally recommended because of the very noticeable color shift that occurs when a lamp is dimmed.

(e) Lamp Ballasts

Metal-halide lamps operate satisfactorily on a simple reactor ballast, although a separate ignitor is usually required to start the lamp. These ballasts have a low power factor, of about 50%, which is undesirable from the perspective of energy conservation, wiring economy, electrical losses, and component heating. High-power-factor magnetic ballasts (pf > 90%) are also available. Magnetic ballasts are large, heavy, and often noisy. The last characteristic can be improved by the use of a pot ted (epoxy-filled) ballast. Electronic ballasts are also readily available, along with dimming and multi level ballasts. However, as noted previously, dimming ballasts are not frequently used because of the large shift in lamp color that they cause.

21. SODIUM-VAPOR LAMPS

The highest-efficacy general-purpose HID source available is the high-pressure sodium lamp (HPS). The basic construction of this type of lamp (also designated SON) is illustrated in FIG. 27, which shows schematic drawings of the design. Typical performance data for various types of HPS lamps are given in TABLE 9. Construction is quite different from that of mercury-vapor and metal-halide lamps. The characteristic color of HPS lamps stems from the spectral absorption phenomenon of the sodium contained in the arc tube-with a resultant pronounced yellow-tinted light.

(a) Primary Characteristics of HPS Lamps

Standard HPS Lamps. The extremely low CRI (20 to 22) is not acceptable where any degree of color rendition is required, thus limiting the standard HPS lamp to exterior areas and road lighting.

Color-Corrected HPS Lamps. Color can be improved considerably by increasing the pressure inside the arc tube. This causes some of the sodium in the arc to be reabsorbed, and the radiated light widens its spectrum into the red range (at the expense of efficacy and lamp life).

"White" HPS Lamps. A still greater increase in lamp pressure improves lamp color and yields a "whiter" lamp color in a limited wattage range.

These low-wattage, reduced-efficacy, shortened life lamps are normally operated with small, light weight electronic ballasts.

Because of its high output and narrow linear arc tube, any HPS lamp is a potential glare source, and wattages of 150 W and higher must be either completely shielded or mounted at sufficient height (if in an open reflector) to be above the near field of vision. These glare-prevention strategies also apply to metal-halide lamps.

If a diffusing coating is added to a sodium-vapor lamp, the entire glass envelope becomes the light emitting source. This reduces lamp luminance, and therefore glare potential, drastically but also reduces output (and efficacy) by 6% to 8%. Light distribution from a coated lamp in an open reflector is vastly improved, as can be seen in FIG. 28.

FIG. 27 Simplified drawings of the internal construction of two high-pressure sodium (SON) lamp designs. (a) Non-cycling lamp (in an E-shaped glass bulb) designed to indicate by a special color when the lamp has reached the replacement stage. Unlike standard HPS lamps, this lamp will not cycle on and off at the end of its useful life. (b) Retrofit HPS (SON) lamp in a BT bulb, intended for direct replacement of an existing mercury lamp. This lamp operates efficiently on a mercury lamp ballast. (Osram-Sylvania.)

(b) Other Operating Characteristics

In contrast to both mercury-vapor and metal-halide lamps, HPS lamps do not emit any appreciable UV radiation, do not tend to rupture violently, and can be installed in any position without affecting operating characteristics.

Like all discharge lamps, HPS lamps require a ballast for ignition and arc control. Due to the extremely high voltage required for lamp ignition, both magnetic and electronic ballasts contain an electronic ignition circuit. Because of this, an HPS lamp must be used with a compatible ballast that carries the same ANSI designation as found on the outer glass bulb. If it is used with an incompatible ballast (and fixture), the lamp may rupture and constitute a serious safety hazard. As with metal-halide lamps, ballasts are available for HPS lamps that provide instantaneous restrike after a power interruption. Light output on restrike is inversely proportional to the length of the outage; for example, after a 10-second outage, restrike light output will be 85% of full capacity, whereas after a 2-minute outage, lamp output will be only 10%. It then takes approximately another minute to regain full output.

FIG. 28 (a) HPS retrofit and non-cycling lamps. (b) A narrow linear arc tube of an HPS lamp is not suitable for use in an open reflector designed for a larger source such as a phosphor-coated mercury lamp. (c) A coated HPS lamp creates a large, diffuse source with a highly improved light distribution. (Osram-Sylvania.)

(c) Lamp Design Types

In addition to standard HPS lamps, including the color-corrected types, three additional special lamps have been developed to solve specific problems. They are:

1. Non-cycling lamp. As an HPS lamp ages, its arc voltage rises. Eventually, the ballast is unable to sustain the arc and the lamp extinguishes.

After cooling, the lamp lights to full brightness and soon thereafter extinguishes again. This on-off cycling is characteristic of an HPS lamp at the end of its life. To eliminate this condition, a special non-cycling lamp was developed that uses very little sodium amalgam in the arc tube and is more environmentally friendly because of this reduced content. To enhance this environmental aspect, these lamps are made with a lead-free base and lead-free solder. Photometric characteristics are similar to those of standard lamps. The E-shaped lamp in Figs. 12.27 and 12.28 is of this design.

2. Standby lamps. As noted previously, HPS lamps require a minute or more to restrike after being extinguished. A crowded public area plunged into complete darkness is a recipe for disaster; therefore, such areas must always be furnished with instant-on emergency lighting. Standby lamps have two arc tubes. When the lighted one is extinguished due to a momentary power loss, the second (cool) arc tube immediately begins to glow-assuming that voltage has returned. This arrangement may be accept able to some code authorities as a substitute for instant-on replacement lighting following a momentary power outage. It is not acceptable as emergency lighting as required by NFPA 101: Life Safety Code.

3. Retrofit lamps. These are HPS lamps designed as a direct replacement for mercury-vapor lamps of the same wattage. They are enclosed in BT shaped envelopes of the same size as the lamp being replaced, and they operate properly on mercury-vapor lamp ballasts. A retrofit lamp is illustrated in FIG. 28.

22. LOW-PRESSURE SODIUM LAMPS

The low-pressure sodium (SOX) lamp produces light characteristic of sodium's monochromatic saturated yellow color, making it inapplicable for general lighting. Because of its very high efficacy of more than 150 lm/W, including ballast loss (but with a CRI of 0), it can be applied wherever color rendition is not an important criterion but energy efficiency is. Thus, SOX lamps are used for street, road, parking lot, and pathway lighting. SOX lamps are used around astronomical observatories because the yellow light can be filtered out of a telescope. Another desirable aspect of SOX lamps is their 100% lumen maintenance. This, coupled with a discharge lamp's typically long life (18,000+ hours), makes SOX lamps fairly economical in terms of life-cycle cost.

OTHER ELECTRIC LAMPS

23. INDUCTION LAMPS

The induction lamp is filled with low-pressure mercury vapor. When ionized by a high-frequency induction coil inside the lamp, the mercury vapor produces UV radiation, which then strikes a phosphor coating on the inside of the lamp, producing light. Similar to the light-producing process used by standard fluorescent lamps, the difference is that the gas is ionized by an induction coil (rather than an electron stream)-thus the name induction lamp.

Two such designs are shown in Figs. 29 and 30.

Characteristics of the design shown in FIG. 29 are as follows:

The extraordinarily long life, excellent lumen maintenance, and instantaneous restrike time make the induction lamp suitable for illuminating public areas. As with all light sources, a comparison with other sources can only be made for a given proposed usage based on project-specific considerations.

A self-contained induction lamp of lower watt age is illustrated in FIG. 30. This lamp is built into a modified R-shaped envelope with a standard Edison base, and is therefore readily applied as a direct replacement for a 100-W incandescent reflector lamp. The lamp's published performance figures are:

24. LIGHT-EMITTING DIODES

FIG. 29 Schematic diagram of an induction lamp (rated 85 W, including all losses) shows its operating principles. The lamp is 4.3 in. (109 mm) in diameter and 7.5 in. (191 mm) high overall. The high-frequency generator (C) produces a high-frequency current, which circulates in the coil on the power coupler (B). This ionizes the mercury vapor inside the lamp (A), producing UV radiation.

The UV radiation strikes the fluorescent coating inside the lamp, producing light. (Illustration Philips Lighting Company.)

FIG. 30 Cutaway of the GE induction lamp showing the essential elements: induction coil, phosphor-coated bulb with mercury vapor fill, and electronic ballast. This 23-W lamp in a modified R (reflector)-shaped bulb has a height just under 5 in. (127 mm) overall and a 3-in. (76-mm) maximum bulb diameter. (GE Lighting.)

Light-emitting diodes (LEDs) (or light-emitting semi conductors) have been used since the 1960s, in a wide range of applications, such as in lamps (medical instruments, bar coders, fiber-optic communication), mobile technologies (phones, digital cameras, laptops), consumer appliances, automotive (instrument panels, signal lights, courtesy lighting), and signals (traffic, rail, aviation). Common LED uses in architectural illumination include signage, retail displays, emergency lighting (exit and emergency signs), and accent lights for pathways. Recently, LEDs are seeing increased use in residential projects and for specialty applications in commercial projects. They are easy to install, should last longer than incandescent or fluorescent lamps, are very efficient (lumens per watt), and do not produce nearly as much heat as an incandescent lamp.

LEDs are available in a full range of colors, are small (18 in. [3mm] across), have a very low light output, use very little power, and have a fast response. They emit light in proportion to the for ward current through the diode. The color of the light emitted by a LED depends upon the band gap material of the semiconductor, such as gallium arsenide phosphide (GaAsP) or gallium phosphide (GaP). GaAsP emits from red to amber, depending upon the concentration of phosphorus. GaP emits green light and can emit red light.

25. SULFUR LAMPS

The principle of microwave energy excitation has been applied successfully to another recently developed lamp that consists of a golf ball-sized globe filled with an inert gas and a few milligrams of sulfur. In contrast to the induction lamp described previously, no mercury vapor is used in this lamp, and the radiation is full-spectrum light with very little UV or infrared radiation. The originally developed lamp was a 6-kW unit that emitted more than 400,000 lumens. The most recent (1998) version of the lamp has the following characteristics:

The current sulfur lamp technology provides approximately the same output as a 1000-W HPS lamp but with much better color characteristics.

It also compares to about 1200 W of metal-halide lamp(s) with similar efficacy, but again, because of its full-spectrum radiation, it has superior color characteristics. Initial trials of the lamp as a driver for a "light pipe" have been highly successful. Because the lamp is still in development at this time, its eventual commercial usage is difficult to predict.

26. FIBER OPTICS

Although optical fibers have been available since the 1920s, practical applications, in the medical field, were not developed until the late 1950s and early 1960s. Bundled fibers used as a diagnostic tool can deliver light to remote regions of the body and carry coherent (understandable) images back to a doctor. In recent years, fiber optics have made their most significant advances in the communications field: long-distance telephone cables or thousands of paired wires have been replaced by a single-fiber cable. Architectural applications of fiber optics in recent years have included alternatives to directly replace recessed ceiling downlights, track and display case lighting in museums, and lighting for pools/spas, supermarkets, and other commercial buildings.

Illuminators for fiber-optic systems utilize a variety of lamp types. The primary advantages of using a fiber-optic system are that no heat is produced where the light exits the fiber, and no UV radiation is transmitted through the fiber.

References / Resources

ANSI. 2004. ANSI C82.11: Lamp Ballasts-High Frequency Fluorescent Lamp Ballasts-Supplements.

American National Standards Institute. New York.

ANSI. 2002. ANSI C78.1451: Electric Lamps-Use of Protective Shields with Tungsten-Halogen Lamps- Cautionary Notice. American National Standards Institute. New York.

CIE. 1970. Daylight, International Recommendations for the Calculation of Natural Daylight (Publication No. 16 E-3.2). Commission Internationale de l'Eclairage. Paris.

Eley Associates. 1993. Advanced Lighting Guidelines: 1993 (EPRI TR-101022s Rev. 1). Electric Power Research Institute. Palo Alto, CA.

Federal Energy Management Program (energy efficient fluorescent ballasts): IESNA. 2000. IESNA Lighting Handbook, 9th ed.

Illuminating Engineering Society of North America. New York.

Krochman, J. 1963. "Über die horizontal Beleuch tungsstärke der Tagesbeleuchtung." Lichtechnik, Vol. 15, No. 11.

NFPA. 2009. NFPA 101: Life Safety Code. National Fire Protection Association. Quincy, MA.

Rennhackkamp, W. M. H. 1967. "Sky Luminance Distribution in Warm Arid Climates." Proceedings 16th International Conference on Illumination, Washington, DC.

U.S. Department of Energy High Performance Buildings Program: U.S. Environmental Protection Agency, Energy Star-labeled lamps and luminaires

Prev. | Next

Home  Similar articles   top of page