Guide to Mechanical / electrical equipment for buildings: PART III: ILLUMINATION: Electric Lighting Design

1. DESIGN CONSIDERATIONS

(a) General

At this point in the lighting design process, where design development of the electrical lighting system occurs, lighting hardware is chosen based upon considerations brought forth from earlier design stages, and appropriate validation calculations are performed. Some spaces require overall uniform illumination. These spaces are calculated by the lumen method, which yields aver age illuminance. Other spaces utilize local lighting alone, or local lighting in addition to general lighting, requiring point-by-point illuminance calculations or some other method for restricted-area calculation. Additional considerations at this design stage are the control strategy; type of ceiling system (e.g., modular, movable fixture, and integrated service); and ancillary considerations of ballast noise, luminaire heat distribution, and maintenance. Also decided here is whether to utilize workstation-mounted or built-in lighting, both of which are principally applicable to open-plan spaces.

(b) Luminaire Characteristics

The purpose of a luminaire or lighting fixture (the terms are synonymous) is twofold: physically, to hold, protect, and electrify the light source; and photometrically to control the lamp output (i.e., to redirect the light produced) because most common light sources emit light in substantially all directions.

The means by which this beam-shaping is accomplished are well known. The characteristics of reflectors, baffles, lenses, and louvers that perform these functions are discussed in some detail in the following sections. However, the problem in luminaire selection is that the requirements are, in many respects, incompatible, and therefore a trade-off between various luminaire characteristics must be made. Thus, for instance, high efficiency installations can entail high fixture luminance with resultant glare. Desirable wall lighting means high-angle luminaire out put with resultant direct glare. Low-angle light means minimum direct glare but possible veiling reflections. A high shielding angle (>35º) means good visual comfort but reduced efficiency, and so on. It is therefore obvious that

• No single luminaire design is ideal for even a majority of applications.

• To make an intelligent selection among the hundreds of lighting fixtures commercially available, it is absolutely necessary that the designer understand both the specific requirements of the application and the light control characteristics of the luminaire being considered.

FIG. 1 Semi-direct fluorescent fixture crosswise distribution (two lamps, 32 W each, prismatic enclosure). Note the sharp cutoff and wide, horizontally even distribution of (a) in contrast to the diffuse, broad, and horizontally uneven distribution of (b).

FIG. 2 Methods for shielding downlights using circular shields for vertically symmetrical sources such as incandescent and HID lamps. Halogen lamps mounted vertically in R lamp envelopes are symmetrical; horizontally mounted units in PAR or other reflectors are not. Baffled downlights (a-c) control unwanted high-angle light by cutoff as illustrated. Black baffles aid by absorbing light and appearing dark. Other colors give a ring of light at the baffled edge. Cones (d, e) control brightness by cutoff and by redirection of light due to their shape. They are either parabolic or elliptical. A light specular finish such as aluminum appears dull; a black specular finish appears unlighted. Black finishes require high-quality maintenance because dust shows as a bright reflection. CFL lamps in reflectors are not normally a serious direct glare concern and are considered to be vertically symmetrical. A shielding angle of 45º minimum is recommended for high-luminance lamps.

2. LIGHTING FIXTURE DISTRIBUTION CHARACTERISTICS

At this point, review of intensity distribution curves by the reader would be useful.

The two distribution curves shown in FIG. 1a are actual test results of two 2-lamp, 1-ft-wide by 4-ft-long (0.3 by 1.2 m) semi-direct fluorescent fixtures with prismatic enclosure. The flat bottom of the curve in FIG. 1a indicates even illumination over a wide area, therefore permitting a high spacing to mounting-height ratio (S/MH) (1:4) for uniform illumination. The rounded bottom of the curve in FIG. 1b indicates uneven illumination and closer required spacing for horizontal uniformity as defined in Section 15.5.

The straight sides of the curve in FIG. 1a show a fairly sharp cutoff, and the small amount of light above 45º means high efficiency, probably insufficient wall lighting, barely adequate diffuseness, and very little direct glare potential but a distinct possibility of veiling reflections. Conversely, the curve in FIG. 1b shows a large amount of horizontal illumination (above 45º) with resultant direct glare potential, diffuseness, and relative inefficiency, because horizontal light output is attenuated by multiple reflections before reaching the horizontal working plane. Here, however, low output below 45º minimizes reflected glare potential. The uplight component of luminaire (a) is directed outward to cover the ceiling and will not cause hot spots; the corresponding light from fix ture (b) is concentrated above the fixture and gives uneven illumination of the ceiling.

FIG. 3 Shielding of fluorescent lamps is less critical due to lower lamp luminance. For T12 lamps, 45º × 35º crosswise/lengthwise shielding as shown is excellent, and 35º × 30º is satisfactory. For T8 lamps, 45º × 35º should be used. Because fixture luminance is higher in the transverse direction (a) than lengthwise (b), a better cutoff angle is required. The shielding elements may be louvers or baffles. Opaque shielding elements have a higher visual comfort rating than translucent plastic units.

These conclusions were reached based on the following observations:

1. Uniformity of illumination requires that the intensity at angles above the nadir (0º from the vertical) be greater than the intensity at 0º so that location-points distant from the fix ture centerline obtain the same illumination as those below the fixture (because illuminance varies inversely with the square of distance). This is exactly the case with the flat-bottom characteristics of FIG. 1a. Therefore, such fixtures can be spaced more widely than the units of FIG. 1b.

2. High efficiency is achieved by directing the luminaire output to the work plane (i.e., from 0º to 45º from the vertical). Light above 45º is directed to the walls and reaches the working plane only after multiple attenuating interreflections.

3. Diffuseness exists when light reaches the work plane from multiple directions. This requires that light be reflected from walls and ceilings to the work plane, which in turn requires luminaire light output above 45º from the vertical.

4. Direct glare is caused by light output at high angles (i.e., above 45º from the vertical). Direct glare from linear fluorescent fixtures can be minimized by placing the long axis parallel to the line of sight, because such fixtures normally have low endwise high-angle output.

5. Reflected glare is caused by reflection of low angle output from the task. Therefore, fixtures with control means that limit output between 0º and 45º minimize the potential for veiling reflections.

6. Shielding is a function of the shape of the fixture housing plus any additional lamp concealment means, such as louvers or baffles. The shielding angle is defined as the angle between a horizontal plane through the louvers or baffles and the inclined plane at which the lamp first becomes visible as one approaches the fixture. Cutoff angle is usually defined as being synonymous with the shielding angle.

However, because some sources define it as the complement of the shielding angle, it is best to avoid the term and use only shielding angle.

7. Ceiling illumination is produced by light above the horizontal. As with light below the horizontal, a spread characteristic (FIG. 1a) means good ceiling coverage, no hot spots, and, ultimately, good diffuseness. Concentrated uplight means a potential hot spot if the fixture suspension hanger is too short, and in any event it yields uneven ceiling illumination.

Thus, we see that a rapid inspection of a fixture curve performed by an informed person can yield a large amount of information on the fixture's performance. The reader is encouraged at this juncture to review the comments on the two distribution curves of FIG. 1 and then analyze similarly other distribution curves in manufacturers' catalogs.

FIG. 4 Shielding elements. (a) The most basic shielding element is the lamp reflector, which may double as the fixture body. (b) Shields perpendicular to the long axis of a linear fluorescent lamp are normally called baffles. They are less important than lengthwise shields (louvers) because endwise lamp luminance is lower than crosswise luminance. (c) Two-way shielding is most effective but significantly lessens luminaire efficiency.

3. LUMINAIRE LIGHT CONTROL

(a) Lamp Shielding

Except where it is desired to use a bare lamp as a source of sparkle, such as in chandeliers and other decorative fixtures, all lamps in interior fixtures should be shielded from normal sight lines (i.e., sight lines in a head-up, eyes-straight-ahead position; see Fig. 11.19). The reason is obvious; bare lamps are so bright that they usually constitute a source of direct or even disabling glare, depending on the apprehended angle (closeness to the eye and size of the lamp) and eye adaptation level. The range of permissible luminaire luminances (listed in Table 11.3) of 1000 to 7000 cd/ m^2 depends upon these two variables (apprehended angle and adaptation level).

As a general rule, exposed incandescent lamps, 6 W and larger, are sources of direct glare and should be avoided. Note that the upper direct-glare limit corresponds to the luminance of a bare 34-W T12 fluorescent tube, which accounts for the fact that bare-lamp fluorescent fixtures are well tolerated. However, when such a fixture is re-lamped (and re-ballasted) with a more efficient, better CRI, T8 lamp whose luminance exceeds 10,000 cd/ m^2, it becomes a source of annoying direct glare that actively impairs visual ability.

Shielding of lamps is accomplished with the fixture housing/reflector or with baffles and louvers, as mentioned previously (see Figs. 2, 3, and 4). Fluorescent fixtures require shielding most when placed crosswise to the line of sight, thus exposing the entire length of the lamp to the field of view. Such fixtures require longitudinal baffles, deep housings, or louvers to provide the necessary shielding. Alternatively, when at all possible, place fixtures with their long axis in the direction of sight lines.

(b) Reflectors

It is important to understand the action of luminaire reflectors. The basic shapes and beam patterns are illustrated in Figs. 5 through 7. Note from FIG. 6 that the so-called pinhole down light requires an elliptic reflector to focus the light through this hole at point f2 in order to maintain even minimal fixture efficiency. Elliptic reflectors are large, and frequently the space above the ceiling is too restricted for their use. Lamps with integral elliptic reflectors can be utilized with a standard baffled reflector to achieve roughly the same effect.

FIG. 5 Parabolic reflector action shown with the fixture below: (a) with the source at the focal point, rays are parallel; (b) with the source below the focal point, they converge; (c) with the source above the focal point, they diverge. This focusing action is illustrated by fixtures correspondingly designated. Note that type (c) requires a large ceiling opening to achieve even minimal efficiency.

FIG. 6 (a) Action of an elliptical reflector section. With the light source at focal point f 1 the light converges at the other focal point, f 2. This effect is useful in fixture design, as in (b). By projecting light up only (through the use of a silvered bowl lamp) the output light can be redirected through a constricted aperture at the other focal point, with little loss. This design is the basis of high-efficiency "pinhole" downlights.

FIG. 7 The extended section reflector allows the source to be concealed (shielded) while projecting its light directly down, but horizontally displaced, from the source.

(c) Reflector Materials

Until fairly recently, reflector materials were of two types: white gloss paint for portions of fixture body interiors that acted as reflectors, and formed anodized aluminum sheet for the shaped reflectors of the types shown in Figs. 5 through 7. The reflectances (reflection factors) of both of these materials are approximately the same, varying between 0.84 and 0.88 when new and clean. Neither, however, is truly specular; the paint finish is actually primarily diffuse, whereas the aluminum is principally specular. Where shaped reflectors are not used, as in the case of a fluorescent troffer, the lack of specularity is essentially immaterial because, at worst, the diffuseness will reduce luminaire output slightly by increasing the number of inter-reflections within the fixture body (FIG. 8a). The idealized specular reflections for shaped reflectors shown in Figs. 5 through 7 are just that; in reality, the reflectances are considerably more diffuse and become increasingly so with reflector aging and dirt accumulation.

Painted fixture body interiors lose their high reflectance by rapid aging due to elevated temperatures and accumulation of dust and dirt. This causes a decrease in overall reflectance, which is compensated for by initial overdesign, as explained in Section 15.20. However, because a rectangular fixture body is not an accurately shaped reflector, the result of overall reflectance reduction is simply an overall reduction in output while maintaining the same photometric distribution characteristics.

Energy conservation programs and utility rebates (now generally terminated) led to the introduction in retrofit of very-high-reflectance auxiliary reflectors that were added to existing fluorescent troffers approximately as shown in FIG. 8b. Unfortunately, many of the claims of highly increased efficiency were based on retrofitting aged, dirty luminaires, and their results can be very misleading. A reasonable estimate of the possible improvement in luminaire efficiency can be made by considering these facts:

• Approximately 40% of a lamp's output in an open luminaire is directed downward and is therefore completely independent of any reflector action.

• The difference in reflectance between a new, clean, painted surface and an old, dirty surface is, at most, 50%. That means that the maximum light loss of an open fixture due to poor maintenance is 50% of 60% (the maximum reflected light component), or 30% of the overall light output. Reference is always to an open-bottom fixture.

• The maximum reflectance of the best (and most expensive) silver reflectors is about 95%, comprising 93% specular and 2% diffuse. This is only 10% higher than the original minimum 85% paint reflectance. Therefore, at most, retrofit of a very dirty, old fixture with a high quality (expensive) silver reflector improves performance by the 30% lost to dirt plus 10% of the 60% reflected light for a total of 36% maxi mum. Relamping, of course, also improves out put, but that is not connected with fixture body reflectance.

• Simple cleaning of a very dirty fixture body restores 20% to 25% of the light loss. The remaining loss is due to aging of the paint. A cost analysis is required to determine whether the 10% to 15% differential in light output between simple cleaning and silvered reflector addition is economically feasible.

FIG. 8 (a) Approximately 40% of lamp output in an open linear fluorescent fixture is unrestricted. The remainder leaves the fixture after one or more reflections. (b) A mirrored reflector narrows the distribution pattern of the luminaire by specular reflection and increases output somewhat.

An important consideration in retrofit and new construction is the photometric characteristic of high-reflectance linear fluorescent luminaires.

In general, they are shaped like the curve of FIG. 1b. This means, as explained in Section 13.16, that light is concentrated downward, resulting in a requirement for closer luminaire spacing to obtain uniform illumination. In new construction, this can be considered in design, although it can be a serious economic penalty. In retrofit work, it can result in unacceptable lack of illuminance uniformity, requiring additional luminaires and expensive relocation of existing units.

Another factor to be considered is the degree of maintenance required to keep "super" reflectors in pristine condition in order to achieve the 15% ± maximum output differential. To determine this, designers should request an aging test and inspect a previously retrofitted installation with ambient conditions and cleaning schedules similar to those of the area under consideration.

4. LUMINAIRE DIFFUSERS

Diffusers are the devices placed between the lamp(s) and the illuminated space, that function to diffuse the light, control fixture brightness, redirect the light, and obscure (hide) and shield the lamps. As most of these devices perform multiple functions, they are discussed individually.

FIG. 9 Comparison of typical candlepower distribution curves for common linear or PL lamp fluorescent luminaire diffuser elements (for a full description of types [a]-[e], see text). Note that for a given geometry of viewer and luminaire, the severity of veiling reflections depends entirely on the fixture's photometric characteristic. In the individual figures, the potential to produce reflected glare is indicated by the weight of the line representing fixture output and reflectance from the work task. The batwing distribution (e) concentrates its output in the 30º to 60º range, which minimizes both direct and reflected glare.

(a) Translucent Diffusers

Because these do not redirect the light but merely diffuse it, the distribution characteristic is circular, as seen in FIG. 9a. Typical examples of this type are white opal glass, frosted glass, and white plastics such as Plexiglas, polystyrene, vinyl, and polycarbonates. The distribution is basically the same as it would be for bare lamps. Lamp-hiding power is good. Depending on the material, direct glare can be a problem (visual comfort probability [VCP] is poor). Veiling reflections are high.

The S/MH does not exceed 1.5. The fixture is generally inefficient. Wall illumination is good because of a large component of high-angle light (which reduces VCP). The net result of using this type of diffuser is to lower lamp luminance by distributing the lamp output over a larger diffusing area. Applications include corridors, stairwells, high-ceilinged spaces, and other areas without demanding visual tasks.

A special type of flat plastic panel that polarizes the transmitted light was introduced some years ago. These panels held great promise because they produce a marked decrease of veiling reflection at an angle of 60º, but much less at other angles. Because most viewing is in the range 20º to 40º, using these panels does not result in any appreciable reduction in reflected glare in normal office work situations. From experience in a drafting room equipped with luminaires utilizing high efficiency multilayer polarizers, it can be stated that visual discomfort from reflected glare, as personally experienced and as reported by a large staff, is not noticeably reduced.

(b) Louvers and Baffles

These are generally rectangular section, metal or plastic, and serve primarily to shield the source (see Figs. 2, and 3) and to diffuse the output, particularly when plastic translucent louvers are used.

Candlepower distribution curves are shown in FIG. 9b. The exact curve shape depends on the shielding angle, design of the louver, and its finish. Louvers finished in specular aluminum or dark colors exhibit low direct glare. The large downward light component can cause serious veiling reflections.

Overall fixture efficiency is average.

The S/MH, a luminaire metric that indicates the maximum spacing permissible for a given luminaire mounting height that yields uniform illumination and is given as a dimensionless ratio, is fully explained in the next section. For this diffuser type, it does not exceed 1.5 and varies inversely with the shielding angle. This is because the basic circular distribution is changed to an egg shape by cutoff and redirection, reducing the high-angle light.

Thus, a 45º shielding angle results in lower direct glare potential but requires closer spacing.

A special design in this category is the miniature eggcrate parabolic wedge louver shown in FIG. 10. These units redirect a large portion of the light directly downward, and because of this redirection and their specular finish, they appear completely dark-darker, indeed, than the unlighted portion of the ceiling when viewed obliquely. Fixtures using these louvers have low efficiency due to trapped light, with a maximum coefficient of utilization of about 0.5 (see Section 15.20). VCP is very high, but veiling reflections can be troublesome; S/MH varies between 1.0 and 1.5. The shielding angle is usually 45º. A typical candlepower distribution curve is shown in FIG. 9c. When these units are used, additional wall lighting is almost always required. The luminance of fixtures and surfaces can be easily checked in the field using a portable luminance meter as in FIG. 11 Because of the low efficiency of the parabolic louver design shown in FIG. 10a caused by the wide light-trapping tops of the louvers, an improved version was developed with the tops shaped to reflect incident light efficiently. This improved design, which is shown in FIG. 10b, increases luminaire efficiency by about 20%.

FIG. 10 (a) Section through a conventional, miniature parabolic wedge, eggcrate type of louver. These units give exceptionally low brightness when seen at a normal viewing angle. Most such units are made of aluminized plastic. Fixtures equipped with these units exhibit low overall efficiency due to the large amount of light trapped by the broad top of each parabolic wedge. (b) A modified wedge design uses a curved top on each wedge to redirect and utilize light striking the top. (c) Solid lines represent light rays redirected by the bottom curve, whereas dotted lines show light redirected by the top curve, which was lost in the design of (a). Typical louver cell dimensions are a ½-in. (12-mm) cube for design (a), with a consequent 45º shielding angle, and 5 6 to ¾-in. (21 to 19 mm) square by ½ in. (12 mm) high for design (b), giving a 35º to 45º shielding angle.

FIG. 11 Checking the luminance of a fluorescent fixture with a luminance meter.

FIG. 12 Action of a Fresnel lens. With a Fresnel lens fixture, a smaller housing without a reflector can be used while still maintaining beam control. The lens performs the same function as a reflector, controlling the beam as a function of source placement. By utilizing a lens fixture, the curved reflector (a) can be largely eliminated, yielding a smaller fixture while maintaining accurate beam control. A common design (b) uses a regressed lens to provide shielding, although lens brightness is not normally objectionable.

(c) Prismatic Lens

Many designs are available with varying distribution characteristics. Figures 15.1a and 15.9d can be taken as typical of this genre. They produce an efficient fixture (high coefficient of utilization), good diffusion, wide permissible spacing-an S/MH as high as 2.0-and low direct glare (high VCP). Veiling reflections can be troublesome, depending upon viewing angles and position.

(d) Fresnel Lens

The action of this lens is similar to that of a reflector.

Lamp-hiding power is poor, but efficiency is high and visual comfort is good. S/MH is seldom more than 1.5.

FIG. 13 (a) Linear batwing distribution with extremely sharp cutoff in the upper and lower ranges. The curve is taken across the lamp axis for a single-lamp unit. (b) Distribution curves for a radial batwing distribution lens. Note that the perpendicular, parallel, and diagonal curves are almost identical. Zonal flux is maximum in the 30º to 60º range and drops off at both extremes, as desired.

(e) Batwing Diffusers

The theory behind this type of diffuser is covered in Section 11.31(c). A typical characteristic is shown in FIG. 9e. There are two fluorescent luminaire designs that produce the batwing shape distribution characteristic-a prismatic lens and parabolic reflectors and baffles.

1. Prismatic batwing diffusers. These are either linear or radial; that is, they produce the batwing distribution either in one direction or in all directions. Typical characteristics of both types are shown in FIG. 13. Note that the characteristic shape is more pronounced in the linear diffuser, which indicates better control of veiling reflections in that direction (usually cross wise). Fixtures equipped with these diffusers have good efficiency, low direct and reflected glare, and good diffusion. As with all enclosed un-gasketed fixtures, the lens acts as a dust trap, necessitating frequent cleaning to maintain high output.

2. Deep parabolic reflectors. These luminaires (FIG. 14) produce modified versions of the characteristic batwing distribution in the normal (crosswise) direction. Distribution in the parallel or lengthwise direction is circular (diffuse), indicating minimum beam control in that direction. These fixtures, like the batwing lens-type diffuser units, have high efficiency, high S/MH, low reflected glare, and low to very low surface brightness, making them usable in visual display terminal (VDT) areas. They are normally applied with the long axis in the direction of sight lines.

5. UNIFORMITY OF ILLUMINATION

In any space intended to be lighted uniformly with multiple, discrete, ceiling-mounted direct lighting system light sources, it is necessary to establish a fixture spacing that gives acceptable uniformity of illumination. A ratio of maximum to minimum illuminance on the working plane of 1:1.3 is readily acceptable because lesser ratios are not easily noticed. For general background or circulation lighting, a ratio of up to 1.5 is acceptable. The recommended S/MH (above the working plane) given by manufacturers (see the figures immediately above the distribution curves for each fixture in TABLE 1) are generally based upon a 1.0 illuminance ratio (FIG. 15). Therefore, the S/MH recommendation may be exceeded somewhat without significantly affecting uniformity.

FIG. 14 A 2-ft (610-mm) square, deep parabolic reflector luminaire designed for three F40, 22.5-in. (572-mm) twin-tube CFL lamps, with a rated output of 3150 lm each. Total power input is 134 W and ballast factor (BF) is 0.9, giving a luminaire efficacy rating (LER) of 42. The typical modified batwing crosswise distribution and circular lengthwise distribution are clearly shown. Photometric data of interest to lighting designers are also shown. ( Columbia Lighting.)

TABLE 1 Coefficients of Utilization for Typical Luminaires with Suggested Maximum Spacing Ratios

FIG. 15 The ratio of maximum to minimum illuminance should not exceed 1:1.3 in areas requiring uniform illumination.

When the luminaire's distribution characteristic is symmetrical in all directions, as is generally the case with small-source lamps such as incandescent, compact fluorescent (CFL), and high intensity discharge (HID), only a single S/MH figure is required. However, with the asymmetrical distribution of most fluorescent fixtures, an S/ MH ratio is required both crosswise and length wise. Due to the characteristic of the lamp itself, the transverse (crosswise, perpendicular) ratio is almost always considerably higher than the longitudinal (parallel, endwise, lengthwise) ratio.

See, for instance, fixtures 26, 28, and 42 in TABLE 1. The S/MH (also called spacing criteria [SC]) for a specific luminaire is determined by measuring the distance between two test luminaires that yields the same illuminance on the working plane midway between them as directly under each one. This ignores (deliberately) any contributions from other fixtures in a multifixture installation and from interreflections, accounting only for the direct component of illuminance from the two test fixtures. Therefore, in an actual installation with several rows of fixtures, the illuminance at point P1 in FIG. 16 is higher than the average by 20% to 30% because of the other fixtures and interreflections, and the illuminance at point P2 is approximately equal to the room average. The illuminance levels along the walls, assuming a distance between the last row and the wall equal to one-half of the side-to-side spacing, range from 60% of average at point P3 down to 50% at point P4, assuming light-colored walls.

FIG. 16 The diagram shows lighting fixtures installed according to the manufacturer's recommended spacing criteria (spacing to mounting height above the work plane ratio), with a row-to-row spacing, D, and a row-to-wall spacing, D/2, as shown at the left wall. Assuming a high-reflectance finish on the wall (light color), illuminances P3 and P4 are at least one-half of the illuminance directly below a fixture. At a dark wall, as on the right, illuminance would fall below this value. As a consequence, the designer would move the right row of luminaires closer to the wall, as shown.

FIG. 17 Typical distribution curves for approximating the ratio of fixture spacing to mounting height (S/MH) above the working plane for direct distribution luminaires. Although these curves were developed for point sources such as incandescent (and HID), they can also be used with asymmetric distribution luminaires such as for fluorescent lamps. The permissible S/MH is somewhat higher than the curves indicate because this is a semi-direct distribution and the ceiling light component permits wider spacing between units.

When walls are dark due to paint or aging, bookshelves, dark wood paneling, and the like, the illuminance levels drop to less than 50% of average, which is obviously insufficient as task lighting. To counteract this effect, particularly when placement of furniture is such that visual tasks will occur near walls, the designer has three choices:

1. Reduce the distance between the last row of fixtures and the wall to a third or less of the row to-row spacing. This also provides required wall lighting.

2. Provide some type of continuous perimeter lighting or wallwash units, both of which increase illuminance levels at the walls.

3. A combination of choices 1 and 2.

Because endwise illumination from linear fluorescent fixtures is considerably lower than crosswise illumination, end walls have lower illuminance than side walls and greater illuminance variation. It is, therefore, particularly important to provide some additional illumination, as discussed previously, and terminate fixture rows no more than 1 ft from an end wall. This is all the more important where visual tasks without supplementary task lighting occur at these walls.

We mentioned previously that the fixture in FIG. 1a had a high S/MH because of its flat-bottomed curve. This ratio, when not given by the manufacturer, may be approximated from FIG. 17. An accurate method of calculating maximum to minimum illumination ratios is available (see the IESNA Lighting Handbook, 1993).

The foregoing discussion of illumination uniformity was concerned with uniformity on a horizontal work plane. Occasionally, it is necessary to know the degree of uniformity vertically-that is, on horizontal planes at different elevations directly below the fixtures. Four different lighting situations are normally encountered.

They are point sources, such as point source downlights; line sources, such as continuous row fluorescent fixtures; infinite sources, such as luminous ceilings-whether trans-illuminated or indirect; and parabolic reflector beams, such as from parabolic aluminized reflector (PAR) lamps.

The vertical uniformity of each type is shown graphically in FIG. 18.

FIG. 18 Variation of illuminance vertically, directly below the fixtures, for different source types. (a, b) Illuminance directly below the fixture varies inversely with the square of the distance for a point source and inversely with the distance for a line source. (c, d) Illuminance remains constant at all distances from either an infinite (or nearly) source or a parabolic reflector.

FIG. 19 Mounting height of fixtures may be lower in a small room than in a large room because of the illusion of lowness created in a large room.

FIG. 20 Coefficient of utilization (CU; lighting system efficiency) as a function of pendant length for various distributions.

With a substantial downward component, as in direct-indirect or general diffuse lighting (a), system efficiency rises slowly as the fixture descends (pendant length increases). Maximum differentials occur in small rooms and can reach 20%. Where the ceiling is the light source, as in indirect and semi-indirect systems (b), the pendant length does not change the room illumination. This curve can be used to estimate CU for indirect and semi-indirect luminaries in the absence of manufacturer's data.

6. LUMINAIRE MOUNTING HEIGHT

The mounting height of luminaires is normally established before their spacing. In arriving at a mounting height for fixtures with an upward component, a balance must be struck between low mounting, which controls ceiling brightness and gives good utilization of light, and a reluctance to dominate an area, particularly a large room, by using such a low mounting height that the apparent ceiling height is affected (FIG. 19). General rules for mounting height are:

1. Indirect and semi-indirect luminaires should normally be suspended no less than 18 in. (460 mm) from the ceiling and preferably 24 to 36 in. (610 to 915 mm). Single-lamp luminaires with a very wide distribution (inverted batwing) may be suspended as little as 12 in. (305 mm) from the ceiling. Manufacturers' recommendations should be sought on this point.

2. Direct-indirect and semi-direct fluorescent fixtures should be suspended not less than 12 in. (305 mm) for two-lamp units and 18 in. (460 mm) for three- and four-lamp units.

The effect of pendant length on the coefficient of utilization (efficiency; see Section 15.11) is given in FIG. 20.

7. LIGHTING FIXTURES

Before proceeding further with design, we will discuss the principal item of lighting hardware: the luminaire itself. This section and the sections that follow cover luminaire construction, installation, and appraisal. The architect should consider that electric lighting fixtures constitute 25% to 30% of the electrical budget or 4% to 5% of the overall building budget. Because the difference between a quality unit and an inferior one is often not readily visible to the casual observer, particular care must be taken in the specification of lighting fixtures and in the examination of shop drawings and samples. All fixtures, if applied properly, give a sufficient quantity of light, but only a good unit combines quantity with good quality, ease of installation, ease of maintenance, and indefinite life. In addition, installation must be proper to ensure mechanical rigidity and safety, electrical safety, freedom from excessive temperatures, and requisite accessibility of component parts and of the fixture outlet box. The following material is a combination of National Electrical Code (NEC) minimum requirements and factors beyond these minima that the authors have found important.

8. LIGHTING FIXTURE CONSTRUCTION

1. All fixtures should be wired and constructed to comply with local codes, NEC (Article 410), and the Underwriters Laboratories (UL) Standard for Luminaires, and should bear the UL label where label service is available.

Reflector Luminaire Manufacturers (RLM) standards should be adhered to for all porcelain-enameled fixtures.

2. Fixtures should generally be constructed of 20-gauge (0.0359-in. [0.9 mm])-thick steel minimum. Cast portions of fixtures should be no less than 1/16 in. thick.

3. All metals should be coated. The final coat should be a baked-enamel white paint of at least 85% reflectance, except for anodized, aluminum, or silvered surfaces.

4. No point on the outside surface of any fixture should exceed 90ºC after installation and on continuous operation. For an exception, see NEC Article 410 M.

5. Each fixture should be identified by a label carrying the manufacturer's name and address and the fixture catalog number.

6. Glass diffuser panels in fluorescent fixtures should be mounted in a metal frame. Plastic diffusers should be suitably hinged. "Lay-in" plastic diffusers should not be used.

7. Plastic diffusers should be of the slow-burning or self-extinguishing type with a low smoke density rating and low heat-distortion temperatures. The latter should be low enough so that the plastic diffuser distorts sufficiently to drop out of the fixture before reaching ignition temperature.

8. It is imperative that plastics used in air handling fixtures be of the noncombustible, low-smoke-density type. These requirements also apply to other nonmetallic components of such fixtures.

9. All plastic diffusers should be clearly marked with their composition material, trade name, and manufacturer's name and identification number. Results of ASTM combustion tests should be submitted with fixture shop drawings. The characteristics of many plastic diffusers change radically with age and expo sure to UV radiation.

A brief survey of the most common standard transparent and translucent lighting fix ture diffusers follows.

Glass. Transparent to translucent, scratch resistant, easily cleaned, nonflammable, available in all grades of impact resistance, non-yellowing, usable with all sources, unaffected by and effective in blocking UV, readily formed into desired patterns. Heavy and expensive.

Acrylic (Plexiglas). Clear to translucent, easily scratched, slow burning, low impact resistance, resistant to yellowing, and available in a special low-yellowing composition (UV grade). Usable indoors and outdoors. Not usable at the elevated temperatures (>90ºC [194ºF]) found in some HID applications. Does not readily embrittle, warp, or craze. Molds well.

Polycarbonate. Initially very clear and highly impact resistant, but with a tendency to opacity and strength loss with age. Good scratch resistance and excellent thermal resistance. Usable indoors and outdoors with all sources. Readily molded to prismatic forms. Self-extinguishing (burning rate). Expensive.

Polystyrene. Usable only indoors because of rapid yellowing when exposed to exterior UV radiation. Slow burning, but smoke generation properties problematic with some fire codes. Not usable in the long term because of discoloration, particularly with UV-producing sources. Good thermal resistance. Readily molded. Cheap. Not scratch resistant.

As can readily be seen, no ideal diffuser material exists. A designer must select the material that best suits the use and budget, considering both initial and replacement costs as well as long-term optical properties.

10. Ballasts should be mounted in fixtures with captive screws on the fixture body to allow ballast replacement without fixture removal.

11. All fixtures mounted outdoors, whether under canopies or directly exposed to the weather, should be constructed of appropriate weather resistant materials and finishes, including gasketing to prevent entrance of water into wiring, and should be marked by the manufacturer "Suitable for Outdoor Use."

9. LIGHTING FIXTURE STRUCTURAL SUPPORT

Although some codes allow fluorescent fixtures weighing less than 40 lb (18 kg) to be mounted directly on the horizontal metal members of hung-ceiling systems, experience has shown that vibration, member deflection, routine maintenance operation on equipment in hung ceilings, and poor workmanship can cause such fixtures to fall, endangering life. It is therefore strongly recommended that all fixtures-surface, pendant, or recessed-whether mounted individually or in rows, be supported from the ceiling system sup port (purlins) or directly from the building structure, but in no case by the ceiling system itself.

This is particularly important in the case of an exposed "Z" spline ceiling system.

10. LIGHTING FIXTURE APPRAISAL

The intense competition in the lighting products field necessitates close scrutiny of the characteristics of luminaires and all accessories. To compare the relative merits of similar lighting fixtures manufactured by different companies, complete test data plus a sample in a regular shipping carton from a normal manufacturing run are needed.

The following list should be used as a basic guide, with additional items added according to job requirements:

1. Photometric and design data. Manufacturers should furnish complete test data, including candlepower distribution curve(s), coefficients of utilization, wall and ceiling luminance coefficients, luminance data from 45º to 85º, a table of VCP, energy data including LER (see Section 15.12), and recommended S/MH (SC). These data should come from a reliable independent testing laboratory, not from the manufacturer's test facilities. In addition, many manufacturers either regularly publish or make available on request various design aids such as isolux (isocandle) curves and point-by-point computer printouts for different layouts.

2. Construction and installation. The designer should check the sample for workmanship; rigidity; quality of materials and finish; and ease of installation, wiring, and leveling. Installation instruction sheets should be sufficiently detailed. Results of actual operating tempera ture tests in various installation modes should be included. Air-handling fixtures should be furnished with heat-removal data, pressure drop curves, air-diffusion data, and noise criteria (NC) data for different airflow rates.

3. Maintenance. Luminaires should be simply and quickly re-lampable, resistant to dirt collection, and simple to clean. Replacement parts must be readily available.

11. LUMINAIRE-ROOM SYSTEM EFFICIENCY: COEFFICIENT OF UTILIZATION

Because of internal reflections inside a luminaire, some of the generated lumen output of the lamp is lost. The ratio of output lumens to lamp (input) lumens, expressed as a percentage, represents the luminous efficiency of the fixture. This characteristic has little meaning by itself, however, as the efficiency of a luminaire in doing a particular lighting job depends on the space in which it is used.

To illustrate, let us consider the case of a large room with a high, dark ceiling. If we were to use a high-efficiency (say, 80%) indirect lighting unit in such a room, most of the light directed upward would be lost (absorbed), and the actual illuminance on the working plane would be very low. If, however, the same room were illuminated with 50% efficiency direct lighting units utilizing the same wattage, the illuminance on the working plane would be considerably higher than in the first case.

Similarly, if we consider a small room with dark walls and ceiling, lighted alternatively by dif fuse lighting and direct lighting units of the same wattage and unit efficiency, the horizontal-plane illumination is higher for the direct units because of the large loss of the horizontal and upward components of the diffuse lighting on the walls and ceiling. Fixture efficiency alone is not sufficient; the overall luminous efficiency of a particular unit in a particular space is the required figure of merit. This number, inasmuch as it describes the utilization of the fixture output in a specific space, is known as the coefficient of utilization (CU). It is defined as the ratio between the lumens reaching the horizontal work plane and the generated lumens. As each luminaire has a different coefficient for every different space in which it is used, a system of standardization has evolved utilizing room cavities (explained later) of certain proportions and various surface reflectances. The fixture coefficients are then computed and tabulated as shown in TABLE 1. It should be emphasized that the figures given in this table are for generic fix ture types only; in an actual job, luminaire data as found in manufacturers' catalogs should be used.

To summarize, CU is a factor that combines fixture efficiency and distribution with room proportions, mounting height, and surface reflectances.

12. LUMINAIRE EFFICACY RATING

As a result of a U.S. Energy Policy Act (EPACT) man date calling for an industry-wide testing and information program designed to improve lighting fixture energy efficiency, a collaborative effort produced NEMA Standard LE5, Procedure for Determining Efficacy of Luminaires. Unlike the CU discussed in the previous section, which defines the illumination efficiency of a particular luminaire in a particular space, Standard LE5 determines the energy efficiency of the luminaire alone. Because this efficiency is expressed in lumens per watt (lumens output per watt input), it uses the same descriptive term used for light sources (i.e., efficacy). This metric takes into account all power used by a luminaire, including ballast, and includes the ballast factor, which is itself a ballast energy efficiency metric. The expression used to calculate the luminaire efficacy rating (LER) is LER photometric efficiency ballast factor

luminaire

=

× input watts

An LER metric applies to a specific type of fluorescent luminaire and is identified by an abbreviation as

FL = fluorescent lensed

FP = fluorescent parabolic

FW = fluorescent wraparound

FI = fluorescent industrial

FS = fluorescent strip light

Thus, the energy figures shown in FIG. 14 are specifically labeled "FP" to denote a fluorescent parabolic luminaire. This enables a designer to com pare LER figures for different fixtures on a common basis. This commonality also extends to ballast type (i.e., magnetic or electronic). In addition, Standard LE5 lists benchmark LER figures that are considered to represent an acceptable luminaire. An additional item of data useful in economic comparisons that is included in the NEMA standard is the yearly cost per 1000/lm, based on 3000 burning hours and $0.08 per kilowatt-hour (see FIG. 14). The actual cost is easily calculated from this figure.

The LER approach has been expanded beyond fluorescent fixtures to include "Commercial, Non residential Downlight Luminaires" (NEMA LE5A) and "High-Intensity Discharge Industrial Luminaires" (NEMA LE5B).

LIGHTING CONTROL

13. REQUIREMENT FOR LIGHTING CONTROL

The term lighting control means all the techniques by which a lighting system can be operated, and covers both manual and automatic controls. The control strategy must be decided on simultaneously with the lighting design because the control scheme must be appropriate to the light source. In turn, the system's accessories and arrangement depend on the control scheme. For instance, if dimming is decided on, using a fluorescent light source, then the range of dimming determines the type of bal last, the ballast switching points, and the degree of dimming flexibility.

The primary purposes of lighting control are flexibility and economy: flexibility to provide the modifications of luminances and patterns desired by the designer, and economy of both energy resources and monetary resources (see Appendix I for treatment of economic analyses). A properly designed lighting control system can reduce energy usage up to 60% over a simple on-off system installation. In addition, financial operating economies result from

• Reduced energy use

• Reduced air-conditioning costs as a result of lower lighting waste heat

• Longer lamp and ballast life due to lower operating temperatures and lower output

• Lower labor costs due to control automation

As noted previously (Section 13), lighting in most new nonresidential construction is designed within the energy constraints of ANSI/ASHRAE/ IESNA Standard 90.1.

This standard has a system of lighting power credits for lighting control systems designed with automatic energy-conserving controls. These credits permit the effective connected load to be reduced by factors that increase with energy conservation effectiveness. Thus, for example, circuits with a simple on-off mode initiated by a daylight sensor have a smaller power credit than daylight sensing with continuous dimming, because the latter is more energy economical (but much more expensive initially). To avoid confusion, particularly in view of overlapping and sometimes inaccurate terminology, it is necessary to differentiate between control functions, control devices, and control systems.

For lighting, the only control functions are switching and dimming. The control devices are the means by which the switching and dimming functions are accomplished. They are numerous, ranging from a simple wall switch to time switches and dimmers of all sorts. Generally also included in this category are control initiation devices, such as occupancy sensors and photocells. The control system is the entire assembly of control and signal initiating equipment together with their interconnections. Included here also are microprocessors and programmable controllers. The system can be a stand-alone arrangement or, alternatively, as is the case in large facilities, it can be part of an energy management system ( EMS), a building automation system (BAS), or both. The difference in operation in these instances lies in the control algorithm, which is primarily energy-oriented for an EMS, and overall building function-oriented for a BAS. In the discussion that follows, the control criteria are energy conservation, cost reduction, and operating flexibility.

FIG. 21 Schematic diagram of switching arrangements to achieve multiple discreet lighting levels with three-lamp fluorescent lighting fixtures. Two-lamp ballasts are used in the interest of energy conservation and financial economy. In scheme (a) ballasts are switched, thus removing either one or two lamps from service. Finer control is achieved by using two-level ballasts or by introducing impedance (b) into the circuit, either in a block for an entire circuit or distributed in each fixture.

14. LIGHTING CONTROL: SWITCHING

There are two basic control functions-switching and dimming. Switching is an on-off function. By selecting the number of lighting elements to be switched in each switching action, the designer can establish the number of control levels. The more levels, the finer the control. Thus, in a space requiring several levels of uniform illumination for different functions, the designer has many control alternatives. He or she can switch entire fixtures, but this adversely affects uniformity. Taking three-lamp fluorescent fixtures as an example, the designer can obtain better uniformity and four levels of illumination by switching the ballasts (assuming one two lamp and half of a two-lamp ballast per fixture):

All ballasts on 100% illumination Two-lamp ballast on 66% illumination Half of two-lamp ballast on 33% illumination All ballasts off 0% illumination (Magnetic ballasts have been preferred in this arrangement because a single electronic ballast is often used for up to four lamps as one means of obtaining its benefits while reducing its cost per lamp.) This type of switching has the advantage of light reduction in relatively small steps at low cost.

A typical arrangement is shown in FIG. 21. Use of split-wired two-lamp units is advantageous from the cost and energy viewpoints. Even more uniform light reduction and finer control are possible with two-level ballasts (at increased cost). There, each lamp remains lighted but at either full or half output. Thus, the designer could have 0%, 50%, and 100% levels or, by combining alternate bal last switching, with two-level ballasts could have 0%, 17%, 33%, 50%, 67%, 83%, and 100% levels.

However, if that degree of control is desirable, dimming is probably preferable. The choice depends on the type of space and on the situation economics, as discussed later. An alternative method of achieving lower lighting levels in discrete steps, by switching, is to introduce impedance into the lighting circuit. This acts to reduce circuit current and light output.

Such devices are readily available for control of fluorescent lamps (see FIG. 26).

Recognition of the fact that an increased number of control (switch) points makes possible finer control, and therefore energy conservation, led to requirements in ANSI/ASHRAE/IESNA Standard 90.1 relating to the number of control points in a space and their types. Many field studies have demonstrated conclusively that reliance on manual switching is not an effective long-term energy conservation strategy, regardless of the good intentions of the space occupants. Indeed, studies have shown that space "ownership" affects even the low level of conservation possible with manual switching (i.e., lighting in private offices, conference rooms, and small storage spaces may be switched off, whereas lighting in multi-occupancy and common use spaces such as libraries, large office spaces, break areas, and the like, may not). As a result, Standard 90.1 first specifies the minimum number of control points required in a space and then awards automatic switching or dimming a higher number of equivalent control points.

The basic requirement is for one control point for every 450 ft 2 (42 m^2) or fraction thereof of enclosed lighted space plus one control point for each task (or group of tasks) located in the space.

Thus, the classroom in Example 1 would require two control points for its 517 ft^2 (48 m^2) of area plus one control point for all similar tasks grouped in the room, for a total of three control points. This requirement could be met with three wall switches, or one wall switch and an occupancy sensor, or an automatic dimming system probably initiated by daylight sensors. The point is, of course, to encourage use of automatic controls, which, as has been pointed out repeatedly, is the only proven method of attaining significant energy conservation.

15. LIGHTING CONTROL: DIMMING

The techniques and equipment required for dimming each of the different light sources, as well as the effect on the color of the light produced and on the lamp, are discussed in Section 12. Figure 22 shows typical lumen output versus power input curves for common light sources. Note that for fluorescent lamps, even with conventional magnetic ballasts, dimming down to approximately 40% of output is possible without reducing efficacy. This desirable characteristic can be exploited in control schemes where it is desired to change light output gradually without sacrificing efficiency. Due to efficacy reduction below 40% output, an economical and efficient control scheme combines dimming and switching of multi-lamp fluorescent fixtures to yield an almost stepless output range of 13% to 100% output. Continuous dimming over a 10% to 100% range is practical with special individual magnetic dimming ballasts or with electronic ballasts. As discussed in Section 12, electronic ballasts are much more energy-efficient than conventional ones and must be considered for all new installations, dimmed or not. For retrofit work, silicon-controlled rectifier (SCR) or triac dimmers give excellent results with existing conventional core-and-coil ballasts down to 40% output.

FIG. 22 Typical dimming curves for generic light source types.

Note that fluorescent lighting is efficient and approximately linear down to 40% output. All other sources have reduced efficacy when dimmed.

16. LIGHTING CONTROL: CONTROL INITIATION

Control initiation is either manual, automatic, or a combined manual-automatic function. The last is usually in the form of an override function: manual override of an automated procedure to ensure adequate control for special or unusual situations and automatic reset of the manual function to re-establish normal or steady-state operation.

(a) Manual Control Initiation

Numerous studies dating back at least 50 years have indicated increased employee satisfaction when at least a degree of control of the working environment is in his or her hands. This satisfaction is frequently accompanied by increased work output, at least in the short term. Unfortunately, manual control of lighting levels has been demonstrated to be wasteful of energy due to the tendency to leave lights on at maximum level even when day light is supplying more than enough light or when leaving a room for an extended period. A modicum of energy conservation in the latter instance is possible with the installation of time-out switches in spaces normally used for short periods, such as supply closets (see Fig. 26.48). However, for nor mal working spaces, even installation of manual dimmers in private offices is not effective because of the need to go to the dimmer control location on the wall to readjust. Manual dimming in multiple occupancy spaces is effective only in creating personnel dissatisfaction and friction.

Modern electronics has made practical a remote-control dimming system that can control single or multiple luminaires, making it applicable to all occupancies. Figure 23 illustrates the use of the system in an open multiple-occupancy space. This enables individual workers to adjust the output of luminaire(s) closest to their work station without disturbing other employees. This adjustment, which is simply accomplished (see FIG. 23a) by remote control, can alleviate direct and/or reflected glare or can be a temporary expedient suitable to the task at hand, as, for instance, work with a VDT. Figure 23b-d shows the system components. Wiring of the system can be arranged so that a single receiver dimmer can control up to 20 ballasts or, conversely, multiple dimmers are connected on a single circuit. The latter arrangement would be used in wiring a group of small single- or double occupancy rooms. Such rooms are frequently wired with a wall-mounted dimmer, which can then be activated by the remote control, as shown in FIG. 24.

(b) Automatic Control Initiation

Automatic controls are of two types: an open-circuit type and a closed-loop feedback type, also known as static and dynamic control, respectively. The former initiates a control function that is independent of the actual lighting situation, whereas the latter reacts to the condition of the lighting situation it controls via a feedback loop.

FIG. 23 (a, b) Ceiling-mounted fluorescent fixtures within an 8-ft (2.4-m) radius of the optical sensor can be dimmed by remote control and restored almost instantaneously. (c) The handheld IR remote control device (shown with a permanently mounted holster and a keeper-cord) arranged for continuous dimming and for dimming override for maximum or minimum light. (d) High-frequency electronic dimming ballast for one to four lamps can provide full-range dimming down to 1-5% of output or down to 10% of output, depending on the type. (e) The IR-activated dimmer/receiver, which fits into a 4-in. (100-mm) square standard outlet box, can be mounted inside the luminaire as in (a) or adjacent to it. The optical tube that carries the IR signal must be exposed on the ceiling. It may be mounted on the underside of an open luminaire or close by. (Lutron Electronics Co.)

1. Static control. The most common type of open-circuit lighting control is the programmable time-base controller. These vary from small, relatively simple units that replace wall switches and fit into a common device box to the more sophisticated units shown in Figs. 26.13 through 26.15.

These devices are available in myriad designs and capacities, but all perform the same basic function-(remote) control of loads and circuits on a preprogrammed time basis. The programming, in turn, is determined after analysis of operating schedules, task requirements, and field conditions.

With "tight" programming, energy savings of up to 50% over an uncontrolled installation are possible.

Because these devices act only on a time base and are insensitive to actual field conditions, an override feature must be incorporated to permit accommodation of special conditions. Thus, if a timer is set to shut off a row of fixtures adjacent to windows between 10:00 A.M. and 3:00 P.M., local override must be provided to accommodate dark rainy days and the like. Similarly, if lighting is shut off during nonworking hours, provision must be made for persons working overtime. The over ride arrangement can be entirely local, in which case it may lead to energy waste because it depends on local cancellation; it can be local with time-out, which can be a nuisance to a person working for an extended period; or it can incorporate an override feedback link to the controller, usually operated by telephone lines. In general, programmable time controls are best applied to facilities with regular, repetitive schedules and few exceptional situations.

2. Dynamic control. The second type of automatic control initiation responds to sensor indicated field conditions via an information feed back loop. It is frequently referred to as dynamic control because the initiation of a control function depends not on a fixed programmed parameter such as time, but on real-time field parameters (i.e., as measured at that instant). In the case of lighting control, these parameters may be ambient illuminance, time, system kW demand, kWh usage in a time period, or space occupancy, singly or in combi nation, depending on the programming algorithm (i.e., how the controller's microprocessor has been programmed). The control device in its entirety is called a programmable controller (see Section 26.16), which, in combination with the field sensors and the interconnecting wiring, constitute the control system. Some systems are wireless, using high frequency signals impressed on the power wiring system to transmit control signals. This arrangement is known as a power line carrier (PLC) system and is discussed in Section 26.34. Other systems are completely wireless, using radio frequencies and a system of wireless transmitters and receivers.

In addition to its CPU (microprocessor), a programmable controller contains input/output (I/O) interfaces, memory, and means for programming (and reprogramming). An operational block diagram is shown in Fig. 26.16. The controller accepts not only the usual time-based signal function, but also information (feedback) from field devices via its I/O device. It then "processes" the signal in its CPU, which consists of logic and storage memory, and sends out a resultant processed control signal.

Large lighting control systems use a computer in lieu of a programmable controller and usually have additional control facilities, such as telephone interfaces and local control/relay/switching centers, which process local sensor input and control local lighting blocks (see Figs. 15.25 and 26.56). Identical systems are used for building automation (Section 30), HVAC control, energy management, and the like, the difference being the types of sensing devices and the control algorithm.

FIG. 24 Remote controller may be used to operate a wall mounted dimmer.

FIG. 25 Schematic arrangement of a large lighting control system. The controller can schedule and supervise thousands of control points via the local control panels. The local control panel accepts coded commands from the controller and operates individual devices and circuits. It also accepts local control signals from manual (switches) and automatic devices. Override is provided via a telephone command to the controller, and all functions, including overrides, appear as hard copy on the printer.

FIG. 26 This type of compact solid-state electronic circuit reduces ballast power input by approximately 30% and lamp out put by about 28%, resulting in a net gain in efficacy. The power factor remains above 90%. Similar units are available for larger power decrease. They can be mounted on the lamp end or in the fixture channel, as shown. An ancillary benefit is cooler ballast operation, resulting in extended life.

17. LIGHTING CONTROL STRATEGY

It is apparent that a good lighting control system varies the lighting supplied to match the lighting required, as the requirement varies; thus, overlighting and under-lighting are avoided. In addition, the control system must be capable of permitting initial adjustments and of accommodating external, non lighting-connected constraints such as commands from a peak-demand controller. The common lighting system situations addressed by the control system follow.

(a) System "Tuning"

In every lighting installation there is a difference between the design intent and the field result. This is due to assumptions and imprecision in calculation, differences between specified and installed equipment, equipment location changes, and so on. The responsible lighting designer "tunes" the lighting system in the field to attain the intended design. This usually means reducing lighting levels in nontask areas because spill light is frequently sufficient for circulation, rough material handling, and the like. This tuning can result in an energy reduction of 20% to 30%, depending on the control technique.

The smaller the group of light sources con trolled, the more accurate the tuning and the larger the energy saving proportionally. Lighting system retuning is also required when the function of an entire space is changed, or when a single area is altered by a furniture move or by a task change.

An ancillary benefit of field tuning is glare reduction, which frequently improves task visibility dramatically. Tuning is a one-step function in the sense that, once accomplished, it does not require change unless the space function changes. There fore, it should also be reversible to accommodate such changes. Tuning should not be confused with lumen maintenance, described in Section 15.17d, which is a control strategy designed to compensate for normal system output decline and its corollary-initial system overdesign.

As stated, the tuning action most often required is a reduction in illuminance. This can be accomplished by:

1. Making appropriate field modifications to adjust able fixtures (aiming, lamp position, etc.).

2. Replacing lamps with others of lower watt ages and replacing fluorescent tubes with low wattage tubes. These changes reduce light out put reversibly.

3. Replacing fluorescent ballasts with low-current ballasts, thereby lowering output.

4. Adding current-limiting impedances to luminaires or lighting circuits, as mentioned in Section 15.14(a) (FIG. 26).

5. Ballast switching or use of multilevel ballasts.

6. Dimming by adjustment of a potentiometer at individual fixtures so equipped.

7. Replacing standard wall switches with time out units, programmable units, or dimmer units.

FIG. 27 Actual plot of energy usage on a typical floor of a high-rise building in New York City, with three degrees of lighting control. Note the importance of override by occupants when tight scheduling is used. Override was provided on a 1000-ft^2 (93- m^2 ) zone basis.

(b) Variable Time Schedule

No normal task area has a constant 24-hour, 365-day lighting requirement. In commercial and industrial spaces, work areas have regularly scheduled periods during which task lighting is not required. These include coffee and lunch breaks, cleaning periods, shift changes, and unoccupied periods. Programmed time controls can readily save 10% to 25% of the energy use compared to relying on occupants to manually operate controls. The action of such a controller in an actual installation is shown in FIG. 27. "Tight" scheduling took account of lunch hour and provided lighting only in restricted areas being cleaned rather than whole floors. The payback period for the investment in control equipment varies between 1½ and 5 years. Note that static control that is insensitive to actual field conditions has only limited ability to conserve energy.

(c) Occupancy Sensing

Within a normal 9:00 A.M. to 5:00 P.M. working schedule, offices in commercial spaces are unoccupied for 30% to 60% of the time. The reasons are manifold-coffee breaks, conferences, work assignments, illness, vacations, and reassignment to a different work location are a few. Occupancy sensors can operate relays to turn off lights after a preset minimum period of about 10 minutes or can dim the light level to a minimum in areas such as corridors, which always require some light. (They can also turn off other energy consumers such as fan-coil units, air conditioners, and fans.) Reestablishment of the original lighting level can be instantaneous, delayed, or manual on the action of the occupant. Another useful function of an occupancy sensor is to provide an automatic override in schedule systems, thus both relieving the occupant of the necessity of using a manual override and limiting the energy use to actual occupancy time. It can also light the occupant's way into a space and shut off the system after he or she has left.

Occupancy sensors that react to a human presence are of three types-passive infrared (IR), ultrasonic, and a hybrid of both technologies. The IR sensor (FIG. 28) reacts to the motion of a heat source within its range. It operates by creating a pattern of beams, and alarms when a heat source (such as a person) moves from one beam to another. It will not alarm to a stationary heat source. Although the IR sensor is quite sensitive, it has several disadvantages:

• Small movements may not be detected, as they may not cross from one beam (zone) to another.

As a result, a person sitting quietly may not be detected and the sensor may shut down/off the lights.

• Very slow movements may not be detected even when they cross zones.

• The IR detector must "see" the heat source.

Therefore, a heat source blocked by furniture will not be detected.

• The beams have a discrete width and depth.

Therefore, there may be "dead" spots under the beams if the units are not carefully selected to have adequate multilevel beams or properly located to give the desired coverage.

Ultrasonic sensors emit energy in the 25- to 40-kHz range, which is well above the range of human hearing. The waves immediately fill a space by reflecting and re-reflecting off all hard surfaces, establishing a pattern that is detected by the sensor. Any movement within the space disturbs this pattern and is immediately detected by the sensor. Ultrasonic sensors have distinct advantages over IR sensors in that they do not require a direct line-of-sight expo sure to the movement, and they detect small movements. The latter characteristic, however, is also a disadvantage inasmuch as movement of curtains and even air movement can trigger a sensor. It is frequently necessary to adjust (reduce) a sensor's sensitivity to avoid false sensing. Unfortunately, this also decreases its coverage.

Hybrid (dual-technology) sensors combine the characteristics of both sensors by usually requiring both sensors to react to turn lights on, but once on, a reaction in either sensor keeps the lights on. In addition, sophisticated electronic circuitry "learns" a space's occupancy patterns and is programmed to react accordingly.

Placement of sensors is very important and should be tested before final installation. Studies have shown that reducing the minimum on period below 10 minutes is counterproductive and frequently causes space occupants to shut off the sensors. Depending on their type and mounting position, sensors cover a maximum area of 250 to 1000 ft^2 (23 to 93 m^2) per unit. The payback period for this equipment runs between 6 months and 3 years, depending on the type of space and the degree of control already existing. Occupancy sensors are best applied in areas that are divided into individual rooms and work spaces. Sensors can be wall or ceiling mounted or mounted on a wall-outlet box in a combined sensor/wall-switch configuration. A few designs are shown in FIG. 29. (See Section 30.2, which discusses the security application of motion detectors and illustrates their operation.) ANSI/ASHRAE/IESNA Standard 90.1 recognizes the great effectiveness of these devices by giving both connected power credit and a high control point equivalency.

FIG. 28 Passive infrared (PIR) occupancy sensors. All sensors can be equipped with an adjustable override to prevent turning on lights when ambient light is sufficient. All units have adjustable delayed-off timing and a flashing LED that indicates sensor operation. (a) Flush-mounted ceiling unit. Note the 360º circular pattern of the lens, which indicates omnidirectional coverage. The unit is approximately 4 in. (100 mm) in diameter. (b) Flush-mounted combination wall switch and PIR occupancy sensor. Operation of the switch overrides the sensor function. (c) Surface-mounted wall PIR occupancy sensor. Note the semicircular shape of the lens, which indicates wide horizontal coverage.

FIG. 29 Ultrasonic occupancy detectors. (a) Ceiling-mounted sensors, bidirectional on the left and unidirectional on the right. (b) Sensors in various designs, intended for differing applications. Left to right: Ceiling mounted two-way sensor intended for large rooms of up to 2800 ft^2 (260 m^2 ). Dual-function sensor and wall switch designed for rooms of up to 300 ft^2 (28 m^2 ); wall mounted on a single- or double-gang wall box. One-way ceiling-mounted sensor designed for use in rooms of up to approximately 1250 ft^2 (116 m^2 ).

(d) Lumen Maintenance

Referring to Section 15.20, we see that in order to maintain a minimum lighting level to the end of a maintenance period, we must deliberately overdesign initially. The extent of the overdesign is the reciprocal of the light loss factor (LLF). With an average LLF of 0.60, this initial overdesign amounts to 1/0.60, or 66%. Assuming a linear light falloff over a 2-year maintenance period, this overdesign results in an average of 33% annual energy waste. (In the next 2-year maintenance cycle, the actual overdesign is slightly less due to a small amount of unrecoverable loss.) Because the light depreciation is a continuous and very gradual process over the maintenance period, the most appropriate control strategy is one that reduces the initial overlighting by the required amount (as measured in the field) and gradually restores it as the system ages. This control strategy, known as lumen maintenance, is accomplished by a dimming system operating in conjunction with local light sensors (photocells). The photocells mea sure ambient light, and in response to their signals the controller(s) operates the dimming units to raise (or lower) the light output. Depending on the size of the installation, the dimmers can either be dispersed or centralized. The modulating action of such a sys tem over the maintenance period is shown in FIG. 30. This strategy, in a purely manual mode (periodic maintenance, initial light reduction in accordance with the length of the maintenance period), gives only about a 10% energy reduction, whereas in an automated system almost all of the 33% annual energy waste can be eliminated.

In a new installation, the choice of whether to use electronic ballasts and full-range dimming, conventional ballasts and partial dimming, or a system of multilevel switching is an economic one and depends on many factors, one of the most important of which is the cost of energy. Often a combined system is advisable. In interior zones, initial lighting reduction does not exceed 30% to 40%, and full range dimming is not required. In perimeter zones, daylight often provides all of the required light, and either full-range dimming, dimming plus switching, or a multilevel switching system is required (see Section 15.14a and the following discussion of daylight compensation). The payback period for a lumen maintenance installation of this type varies from 1 to 5 years. Shorter payback periods can be obtained by using multilevel switching rather than dimming because of its lower first cost.

FIG. 30 Graph of energy use by a system that reduces the initial lighting level to compensate for initial overdesign and gradually increases the level as system output depreciates. In subsequent cycles the energy savings are reduced slightly because of unrecoverable output loss.

FIG. 31 Typical graph of energy savings with daylight compensating lighting control. A full-range dimming system is more effective than one that dims down to only 40%, because daylight often supplies most of the required lighting.

An economic analysis is required to determine whether the additional cost of such a system is justified.

FIG. 32 Typical building plan showing approximate daylight perimeter zones. Exact delineation of zones depends upon latitude, climate, window design, and cost of electric energy.

An additional favorable effect of initial light reduction is the lengthening of effective lamp life and a reduction in the rate of its lumen depreciation. When a lamp is operated at the rated voltage, its lumen output drops during its life (according to the type of lamp). However, if lamps are operated at reduced output, as is the case if lamps are dimmed to compensate for initial overlighting, the lamp life cycle is greatly extended, lumen depreciation is reduced, and lamp energy consumption is linearly reduced.

Typical life extension (to economic replacement) figures are:

Fluorescent 80%

Metal-halide 40%

High-pressure sodium 20%

These figures are for interior zones; for perimeter zones with ambient daylight compensation, they are higher.

(e) Daylight Compensation

It should be apparent that a control system arranged for continuous ambient light compensation, as described in the preceding subsection, is automatically also arranged to compensate for ambient daylight. The difference is that ambient compensation for lumen maintenance due to light loss factors is a very gradual process of increasing output, whereas daylight compensation can be a minute-by-minute variation and generally in the direction of decreased electric lighting. Because of these possible rapid variations, switching systems are undesirable, as the constant on-off or level switching of lamps can be very annoying to occupants and deleterious to lamps. Automatic dimming is the system of choice.

Recognizing that in perimeter areas daylight often supplies all the required light, the system for fluorescent installations must be either full-range dimming with electronic ballasts or partial dimming with magnetic or electronic ballasts. Figure 31 shows the action of both types of dimming systems in a fluorescent installation with daylight compensation. The crucial design element in a daylight compensating system is the establishment of zone areas. Depending on the latitude and climate, the southern and possibly the eastern and western exposures can have an interior (second) perimeter zone that receives sufficient daylight for a large enough portion of the year to economically war rant dimming. The northern exposure has only a narrow perimeter zone (FIG. 32). As a starting point, the size of the zones is established by deter mining the maximum room depth that receives at least half of its illuminance from daylight for several hours a day. A computer daylight and dimming cost study with perimeter zone depth as a variable is an effective approach. Several such programs are available.

FIG. 33 Typical daylight factor curve plotted on a room section with one side window. The photocell control technique for daylight compensation described in the text assumes a constant daylight factor (DF) distribution indoors-that is, fixed luminance sky and no direct sun.

Placement of the control photocells depends on the control system. Where daylight compensation is desired in conjunction with lumen maintenance (see the preceding subsection), area photocells are desirable, as they give a feedback control signal for the specific area involved. Alternatively, a daylight-factor map of a space can be made, prefer ably after the installation is complete and the furniture in place, and photocells located at one point, based on this map (FIG. 33). We know from our study of daylight in Section 14 that the daylight factor at any inside point is constant. Therefore, by measuring actual daylight level at a point either immediately inside or outside the window, shielded from direct sun and from inside electric lighting, we can relate it to any area in the room by the ratio of daylight factors and establish a switching level at which the lighting for that inside area is switched, partially or fully.

An example should make this clear. In FIG. 33, the photocells are mounted at point A. The ratio of daylight between points A and B is the ratio of their daylight factors-that is, 20/10, or 2.0. Therefore, if 500 lux of daylight is required at point B before switching lights in that area, twice that amount (2.0 × 500 lux) or 1000 lux is required at point A. Similarly, a 500-lux requirement at point C corresponds to (20/4)(500) or 2500 lux at point A. Therefore, switching can be initiated by two single-level photocells or a multilevel unit at point A, with settings at 1000 and 2500 lux. Other switching arrangements can be made on the same principle. Dimming initiation can also be arranged in this fashion.

Daylight compensation can reduce energy use in perimeter areas by up to 60%, depending on latitude, climate, depth of perimeter zone, hours of building use, initial power density, and so on. The amount saved for the entire building obviously depends on the building's configuration-that is, the ratio of perimeter to total area. Payback time is usually in the range of 3 months to 3 years.

In summary, a well-designed lighting control system can provide energy conservation of up to 60%, extremely long lamp life, reduced cooling costs, extended ballast life, and reduced maintenance costs. When all these factors are considered, the investment payback period is always short and therefore financially attractive.

An aspect of a centralized lighting control system not discussed previously, because it is not directly concerned with lighting, is its use in connection with peak demand reduction (see Section 25). When interconnected with a demand controller, this approach acts to reduce the electric lighting load in accordance with a predetermined preferential-load schedule and can achieve significant savings.

DETAILED DESIGN PROCEDURES

18. CALCULATION OF AVERAGE ILLUMINANCE

Once a luminaire has been selected on the basis of the foregoing criteria, it remains only to calculate the number of such fixtures required in each space, for uniform general illuminance, and to arrange them properly. Although a number of calculation methods are available, the lumen (flux) method is simplest and most applicable to our need for area lighting calculations. Illuminance calculation from point, line, or area sources is covered in Sections 15.27 to 15.30. Luminance (photometric brightness) calculations are covered in Section 15.32.

Before beginning a detailed description of the zonal cavity calculation method, a general comment on precision is in order. The precision of any calculation should not exceed either the accuracy required or the precision of available data. Thus, there is no point in working to three decimal places if a ±10% rounding of the result is common and acceptable, or if the data available are accurate to only one decimal place. As the reader will see, the lumen (lighting flux) method of average illuminance calculation is replete with assumptions and estimates. Among these are:

1. It is assumed that the space is empty. This is not normally the case.

2. It is assumed that all surfaces are perfect diffusers. This is not the case.

3. All surface reflectances are estimates, ±10%.

4. Maintenance conditions are estimates, at best ±10%.

5. No allowance is made for deviation of the performance of an individual product from its specification.

Any attempt to account accurately for these approximations would enormously complicate the calculation and would serve no useful purpose for this type of average illuminance calculation. For this reason, the procedure presented here introduces some approximations (which we feel are well justified) in the interest of simplification. These approximations are noted wherever used.

19. CALCULATION OF HORIZONTAL ILLUMINANCE BY THE LUMEN (FLUX) METHOD

The lumen method of calculation is a procedure for determining the average maintained illuminance (in footcandles or lux) on the working plane in a room.

The method presupposes that luminaires will be spaced so that uniformity of illumination is provided in order that an average calculation will have validity.

The method is based on the definition of 1 lux of illuminance as 1 lumen incident on 1 square meter of area (1 footcandle incident on 1 square foot), that is, lux or fc lumens a ream or ft

As explained previously, the ratio between the lumens reaching the working plane in a specific space and the lumens generated is the coefficient of utilization, CU.

Or lumens on the working plane = lamp lumens × CU

Therefore, illluance lamp lumens CU area min E =

The coefficient CU is selected from tables provided by the manufacturer of a selected luminaire by a technique known as the zonal cavity method . In the absence of specific CU data, an approximation can be made by using the generic fixture types in TABLE 1.

The illuminance figure so calculated is the initial average illuminance. This initial level is reduced by the effect of temperature and voltage variations, dirt accumulation on luminaires and room surfaces, lamp output depreciation, and maintenance conditions. All of these effects are cumulatively referred to as the light loss factor (LLF):

maintained E = initial E × LLF

(This factor was previously termed the maintenance factor, MF.) The procedure required to arrive at this factor is explained in the following section.

Our final expression for maintained illuminance E as calculated by the lumen method is, therefore,

(1)

where E is lux if area is expressed in square meters, or footcandles if the area is in square feet.

Lamp lumens is the total within the space and is equal to number of fixtures × lamps per fixture

× initial lumens per lamp The formula then becomes

(2)

or, conversely, solving for the number of luminaires required to achieve a target maintained illuminance E:

(3)

For large areas, a much more useful figure is the area illuminated per luminaire:

(4)

For instance, it is much more convenient to know that to maintain, say, 60 fc within a space with a given luminaire, 70 ft^2 (6.5 m^2) per unit is appropriate than it is to know that for an 18,000-ft^2 (1672-m^2) floor, 257 fixtures are necessary. The former figure allows us to establish a pattern, say 7 × 10; the latter figure is too large to be immediately useful. Therefore, for rooms requiring more than a small number of luminaires, the latter calculation should always be used.

20. CALCULATION OF LIGHT LOSS FACTOR

The light loss factor (LLF) is composed of elements that can be categorized as recoverable and nonrecoverable. The former can be improved by maintenance; the latter cannot. The total LLF is the product of all the individual factors. The overview that follows includes approximations. For more precise data, see the IESNA Lighting Handbook (2000).

Among the nonrecoverable loss factors are the following:

(a) Luminaire Ambient Temperature

Light output changes when a fluorescent fixture operates at other than its design temperature. With normal indoor installation use 1.0-that is, no depreciation. For other conditions, refer to technical data on the luminaire involved.

(b) Voltage When a lamp operates at the rated voltage, use 1.0. Details of source sensitivity to voltage are given in Section 12.

(c) Luminaire Surface Depreciation

This factor is proportional to age and depends upon the type of surface involved. The designer must estimate this factor based on knowledge of the luminaire materials.

(d) Components

Losses due to components include ballast factor, ballast-lamp photometric factor, equipment operating factor, and lamp position (tilt) factor. Air troffers also introduce a thermal factor. In the absence of specific data for component factors, use a total of 0.92.

In the absence of reliable data for any of the foregoing nonrecoverable factors, use an overall factor representing the product a × b × c × d of 0.88.

The factors that follow are recoverable-that is, they can be returned to their initial state by maintenance.

(e) Room Surface Dirt

This factor is self-explanatory. Obviously, lighting approaches that depend heavily on surface reflections, such as indirect systems, are more seriously affected than systems that deliver most of their useful light directly. Assuming a 24-month cleaning cycle and normal conditions of cleanliness, use the appropriate factor in the following list. Alter it for other conditions such as infrequent maintenance and unusual cleanliness or dirtiness.

Direct lighting: 0.92 ± 5%

Semi-direct lighting: 0.87 ± 8%

Direct-indirect lighting: 0.82 ± 10%

Semi-indirect lighting: 0.77 ± 12%

Indirect lighting: 0.72 ± 17%

(f) Lamp Lumen Depreciation

This factor depends upon the type of lamp and the replacement schedule. Use the following when exact data are unavailable:

(g) Burnouts

This factor accounts for lamps that produce no output but have not been replaced. It depends upon maintenance schedules and method of replacement. Use the following as a general rule:

Group replacement procedures: 1.0 Individual replacement on burnout: 0.95

FIG. 34 The LDD factor is determined from the category of luminaire (which is an indication of its proneness to dirt accumulation) plus knowledge of room ambient conditions.

(h) Luminaire Dirt Depreciation

The luminaire dirt depreciation (LDD) factor depends upon luminaire design, atmosphere conditions in the space, and maintenance schedule.

The luminaire maintenance category is obtained from the manufacturer's data or from TABLE 1.

The type of atmosphere is determined by considering the space involved. Assuming a 12-month cleaning schedule and normal room cleanliness, use the base number in FIG. 34 and change it to match the conditions of dirt and maintenance. The categories correspond to those used by the IESNA.

Total LLF is the product of all the depreciation factors:

For example, a fluorescent air troffer in a regularly maintained group lamp-replacement, air-conditioned office might typically have an LLF of 0.8.

The same fixture in the same office, but with walls and fixture cleaned only when burned-out lamps are replaced, would typically have an LLF of 0.55.

Thus, if in the first case the maintained illumination is E fc, in the second case it is 0.55/0.80 or 0.69 E fc, that is, a reduction of 31% as a result of poor maintenance. When a detailed determination of LLF is not possible, use the factors given in Section 15.22 (they are somewhat conservative).

21. DETERMINATION OF THE COEFFICIENT OF UTILIZATION BY THE ZONAL CAVITY METHOD

The coefficient of utilization (CU) connects a particular fixture to a particular space by relating the luminaire's light distribution characteristic to the room's size and its surface reflectances. To account for the luminaire's mounting height and its relationship to the working plane, the space is divided into three cavities: a ceiling cavity above the fix ture, a floor cavity below the working plane, and a room cavity between the two (FIG. 35). Given the surface reflectances, the effective reflectances of the floor and ceiling cavities can be obtained. With these, the CU can be selected from tables (either TABLE 1 or from manufacturer's data) and the lumen formula applied to arrive at average illuminance. A step-by-step explanation of the method plus illustrative examples demonstrate the procedure. The reader should follow the steps with the flowchart in FIG. 36 and the calculation form in FIG. 37 in hand.

FIG. 35 Room cavities as used in the zonal cavity method.

STEP 1. First, dimensional data are established. In offices, schools, and many other occupancies, the work plane is 30 in. (760 mm) above the finished floor (AFF). In drafting rooms it is 36 to 38 in. (915 to 965 mm); in shops, 42 to 48 in. (1066 to 1220 mm); in carpet stores and sail-cutting rooms at floor level. The three h terms are the heights of the various cavities. Also identify the initial reflectance of the room surfaces and fill in the sketch in FIG. 37. Utilize the reflectance closest to those given in TABLE 2.

STEP 2. See FIG. 37. This step involves determining the cavity ratios of the room by calculation. The basic expression for a cavity ratio (CR) is CR area of cavity wall area of work plane

(5)

In a rectangular space, the area of the cavity wall is h × (2l × 2w) or 2h(l + w); therefore, CR area of work plane

(6)

For other than rectangular rooms, the area can be calculated as required by geometry. For instance, in a circular room, the cavity wall area = h × 2pr and the work plane area is pr2. Thus,

(7)

For each of the cavities in a rectangular room we have:

(8)

(9)

(10)

For reference, because CR values for a space are related, once one has been determined, the others can be obtained by ratios

(11)

(13)

FIG. 36 Zonal cavity method flowchart.

STEP 3. See TABLE 2 and Figs. 36 and 37.

This step involves obtaining the effective ceiling cavity reflectance (?FC) from TABLE 2. Note that the wall reflectance remains as selected in step 1.

If the fixtures are surface-mounted or recessed, then CCR = 0 and ?CC = selected ceiling surface reflectance.

STEP 4. See TABLE 2 and Figs. 36 and 37.

This step involves obtaining the effective floor cavity reflectance ?FC, as in step 3 for ?CC. If the floor is the working plane, FCR = 0 and ?FC = selected floor surface reflectance.

STEP 5. Select the CU from the manufacturer's data. Note that interpolation may be necessary for CCR (?CC) if it is between the figures in the CU table.

See Example 1 in the next section. CU correction factors for ?FC other than 20% (the standard value in CU tables) are given in TABLE 3.

STEP 6. Calculate the illuminance and the number of fixtures or area per luminaire as in Section 15.19.

Illustrative examples and shortcut methods are demonstrated in the following section. CU coefficients are listed in TABLE 1 for generic luminaire types.

22. ZONAL CAVITY CALCULATIONS: ILLUSTRATIVE EXAMPLES

[...]

FIG. 39 Complete photometric data on a 2-ft (610-mm) square fixture with a 36-cell mirrored parabolic louver. The unit uses two T8 U-shaped 32-W fluorescent lamps. Note that the transverse distribution characteristic has a modified batwing shape typical of this type of louver. The louver's specular finish yields very low brightness at high angles, as can be seen from both the distribution curve and the luminance (cd/m^2 ) data.

FIG. 40 Sheet for illuminance calculation using an approximate zonal cavity method based on a method developed by B. F. Jones.

23. ZONAL CAVITY CALCULATION BY APPROXIMATION

Although the foregoing zonal cavity calculations are straightforward and essentially simple, they can become tedious if more than a few areas are involved.

Two alternatives exist: to utilize one of the many computer programs available or simply to shorten the calculations with reasonable approximations.

Computer assistance is discussed in Sections 15.30 and 15.31 and Appendix M. An effective computational method using approximations is demonstrated in this section, which is based upon a method developed by B. F. Jones.

Fill in the calculation sheet for illuminance calculation by approximation, as shown in Fig. 40. Assume that all rooms are square. To do this for a rectangle, take one-third of the difference in dimensions and add to the smaller dimension to obtain the equivalent width w. Then, for square rooms

[...]

Then, using the calculated side dimension of the square equivalent room, assume:

1. ?CC = 0.80 for a large room, that is, equal to or larger than 30 × 30 ft (10 × 10 m).

?CC = 0.70 for a medium room, that is, between 30 × 30 ft (10 × 10 m) and 12 × 12 ft (4 × 4 m).

?CC = 0.60 for a small room, that is, equal to or smaller than 12 × 12 ft (4 × 4 m).

2. Assume that ?FC = 0.20.

3. Assume that LLF = 0.65 for good conditions, 0.55 for average conditions, and 0.45 for poor conditions.

FIG. 38 Layout of pendant parabolic aluminum reflector luminaires in a typical classroom. The units have a modified batwing distribution in the crosswise direction, which mandates their being hung with their long axis parallel to the line of sight, as shown. Furthermore, the fixtures have lower brightness in that direction, that is, VCP for students in the head-up position is excellent. Note that the distance between the outside fixture row and the window wall is more than one-half of the side-to-side spacing because of daylight contribution during school hours, whereas the inside fixture row to inside wall spacing is less than one-half of the side-to-side spacing in order to maintain sufficient illuminance near that wall. The single fixture to the right of the teacher's desk (to the teacher's left) was so placed to avoid the veiling reflections that would result from a fixture directly above the desk and to take advantage of the fixture's transverse batwing distribution characteristic.

In conclusion, then, with respect to zonal cavity calculations, we can make the following statements:

1. For preliminary and routine calculations of rectangular rooms, with assumed reflectances, use the assumptions listed previously. A modified calculation form is provided in FIG. 40 to assist with this method.

2. For rooms where a high degree of accuracy is desired and actual reflectances are known, use the long method with visual interpolation.

3. For rooms of unusual shape or rooms with special conditions, such as coffered ceilings, mixed-material walls, and partial height partitions, use computer assistance.

4. For spaces in which a number of different solutions are to be tried, use a computer.

24. EFFECT OF CAVITY REFLECTANCES ON ILLUMINANCE

FIG. 41 Effect of surface reflectances on the CU of a luminaire with semi-indirect distribution. As expected, because the ceiling becomes the light source, its reflectance has the most pronounced effect. With this particular unit having a 25% downward component, the floor finish also has an appreciable effect, increasing the CU by an average of 10% for a 30% reflectance floor. The effect of wall reflectance naturally increases as rooms become smaller and the proportion of wall surface becomes larger. The change in CU between a 30% and a 50% reflectance wall varies from 15% for a 400-ft^2 (37-m^2 ) room to 5% for a 4000-ft^2 (372-m^2 ) room.

The reflectances of the various room cavities have a marked effect on the CU because of light reflections within the room. To demonstrate this graphically, we have plotted in Figs. 41, 42, and 43 the effect of varying cavity reflectances on the three principal types of fixture distribution: semi-indirect, direct-indirect, and direct-spread.

Note that as expected, ceiling cavity reflectance has the most pronounced effect with indirect fixtures and floor reflectance with direct units.

Because lighting costs amount to 3% to 5% of the total construction cost for many types of buildings such as offices, a 20% differential in lighting fixtures can amount to as much as 1% of the total cost of a facility. This amount would not only pay for the increased cost of higher-reflectance finishes and materials but would also reduce both initial and operating costs. These data clearly indicate the necessity for the lighting designer to have considerable influence on the selection of room materials and finishes, a situation that, unfortunately, does not usually occur.

25. MODULAR LIGHTING DESIGN

An increasingly large number of buildings are being designed around a modular system, resulting in a need for flexible lighting to fit the module utilized.

In such buildings, once the general lighting scheme and luminaire are established, it is convenient to draw a family of curves for the fixture chosen, thereby facilitating the utilization of the modular unit in various spaces. "Area" may readily be replaced with multiples of modular areas, as shown in FIG. 44.

26. CALCULATING ILLUMINANCE AT A POINT

The lumen (flux) method of horizontal illuminance calculation explained previously is appropriate for spaces in which illuminance is essentially uniform throughout. However, even in such a space, illuminance varies at least ±10% and, near columns, walls, windows, bookcases, and the like, consider ably more. Therefore, to answer the often asked question "How much light will I have on my desk?" the designer must turn to other methods. Three are available:

1. Calculation of illuminance at selected points by computer, as explained in Section 15.31

2. Utilization of one of the design aids explained in Section 15.27

3. Longhand calculation by one of the methods presented in Sections 15.28 through 15.30

These methods also yield results where the lumen method is simply inapplicable. Among such situations are layouts that are intentionally nonuniform; calculation of illuminance on planes other than horizontal (e.g., wallwashers); calculation of illuminance resulting from architectural lighting elements such as coves, valances, and the like; and illuminance calculations for nonstandard light sources for which CU data of the type given in TABLE 1 are not available.

FIG. 42 Effect of surface reflectances (a: floor, b: ceiling, c: wall) on the CU of a luminaire with direct-indirect distribution.

With this distribution, the effects of the ceiling and floor are most pronounced, with an appreciable wall effect only in small rooms.

FIG. 43 Effect of surface reflectances on the CU of a luminaire with direct (spread) distribution. Floor finish is most important, with wall reflectance important only in small rooms.

As these fixtures have no upward component, all ceiling illumination is derived from reflection. Thus, in a room with floor reflectance of less than 20%, ceiling finish has no effect on room illumination.

FIG. 44 Luminaire design chart. For frequently used fixtures, this type of chart gives an easy design figure for various size rooms. As seen from the ordinates, the figures can be translated into number of modules and watts per square foot.

27. DESIGN AIDS

By design aids we mean any of the various curves, charts, plots, or tables either prepared by the designer or made available by luminaire manufacturers, the purpose of which is to simplify and speed lighting design when using a particular lighting fixture. The reliability of data so obtained depends entirely on the manufacturer involved, and their use should be governed accordingly. More recently, it has become customary for major manufacturers to provide computer output charts and tables based on the designers' proposed layout(s). When using these, it behooves the designer to carefully study the data input to the computer program, as the fixture supplier is certainly not a disinterested party, and the program may be "weighted" accordingly. A brief description of common design aids follows.

(a) Isolux Charts

These charts, also called iso-footcandle charts, are based on the type traditionally supplied by manufacturers of outdoor lighting equipment, such as street lights and floodlights, but are equally applicable to interior lighting. Their use is illustrated in FIG. 45. The basic tool is an isolux diagram for a single luminaire. This is either calculated, measured from a full scale mockup (the most accurate if not the most practical method), or obtained from the manufacturer. Inasmuch as the relative positions of the source and illuminated point are reversible-that is, if a source at A causes illuminance E at point B, then the source at B will cause the same illuminance E at point A-placing the center of the isolux chart at the point in question permits direct reading of the illuminance contribution of every other luminaire. It then remains simply to sum the individual contributions to obtain the (scalar) illuminance at the desired point. An example is shown in FIG. 45.

FIG. 45 The ellipses represent isolux lines for a single luminaire at a given height above the work plane. They are centered on the point (the work area of a desk) for which the illuminance must be determined. The total illuminance at that point is the sum of the individual luminaire contributions. The center of the luminaire is the point of reference. Therefore, when two or more isolux lines pass through a fixture, its contribution is determined by the interpolated isolux line passing through its center. Note the symmetry around the vertical axis, necessitating a plot of only half of the ellipses.

FIG. 46 For downlights with symmetrical circular distribution, a "cone of light," as shown, can be drawn. The illuminance at varying distances on the beam centerline directly below the luminaire is given in the center column. A circle at the circumference of which the illuminance is half of this maximum is drawn at each distance from the downlight (2 ft, 4 ft, etc.). The numbers in the left column show the diameter of this (beam) circle.

FIG. 47 Addition of an interior reflector to the downlight of FIG. 46 converts it to a dual-purpose downlight and wallwash unit. The wall illuminances produced by multiple units spaced 3 ft (0.9 m) apart and ceiling-mounted 3 ft (0.9 m) from the wall are given in the chart. Similar charts are available for other luminaire spacings.

(b) Illuminance "Cone" Charts

See FIG. 46. When the light distribution of a direct downlight is symmetrical, as is generally the case, a cone can be drawn showing the illuminance directly under the fixture at various distances. The projected circles are defined by maximum illuminance at the center and half of this illuminance at the edge. This projected circle can be used in the same fashion as the isolux chart in the preceding section, except that only two values are given-that at the center and that on the circumference.

(c) Illuminance Tables and Charts

These take various forms but all give specific illuminance data, in numerical form, for specific points.

The values are obtained from a computer printout or an actual test. Figure 47 shows the illuminance pattern on a wall produced by the wallwasher version of the downlight shown in FIG. 46. The difference between the two fixtures is the addition of an interior reflector in the wallwasher version.

28. CALCULATING ILLUMINANCE FROM A POINT SOURCE

It is well to note at the outset that all of the following methods calculate illuminance at a point.

The answer to the query "How much light will I have on my desk from a luminaire at this location?" is arrived at by taking several points on the desk and calculating illuminance at each one. Shortcuts can be made for symmetry and so on. Most of these calculations are lengthy and are generally performed by a consulting engineer or lighting specialist rather than the architect. They are presented here as background material to describe the technical nature of lighting design.

The basis of point source calculations is the inverse square law developed in Section 11.12:

[...]

where fc, cp, and D are footcandle illuminance, candlepower intensity, and distance, respectively. Refer to FIG. 48. The horizontal illuminance at a point P as shown in FIG. 48 is...

(14)

...and the vertical illuminance at that same point is...

(15)

However, because cos ?? ==

H D R D and sin we have then at point P horizontal illuminance 2

(17)

Inasmuch as the candlepower intensity in the direction of point P is taken from a candlepower distribution curve, and ? is known, these expressions are readily usable. Very few commercial light sources are actually point sources. However, when the maxi mum dimension of the source is less than five times the distance to point P, the equations give satisfactory results. Note that these equations can be used to calculate and plot isolux diagrams for point sources of the type shown in Figs. 15.45 and 15.46.

FIG. 48 Relationship between intensity in candlepower (cp) and illuminance when the source can be considered a point source-that is, when the inverse square law applies. Source major dimension must not exceed 0.2D to be considered a point source. Measurement in feet yields fc; distances in meters yield lux.

FIG. 49 Typical candlepower distribution plot for use in inverse square law calculation.

29. CALCULATING ILLUMINANCE FROM LINEAR AND AREA SOURCES

When the source is too large to be considered a point source (the definition is relative and depends on the distance to the illuminated surface), it is referred to as either a linear source or an area source. The direct component of the illuminance at a point, resulting from such sources, can be calculated by manual graphical or analytical means, both of which are based on an assumed distribution, generally Lambertian (diffuse). Inasmuch as most lighting fixtures being applied today do not have Lambertian characteristics (e.g., parabolic reflectors, prismatic diffusers), the results of such calculations are necessarily approximate. (Skylights and luminous ceilings do have a Lambertian distribution, and for these sources these calculation methods do give reliable results.) In addition to the direct component of illuminance, a reflected component must be added that depends on the point's location in the room and the room characteristics. The calculations involve charts, diagrams, and tables. The interested reader will find a full description of these manual methods in the IESNA Lighting Handbook (2000).

Because these manual methods are laborious and frequently less than reliable, they are not presented here. The ready availability of desktop computers and computer programs that can handle detailed input for a specific light source, without broad approximations, has made these manual procedures obsolete. We recommend that when point-by-point illuminance calculation is desired, such a program be used. Alternatively, one of the design aids described previously, based on a specific light source, can be used.

30. COMPUTER-AIDED LIGHTING DESIGN

As pointed out in previous sections, the use of computers in lighting design is a practical necessity because of the complexity of calculations. Once the desired luminance patterns and illuminance levels have been established, and luminaires selected and located by using either average illuminance (zonal cavity) calculations or one of the design aids previously demonstrated and/or the manufacturer's assistance, the responsible designer will confirm the preliminary design solution with accurate calculations. Furthermore, only a computer analysis will give useful point-by-point illuminance figures plus valuable data on VCP and reflected glare for selected work locations and viewing directions.

In addition, computer analysis gives the designer a degree of flexibility not otherwise possible in that:

1. The calculations are performed accurately and rapidly.

2. The designer is freed for other, less routine work.

3. The designer has the ability to change parameters repeatedly without making the analysis excessively burdensome, as would be the case with hand calculations.

It is this last characteristic that gives the designer greatest flexibility. The ability to run a series of calculations for a pendant fixture with varying pendant lengths, or to change paint colors and reflectances for various surfaces and note the effect, or to test different layout patterns, as mentioned in Example 15.2, gives the designer a very powerful and extremely useful design tool. In addition, computer analysis can consider related items, such as first costs, energy use, operating costs, and impact on HVAC systems- items whose complexity because of interrelations puts them well beyond the pencil-and-hand calculator's ability. A few lighting analysis programs currently available are listed in Appendix M.

31. COMPUTER-AIDED LIGHTING DESIGN: ILLUSTRATIVE EXAMPLE

In order to demonstrate the huge amount of accurately calculated data available at a designer's finger tips when using any of the comprehensive lighting design programs, we have chosen to run Example 15.2 with a typical program: Lumen-Micro 7. The first step, of course, is to input the data and the target illuminance. For convenience they are repeated here:

• Room: 60 ft × 100 ft × 8 ft (18.3 × 30.5 × 2.4 m)

• Reflectances: 0.80 ceiling, 0.50 wall, 0.30 floor

• Required uniform illuminance: 50 fc (500 lux)

• Luminaire: recessed parabolic troffer

Most lighting analysis programs (like this one) have a large library of luminaires that is constantly updated, plus an efficient search program. This obviates the tiresome task of physically searching through several dozen catalogs and the even more tiresome task of keeping the catalogs current. We decided to use a 2-ft (0.6-m) square luminaire in order to reduce the possibility of both direct and reflected glare to persons viewing the fixture cross wise, and further chose to use U-shaped lamps. A very rapid search of the software's extensive library of lighting fixtures resulted in the selection of the luminaire shown in FIG. 50 (as printed from the program's library). On the basis of a preliminary calculation, an area of approximately 75 ft^2 (7 m^2) per fixture seemed appropriate (because in the original longhand calculation an area of about 50 ft 2 (4.6 m^2) was used for a two-lamp fixture). A row spacing of 7.5 ft (2.3 m) and a column spacing of 10 ft (3.0 m) was input to the analysis software, and a complete point-by-point calculation was run for horizontal and vertical illuminance, VCP and relative visual performance (RVP). The numerical results are shown in Figs. 51 to 53. As the room is very large and symmetrical and is being calculated with no daylight contribution, a calculation for one-quarter of the space is sufficient. Figure 51 gives a summary of the results. If daylighting calculations had been done, their results would be shown as well. The average illuminance in the space is 48 fc, compared to the 50-fc target illuminance, and is therefore well within the desired range. Unit power density (UPD) is only 1.29 W/ft^2 (13.9 W/ m^2) which is within the requirements of the major energy codes. The Luminaire Summary shows the fixture chosen; had several units been used, all would be listed. Similarly, a list of luminaires in their positions in the space appears at the bottom of the figure. Inasmuch as only one type was used in this space, the long, repetitive list was truncated. Figure 52 is a printout of the point-by-point horizontal illuminance on the working plane (30 in. [760 mm] AFF), and it immediately demonstrates the value of a point-by-point calculation. Although the average illuminance is 48 fc and therefore satisfies the target requirement, the illuminance drops to as low as 38.5 fc and increases to 65 fc in the body of the space, away from the walls. (The calculation grid is 2.5 ft by 2.5 ft [0.76 by 0.76 m].) In practice, this situation might cause the designer to select a closer fixture spacing or perhaps change the luminaire, because 38.5 fc is 23% below the target illuminance.

On another printout of horizontal illuminance (FIG. 53) we have plotted luminaire locations, desk locations, and one drafting table. (The plan is not scalable because the printout row and column spacings are different.) Note that on desks against the wall, the average illuminance is near 50 fc, whereas on the drafting table it is about 45 fc because of its location. A staggered arrangement might improve uniformity, and therein lies one of the great advantages of computer calculation. With an additional half hour of work, an experienced designer who is accustomed to working with a given program could try three luminaires, each in three different arrays, and compare the results to arrive at an optimum lighting/economic solution.

Figure 54 shows partial printouts of three other metrics of the lighting design. Vertical illuminance (FIG. 54a) is useful for visual tasks on other than horizontal surfaces. In conventional offices where the majority of work is done on the horizontal plane, vertical illuminance is not critical. As pointed out repeatedly in previous sections, diffuse lighting is required to achieve good vertical surface illumination from overhead sources. In this space, with direct lighting only and (effectively) no walls to reflect light and increase diffusion, vertical illumination comes primarily as a component of direct lighting, with a small contribution from light re-reflected from the floor.

Figure 54b indicates uniformly high VCP when looking north (directly ahead). This is to be expected from a parabolic reflector luminaire, which is noted for low luminance and therefore low direct glare. Relative visual performance (RVP) is the metric that replaced equivalent spherical illumination (ESI) as an indication of supra threshold visual response, considering task size, contrast, and background luminance. The task we selected for these calculations is a difficult one: HB pencil on vellum, on a white table. Nevertheless, the calculations indicate perfect RVP. In the discussion of reflected glare, ESI and RVP in Sections 11.29 and 11.30, we expressed serious reservations about the usefulness of RVP as a visual performance metric when considering veiling reflections; the results shown in FIG. 54c seem to reinforce these reservations.

FIG. 50 Photometric report obtained from the computer library of Lumen-Micro 7© as a result of a search for a suitable 2-ft (610-mm) square, recessed parabolic luminaire for the example in Section 15.31.

FIG. 51 Summary of the results of a single layout plan.

Graphical output is often easier to grasp than a raft of numbers. Using the same calculation grid as previously, we printed out horizontal illuminance contours overlaid with the point-by-point calculations in FIG. 55. (They can just as easily be printed separately.) The rectangular boxes in the figure are closed 40-fc contour lines.

32. AVERAGE LUMINANCE CALCULATIONS

FIG. 52 Point-by-point horizontal illuminance results.

The basic equations relating luminance to candela intensity and to illuminance are covered in Section 11.8, which also deals with the calculation of source luminance and reflected luminance when the illuminance is known. Thus, once horizontal illuminance has been calculated by any of the methods described previously and the reflectance of an object is known, its horizontal luminance can readily be calculated (see Section 11.8). However, as explained in detail in Sections 11.32 and 11.33 (which discuss lighting quality) and in Sections 13.10 through 13.16 (which deal with lighting systems and patterns) the luminance impression of a visual environment is affected more by vertical than by horizontal surface luminance. For this reason, it is important to be able to calculate average vertical surface (wall) luminance in the same simple, straightforward fashion used to calculate average horizontal illuminance.

In addition, it is useful to know the average luminance of the ceiling cavity in a space in order to judge the contrast between all luminous objects, including luminaires, which have the ceiling cavity as background.

FIG. 53 Illuminance chart with luminaires and proposed furniture layout plotted on the chart. (Lighting Technologies, Boulder, CO.)

Straightforward calculation of both wall and ceiling cavity luminance (LW and LCC) is possible through the use of luminance coefficients that are similar in concept and application to coefficients of utilization. These coefficients are listed in TABLE 4 for some of the generic fixture types listed in TABLE 1. Others are listed in the IESNA Lighting Handbook (2000). For actual design calculations, it is preferable to obtain coefficients from luminaire manufacturers. The average luminance calculations are parallel to those for illuminance.

Average initial wall luminance (cd/m^2):

(18)

and average initial ceiling cavity luminance in cd/m^2:

(19)

If area is expressed in square feet and p is omitted from these equations, L will be expressed in foot-lamberts.

FIG. 54 Partial printout report for the computer analysis of Example 15.2. (a) Point-by-point calculation of vertical illuminance at the working plane level. Note that it averages less than half of the horizontal illuminance. (b) VCP chart of the room. (c) RVP chart indicating perfect seeing conditions in the entire area looking north. (Lighting Technologies, Boulder, CO.)

FIG. 55 A combined graphic (contours) and point-by-point printout of horizontal illuminance for lighting design. The almost rectangular figures represent closed contours at 40 fc.

To obtain maintained values, an LLF similar to that explained in Section 15.20 is introduced.

It is calculated similarly, except that item 17 (see FIG. 37), room surface dirt, is calculated using the following figures:

Lighting System Wall Luminance Ceiling Luminance Direct 0.82 ± 10% 0.75 ± 10% Semi-direct 0.87 ± 7% 0.82 ± 10% Direct-indirect 0.92 ± 5% 0.85 ± 8% Semi-indirect 0.87 ± 7% 0.88 ± 7% Indirect 0.82 ± 10% 0.90 ± 5%

For ceiling-mounted or recessed luminaires, LCC is the average luminance of the ceiling between luminaires. For pendant luminaires, the calculated LCC is that of an imaginary plane at the height of the luminaires. LCC is useful in determining the brightness ratios when compared to luminaire luminance at the seeing angle involved. The ceiling cavity, like the wall, is assumed to have a Lambertian characteristic-that is, perfect diffuseness-making luminance independent of viewing angle.

It would be instructive to calculate the wall luminance of the office in Example 2. The photometric data in FIG. 39 do not include the wall luminance coefficient because luminance coefficients are not normally published by luminaire manufacturers. However, based on other available data, a figure of 0.22 for wall luminance is a good estimate given RCR, ?CC, and ?W of 0.66, 80%, and 30%, respectively. Initial wall luminance is then

[…]

is within the preferred range of 25 to 150 cd/m^2 (see Table 11.3). In actuality, the average wall luminance would probably be higher because of the practice of placing the last row of luminaires quite close to the wall.

TABLE 4 Wall Luminance Coefficients and Ceiling Cavity Luminance Coefficients for Typical Luminaires (Continued )

EVALUATION

33. LIGHTING DESIGN EVALUATION

The final step in lighting design is evaluation of the design relative to three key aspects-lighting, costs, and energy. The lighting aspects include quantity, quality, luminance ratios, mood, ambience, texture, color, variation, psychological impressions, orientation, and daylight use-in short, a review of all the lighting factors previously discussed in detail. A good deal of experience is required to visualize actual lighting results from design drawings.

The novice designer would do well to have someone with such experience assist in doing the review. The other two aspects of evaluation, cost and energy, can be evaluated readily with the aid of the contractor's estimating figures for cost and a straightforward calculation for energy. The estimates are compared to the cost and energy budget figures developed at the preliminary design stage.

As we have repeatedly stressed, the important cost figures are life-cycle cost, annual operating cost, and first cost for economic comparisons, operating budgets, and construction budgets, respectively. In Section 16 we present lighting recommendations for specific occupancies accompanied by actual cost studies and energy analyses. Detailed cost studies including the impact of lighting on air conditioning, the proportional cost of the wiring system, and the proper apportionment of costs involve the entire building and can be accurately performed only by computer. Studies of this type are generally made by consulting engineers rather than architects, and then only after initial, operating, and total costs have been set in proper perspective for a particular job by the architect and client. This is necessary because often, as in the case with speculative construction, the client's overriding consideration is first cost, thereby rendering a complete cost analysis unnecessary. Any attempt to completely separate costs for lighting, HVAC, structure, and so on is arbitrary because of the intimate interactions among these elements.

Lighting designers are well advised to keep them selves and the construction team aware of this if they are to fulfill their responsibility.

References / Resources

ASHRAE. 2007. ANSI/ASHRAE/IESNA Standard 90.1-2007: Energy Standard for Buildings Except Low-Rise Residential Buildings. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. Atlanta, GA.

IESNA. 1981. IES Lighting Handbook, 7th ed. Illuminating Engineering Society of North America. New York.

IESNA. 1993. IESNA Lighting Handbook, 8th ed. Illuminating Engineering Society of North America. New York.

IESNA. 2000. IESNA Lighting Handbook, 9th ed. Illuminating Engineering Society of North America. New York.

NEMA. 1998. NEMA LE 5B-1998: Procedure for Deter mining Luminaire Efficacy Ratings for High-Intensity Discharge Industrial Luminaires. National Electrical Manufacturers Association. Rosslyn, VA.

NEMA. 1999. NEMA LE 5A-1999: Procedure for Determining Luminaire Efficacy Ratings for Commercial, Non-Residential Downlight Luminaires. National Electrical Manufacturers Association. Rosslyn, VA.

NEMA. 2001. NEMA LE 5-2001: Procedure for Deter mining Luminaire Efficacy Ratings for Fluorescent Luminaires. National Electrical Manufacturers Association. Rosslyn, VA.

NFPA. 2008. NFPA 70-2008: National Electrical Code. National Fire Protection Association. Quincy, MA.

UL. 2004. UL 1598: Standard for Luminaires. Underwriters Laboratories, Inc. Northbrook, IL.

Prev. | Next

Home  Similar articles   top of page