Parkyn 2005
Parkyn 2005
Parkyn 2005
ABSTRACT
The ubiquitous downlight inhabits our ceilings by the millions. Hot, inefficient, and electrically wasteful, it is next in
line for replacement by the latest high-brightness, high-efficacy white LEDs. The conventional downlight configuration
of a large incandescent spotlight in a low-cost, ceiling-recessed metal can, represents the culmination of old technology,
fated never to improve significantly. Incandescent downlights add greatly both to direct and indirect electrical
consumption, with the lamps requiring relatively frequent replacement. The small size of LED emitters means small
optical elements can produce much higher-quality beams than incandescent spotlight-lamps can produce. Herein we
introduce compact high-luminosity LED downlights with lenses that deliver uniform illumination to delimited targets
such as tables. One version utilizes circular lenses and micro-diffuser films to deliver square outputs. The other uses
lenses cut to the target shape. In particular, one of these lenses is the first to offer a semicircular spot suitable for
gambling tables.
Keywords: LED downlights, low-spill down-lighting , square spots, semicirclular spots, uniform illumination..
1 LED DOWNLIGHTS
1.1. Advantages
To see so often the multitudes of ceiling-mounted downlight-cans is to know that their national energy impact is
significant (i.e., tens of billions of Watts in the U.S. alone). Parabolic spotlights, or PAR lamps, seem less frequent than
bare 100-W light-bulbs, which obviously deliver far less onto the intended target than to the inside of the can. Reducing
these high electricity and heat-load expenditures per delivered lumen is only part of the benefit of LED downlights, with
long operating life also attractive. For architects, builders, and users, the compactness and low-voltage operation means
much lower cost and greater versatility of installation. A key feature of LEDs is the compactness of their emission
zones, enabling highly controlled downward distribution of their light, with very low spill into the rest of the room. In
restaurants, one downlight over each table would have both high table illuminance and low ambient illumination, with
very low spill onto patrons or floor. Herein we discuss two types of optical systems for compact low-spill LED
downlights, suitable for both retrofitting and as low-voltage lighting in new construction.
In the United States there is an installed base of 4.4 billion, low efficiency, incandescent and halogen lamps which
annually produce 16% of the total electrically generated light output (4.6 trillion lumen hours). However, due to their
Incandescent lamps can be broken down into two main groups. The first group is the general service type incandescent
lamps which cost less than $0.25 per thousand lumens (klm) output and consume 237 TWh (or 31%) of the total
lighting energy budget annually. The second group consists of moderate to high-priced halogen and standard reflector
lamps ($3-13/klm), and this group of lamps alone consumes 77 TWh of energy or 10 % of the total lighting energy
budget.
In an attempt to compare present-day LED lighting systems with incandescent lamps, let us compare the lumen output
and the cost of commercially available white LEDs. Large white LEDs are commercially available in 1 to 5 watt
packages (i.e. input electrical power) from a number of manufacturers including Lumileds, Osram-Opto, Nichia, and
Cree. To be specific, let us focus our attention on the one watt package which typically consists of an LED that is
approximately 1 mm on a side, running at a current of 350 mA at a voltage of approximately 3.5 V. A typical luminous
efficacy of such packages is currently in the range of 40 to 50 lumens per watt (recent publications by Cree [4] and
Nichia [5] indicate that 55 to 60 lumens per watt packages will be available within the next 12 to 18 months). For
comparison, consider a standard 100 watt incandescent lamp which has a rating of 1500 lumens and costs
approximately $0.50. The corresponding white LED one watt package costs approximately $2 (in large quantities) and
has a 50 lumen output. To obtain a 1500 lumen output, one would require 30 of these one watt white LEDs at a cost of
approximately $60. Consequently, there is today an approximate 120 to 1 cost disadvantage for white LEDs when
compared to a simple incandescent lamp. When one considers higher end halogen lamps with their corresponding
reflectors at a cost of approximately $6 for an MR 16 lamp (20 watts at approximately 18 lumens per watt for 360
lumen output), the corresponding white LED equivalent would require 7 one watt packages at cost of approximately
$14, and now the cost penalty for switching to LEDs is substantially reduced to a 2.5 to 1 disadvantage. However, this
analysis is too simplistic (as will be shown in later sections of this paper) and does not take into account other
characteristic advantages of LEDs that will further reduce these cost penalties by factors of approximately 4. Thus,
based upon this analysis, one can conclude that architectural, accent, and display lighting is cost competitive today and
general service incandescent lamps will probably see LEDs taking their market share away in 3 to 5 years. Further,
LEDs’ penetration into lighting markets such as vehicular lighting, traffic signals, emergency egress, medical/dental,
and backlighting for small aperture LCDs, is more than compelling and should continue to be dominated by LEDs as far
into the future as one can see.
An LED in its simplest form is a semiconductor p-n junction device that, when forward biased with a direct current (dc)
flowing through its p-n junction emits photons as a result of the electrons and holes recombining near the junction. The
energy of the photons is primarily determined by the energy bandgap of the semiconductor where the recombination
occurs. Compound semiconductor materials composed of Group III and V elements (from the Periodic Chart) are the
materials of choice for LEDs because they have direct-bandgap properties and bandgap energies necessary for
efficiently producing visible photons. The best AlInGaP LEDS (red and amber) convert about 40-50% of the electrons
sent into their p-n junctions directly into useful light output. The best InGaN LEDs (UV, blue, green, and white)
convert 25-35% of electrons traveling through their p-n junctions into useful light.
Consider the case of using large white power LEDs to replace a simple 100 watt incandescent lamp, which when new is
rated at 1500 lumens, and see if we still believe that LEDs are a cool light source. At today's efficiency figures of 50
lm/W for the white high-power LED, we would require 30 of these 1 W LEDs to produce 1500 lumens. Assuming that
30% of the power goes into creating light and that 70% goes into creating heat, we would need to dissipate 21 watts of
heat. While this is not a tough task if one can radiate this power away at the fourth power of the incandescent lamp’s
high temperature (3000°K), it is a more difficult task for an LED lamp, which only have convection and conduction for
effective heat removal.
2.4 Conclusion
LEDs are not a cool light source, even though today they have approximately three times the efficiency of an
incandescent bulb, and in the near future probably four times to five times the efficiency of an incandescent bulb. The
ability of incandescent lamps to get rid of excess heat via radiation transfer is their fundamental advantage over LEDs,
and is due to their much greater temperature elevation (thousands °K) over ambient causing much steeper temperature
gradients and higher heat fluxes than ever found in LEDs. As heat sources, LEDs must have clearly defined heat paths
for installers. Illustrations below are shown with rear heat fins.
LEDs by their very nature are quite small as light sources go, each an almost perfect thermodynamic light source, so
they are easily integrated with optical systems. The ability of any optical system to gather up the light from any source
is directly proportional to the size of the optical system relative to the size of light source. Consequently LEDs enjoy a
fundamental advantage over incandescent lamps, fluorescent lamps, and high-intensity discharge sources because of
LEDs’ intrinsically small size and the fact that they do not require a large glass envelope as many of their competitive
light sources do.
Optics can broadly be broken down into the two fields of imaging and non-imaging optics. Imaging optics has been
around for well over 300 years and is the optics of parabolas, ellipses, thick lenses, thin lenses, and Fresnel lenses. The
one characteristic that all of these optical technologies have in common is that they form images of objects and are
frequently used in such things as cameras, movie and 35mm slide projectors, automobile headlights, flashlights,
eyeglasses, etc. As the lighting industry developed over the past 200 years it was natural to incorporate this already
existing imaging optical technology into new lighting systems. A modest thought-experiment, however, will show that
imaging optics is far from optimum. Consider for the moment the simple example of a parabola used in a flashlight to
project a beam. Depending upon the depth of the parabola, only about 40% of the light leaving the light bulb will
reflect off the parabolic mirror be collimated and projected into the beam. The other 60% of the light leaves the
flashlight in an unguided way and is typically not useful and in many applications considered a negative attribute known
as glare. This is particularly true for automobile headlights. It is obvious that the optimum optical system should
Figure 3. Total Internal Reflection (TIR) Lens, highly compact collimator, injection moldable.
One of the most useful features of LED lighting systems is their ability to create virtually whatever output lighting
pattern one desires with optical efficiencies in excess of 80%. This unique characteristic of LED lighting systems is the
After the original invention of computer-generated kinoform at IBM research labs in 1969, a number of industrial
researchers [9,10,11,12,13,14,15,16,17,18,19] , led by H. J. Caulfield, showed how kinoform diffusers could be made as either
volume or surface relief structures [ 9] and recorded by using only optical means, without the complication of having first
to calculate a computer-generated wavefront, as taught by IBM in 1969 [8]. H. J. Caulfield further showed [11] that these
kinoform diffusers were highly efficient in either transmission or reflection and that they were completely tailorable
with regard to their angular dispersion for their far field radiation patterns. This meant that for the first time it was
possible to take a collimated beam of light and by passing it through a kinoform diffuser create whatever far field
radiation pattern one desired, including shapes such as circles ,ellipses, squares, rectangles, alpha characters, or even a
company logo. The first use of these kinoform diffusers for lighting applications was taught in IBM’s Technical
Disclosure Bulletin [10] in August 1971 for the purpose of uniformly lighting an area with a light source. This
homogenizing of the output beam is particularly applicable and useful with LED lighting systems. In what follows, we
will attempt to delineate more clearly the unique attributes of these kinoform diffusers (also known as holographic
diffusers, light shaping diffusers or tailored micro- diffusers).
In general, kinoform diffusers provide the very efficient achromatic forward scattering with low backscatter and
adsorption. The angular scattering pattern may be highly asymmetric in the azimuth and elevation directions, and may
also be skewed away from the incident beam direction. These properties make kinoform diffusers ideal for many
applications. We will now examine in more detail the attributes and characteristics of kinoform diffusers relative to
conventional diffusers [19]
Kinoform diffusers achieve near 100% forward scattering efficiency of the light incident upon them (88–92% for
surface relief kinoform diffusers and 92% for volume kinoform diffusers, with no AR coatings). Conventional diffusers
will have forward scattering efficiencies in the range of 40 to 80%, depending upon the particular type of conventional
diffuser and its method of construction. This is perhaps the most significant advantage of kinoform diffusers.
The angular extent of the forward scattering pattern can be exactly matched to the application. Normally incident light
can be scattered by the kinoform diffusers into cone angles exceeding 90° or less than 5°. Furthermore, the scattering
lobe remains stationary relative to the kinoform diffuser for all angles of incidence that lie within the scattering lobe.
This is very different from the operation of the conventional diffuser in which the output lobe changes direction in direct
proportion to the angle of incidence of the incoming beam
.
The kinoform diffusers discussed above all operate in transmission, however, it should be recognized that reflective
kinoform diffusers, having very similar angular scattering properties, have also been constructed. It is also worth
emphasizing that kinoform diffusers exhibit highly achromatic performance. These diffusers may therefore be applied
to the very broadband light sources such as florescent lamps for LCD backlighting without causing color distortion.
They can also be used directly with white LED light sources and not cause any color breakup in the transmitted beam.
Note also that these kinoform diffusers perform very well into the infrared portion of the spectrum ( to about 1000 nm).
-
/
TX:
Figure 4. Kinoform diffusers can be made to scatter in various aspect ratios [17]
Figure 5. Kinoform diffusers can provide off-normal peak scatter and/or multiple lobes [17]
(reprinted with permission of the Society of Information Display)
The TIR lens shown in Figure 3 is intended for sources with substantial side emission, such as the Lumileds Luxeon.
Many LEDs however, have a Lambertian pattern with zero emission at 90°, causing the periphery of the lens to be
wasted, and the output beam to be overly wide. Raising the lens rim somewhat is thus necessary, albeit at some
efficiency cost. The cumulative flux out to angle θ is sin2θ, so that 88% of the flux is within 70°. Another consideration
is where to truncate the TIR lens so it will pack hexagonally into a circle, as shown in Figure 7.
Figure 7 Nineteen TIR Lenses cut to fit hexagonally, with 6 of the 12 peripheral lenses further cut to fit a circular outline.
C
Figure 8. Cutaway rear view of circular 6” LED downlight with 19 TIR lenses. Beam-shaping diffuser is on the cover.
The TIR-lens downlight in Figure 8 would also have dark vertical partitions running up from the circuit board, with one
between each hexagonally cut lens, in order to prevent spill light. This 19W less should duplicate the illuminance of a
75W incandescent downlight, but in a well-bounded zone with little light outside it.
A companion paper20 details the tailoring method for these lenses. Their basic function is to provide a cos-4 intensity
distribution, which gives constant illuminance on a target plane. Figure 9 shows two cross-section and their patterns.
II
LED
LENS BOTTOM
LENS TOP
Figure 11. Hexagonal lens array cut to a circle. The Div line marksa hexagonal boundary between lenses. The array puts into
the beam 88% of the LED’s light to 63° off-axis.
However, these lenses are cut, they will appear uniformly illuminated across their output face. Thus when they are cut
to a square shape (Figure 12, top), they will produce a square beam (Figure 12. bottom), and will mount in a square
downlight (Figure 13). Also, a semicircular lens (Figure 14) will make a semicircular spot, for gambling tables.
I—
(-4
as
Lfl
as
L.fl
SD
Figure 12 (above) Squared-cut lens and its square-output beam (below) at z = 2.5m
LED. MIRROR
SEMI-LENS
Figure 14. Semicircular lens producing a semi-circular spot, for gambling tables.
1.) R. Steele, “High Brightness LED Market Overview and Forecast”, Strategies in Light Newsletter, June. 2005, by Strategies
Unlimited.
2.) U.S. Lighting Market Characterization, Vol.1: National Lighting Inventory and Energy Consumption Estimate, U.S.
Department of Energy, Office of Energy Efficiency and Renewable, September 2002
3.) Energy Saving Potential of Solid State Lighting: U.S. Department of Energy, Office of Building Technology, April 2001
4.) www.cree.com, see X- Bright Plus LED specifications.
5.) M. Yamada, T. Mitani, Y. Narukawa, S. Shioji, I. Niki, S. Sonobe, K. Deguchi, M. Sano, T. Mukai, Jpn. J. Appl.
Phys., 41, L1431- L1432 (2002)
6.)W.T. Welford, R. Winston, The Optics of Nonimaging Concentrators, Academic Press, New York, 1978
7.)W.T. Welford, R. Winston, High Collection Nonimaging Optics, Academic Press, New York, 1989
8.) L.B. Lesem, P.M. Hirsch, J.A. Jordan Jr.,” The Kinoform: A New Wavefront Reconstruction Device”, IBM J. Res. Develop.,
March 1969
9.) H.J. Caulfield, “ Kinoform Diffusers,” SPIE 25, Development in Holography, 1971
10.) “Uniform Area Illumination Using Kinoforms,” IBM Technical Disclosure Bulletin, Aug. 1971
11.) H.J. Caulfield, “ Optically Generated Kinoforms”, Opt. Comm.,4,#3, Nov. 1971
12.) K. Biedermann, F. Bestenreiner, “ Method of Making a Diffusing Layer”, U.S. Pat.# 3,619,021, issued Nov. 9, 1971
13.) F. Bestenreiner, W. Weierhausen, P. Janovjak, “ Method and Arrangement for Manufacturing Diffusion Lenses of Arbitrary
Indicatrix”, U.S. Pat.# 3,698,810, issued Oct.17,1972
164) D. Meyerhofer, “ Holographic and Interferometric Viewing Screens”, Appl. Opt., 12, p. 2180-2184, Sept. 1973
15) P.F. Gray, “ A Method of Forming Optical Diffusers of Simple Known Statistical Properties”, Optica Acta, 25,
p.765 – 775, 1978
16.) “Light Diffuser with Controlled Divergence”, IBM Technical Disclosure Bulletin, p.276-279, June 1986
17.) J.M. Tedesco, L.A.K. Brady, W.S. Colburn, “Holographic Diffusers for LCD Backlights and Projection
Sceens”, p.29-32, SID Digest 1993
18.) T. Suzuki,” Method for Optically Making a Diffusion Plate”, U.S. Pat # 4,336,978, issued Jun.29,1982
19.) T. Suzuki, “Method for Making a Member Having Microstructure Elements Arranged Thereon”, U.S. Pat. #4,523,807,
issued Jun. 18,1985
20) SPIE Vol 5942 # 3, ‘Illumination-redistribution lenses for non-circular spots’, W.A. Parkyn & D.G. Pelka