C L I N I C A L
A N D
E X P E R I M E N T A L
OPTOMETRY cxo_462
66..76
RESEARCH PAPER
Performance of ‘energy efficient’ compact fluorescent lamps
Clin Exp Optom 2010; 93: 2: 66–76 Gloria S-C Yuen* BEng MBiomedEng Alistair B Sproul† BSc(Hons) PhD Stephen J Dain* BSc(Hons) PhD FCOptom FAAO FIES(ANZ) FMSA * School of Optometry and Vision Science and † School of Photovoltaic and Renewable Energy Engineering, University of New South Wales, Sydney, New South Wales, Australia E-mail:
[email protected]
Submitted: 4 August 2009 Revised: 2 November 2009 Accepted for publication: 6 December 2009
DOI:10.1111/j.1444-0938.2010.00462.x Background: Compact fluorescent lamps (CFLs) have been heralded as highly energy efficient replacements for incandescent light globes, however, there is some public dissatisfaction with the light output and colour of CFLs. Independent examination of the claims made has not been made. Compliance with the interim Australian/New Zealand Standard has not been established by any independent authority. While the total light output (luminous flux) may meet certain standards, luminous intensity distributions of some designs do differ significantly from the incandescent sources that they are intended to replace. Methods: Luminous intensity distribution, luminous flux and spectral energy distribution of CFLs claimed to be equivalent to 75 W incandescent globes and 75 W incandescent globes (pearl and clear) were measured. Luminous flux, luminous efficacy, colour rendering index, correlated colour temperature, wattage and power factor were then calculated and compared with claims made by manufacturers and requirements of the standards. Results: The sources generally complied with the requirements for luminous flux, luminous efficacy, colour rendering index and correlated colour temperature. The claim of 75 W equivalence, which is not regulated in Australia and New Zealand, is justified less than half the time. Luminous intensity distributions of biaxial CFLs are distinctly different from the incandescent lamps they purport to replace. Conclusion: CFLs generally comply with the standards set. The basis on which equivalent wattages are claimed needs to be included in the Australian and New Zealand standard because this is the measure most likely to be relied on by the public. Due to the differences in luminous intensity distribution, CFLs may not necessarily be a direct replacement for incandescent sources without some consideration.
Key words: electric lighting, energy consumption, energy efficiency, lamps, lighting, lighting standards, light sources
Compact fluorescent lamps (CFLs) are replacing incandescent lamps in many applications because they represent significant energy and cost savings. Legislation is increasingly demanding a move to CFLs. Optometrists frequently advise Clinical and Experimental Optometry 93.2 March 2010
66
patients on appropriate lighting and this will increasingly involve recommendations about CFLs. There is no doubt that they are more energy efficient than incandescent lamps but there seems to be no independent validation of the claims made for
CFLs, including the implied magnitude of energy savings. CFLs differ from incandescent lamps in several ways beyond their much publicised improved energy efficiency and longer life. These include: © 2010 The Authors
Journal compilation © 2010 Optometrists Association Australia
Performance of compact fluorescent lamps Yuen, Sproul and Dain
1. They do not reach a stable output for some minutes after being turned on. 2. They are often much bulkier and may extend beyond shades and covers and create a glare source. 3. The large area of light emission from a CFL is much more difficult to direct exactly with reflectors and in luminaire design. 4. Very few of them are dimmable and the dimming systems applicable to incandescent lamps are inappropriate. 5. CFL output is ambient temperature dependent. This project was initiated to provide independent assessment of some of the claims about CFLs and not least to assist in the aim that the advice given to patients by optometrists will be complete and accurate. These other factors (listed above) are important but in the current study the issues of claims on light output will be specifically addressed. In February 2007, the Australian Government proposed to phase out lamps that have a luminous efficacy less than 20 lm.W-1 by 2010.1 This phase-out plan allows for special purpose lamps (for example oven lights and heat lamps) and tungsten halogen lamps (which are marginally more efficient than other incandescent lamps but less efficient than CFLs) to remain available. Industries are at liberty to negotiate longer change-over times. As a result, the lamps that will be considered sufficiently efficient include CFLs and light emitting diodes (LEDs). As LEDs are currently more expensive to purchase, particularly on a luminous efficacy basis, CFLs are currently the lamps in the public’s mind. The technology of CFLs is similar to that of the linear fluorescent tube. CLFs comprise two main components: a ballast and tubing. Modern CFLs use electronic ballasts, which remove the start-up flicker that is often seen in fluorescent tubes.2 The fluorescent tube works by passing a current through mercury vapour, exciting the molecules and causing the emission of ultraviolet (UV) and visible radiation. The UV radiation then excites the phosphor coating on the inside of the tubing, resulting in the emission of light. The colour of
the light depends primarily on the combination of phosphors used to coat the tubing. Most domestic CFLs have electronic ballasts that are integrated with the tubing (self-ballasted) and have the common bases of bayonet caps (B22) or Edison screws (E27). As most CFLs are also small enough to fit into the majority of household fittings, they can easily be used to replace incandescent lamps. There are many designs including double, triple and quadruple biaxial, spiral, globe and reflectors. There are also four main colours: warm white, white, cool white and daylight. The claims made by companies producing CFLs include that the lamps use less power for the same light output as incandescent lamps, resulting in more lumens per Watt. The life of CFLs is also claimed to be longer than that of their incandescent counterparts, resulting in both a longer-lasting lamp and lower running costs. This offsets the higher purchase price of the CFLs, leading to a net saving over the lifetime of a CFL. In the past, anecdotal evidence has suggested that CFLs are disliked due to the visible flickering, low light output and colour of the light. The visible flickering was due to the use of magnetic ballasts, which caused flickering at 50 to 120 Hz,3 however, there have been many modifications since the introduction of CFLs and flicker is now in the kilohertz range (20 to 40 kHz).1,4,5 This is undetectable by the human visual system. While the colour for which CFLs were disliked is still available, there are also other colours available, including ‘warm white’. A warm white colour corresponds to a correlated colour temperature of about 2700 K and closely resembles the colour of light from incandescent lamps. Recently, a survey6 was carried out of the opinions of 600 Australians on CFLs. Various companies and consumer organisations around the world have also conducted their own tests to compare aspects of CFLs with incandescent lamps.7–9 Concerns centre on light quality, size of the lamp and cost.8,10 The Australian survey showed that both awareness and use of CFLs is growing and the public’s opinion
© 2010 The Authors Journal compilation © 2010 Optometrists Association Australia
of them is improving but there remain issues with the light quality and cost. Many people also appear to doubt the claims on energy savings. The survey showed that the public would generally welcome the introduction of an endorsement label, which would allow for the average consumer to compare products better or, at least, know that they are purchasing something that satisfies minimum performance standards. While the survey points to the public’s desire for endorsement labels and performance standards, the Australian Government’s proposed plan makes performance requirements a necessity if the reputation of CFLs, and hence the public’s acceptance, is to be maintained. Currently, there is a number of performance criteria for CFLs around the world. There is an international exercise referred to as the International CFL Harmonisation Initiative, which has the ultimate goal of standardising testing methods and performance of CFLs to enable the delivery of high-quality products worldwide.11 The Harmonisation Initiative created a report in 2006,12 in which existing performance criteria were compared and a tiered system for international harmonisation was suggested that could incorporate the present different standards more easily. The different standards are an issue for Australia, as CFLs are imported and, as a consequence, have passed different performance standards. Understandably, Standards Australia is a supporter of the initiative. A preliminary report5 on Australia’s Minimum Energy Performance Standards (MEPS) options for CFLs has already been issued as well as a new interim Australian Standard.13–14 The first part of this standard describes suitable test methods to obtain performance data on CFLs. The second part, AS/NZS 4847.2(Int)14 contains the finalised Australian MEPS for CFLs and other selfballasted gas-discharge lamps. A technical report prepared for The Australian Greenhouse Office15 also compares the proposed performance criteria of CFLs of various countries. The Efficient Lighting Initiative Institute has proposed some suitable performance criteria and is a voluntary certification that is available Clinical and Experimental Optometry 93.2 March 2010
67
Performance of compact fluorescent lamps Yuen, Sproul and Dain
worldwide.16,17 Its performance requirements are actually contained in the Australian Standard.14 While consumers are likely to be aware of aspects of CFLs such as power usage (in Watts), light output (in lumens) and colour (as correlated colour temperature or a descriptor), they are probably unaware of the power factor (PF). Power factor is a property of alternating current (AC) circuits. Its value ranges from zero to 1. Circuits with purely resistive elements have a power factor of 1. The coiled coil of an incandescent lamp has only a slight inductive effect (being a coil) so, in practice, they may not quite achieve a power factor of 1.0. Inductive or capacitive elements (mainly the ballast of a CFL) will result in a PF of less than 1.18 A typical CFL ballast will have a PF between 0.4 and 0.6.18 A good PF for a ballast would be greater than 0.85, while a PF greater than 0.9 is referred to as a high power factor ballast.18 True power factor is defined as follows:
Real Power Apparent Power Watts = Volts × Amps
True Power Factor =
(1)
Utility companies supply customers with the ‘apparent power’ (referred to as VA, volts x amps) but generally only charge for the real power (W),18 except where power factors in an installation are too low. The difference between the two is referred to as the ‘reactive power’ (VAr).19 True power factor is also explained as a combination of ‘displacement power factor’ and ‘distortion power factor.20 There are many issues indicated by a low PF. For residential or small commercial customers, the electricity bill does not reflect the increased power needed to use low PF appliances, however, there is actually more power being supplied by the utility company, resulting in increased generation and transmission costs and losses.19,21,22 The problem is usually described as a need to supply more current out of phase with the voltage. Voltage is carefully regulated to ensure a constant voltage is delivered to the customer (for example, 240 V rms). The load Clinical and Experimental Optometry 93.2 March 2010
68
consumes the power associated with current in phase with the voltage and this ‘real power’ is billed to the customer. Current out of phase with the voltage (reactive power) has to be generated by the electricity supply and transmitted to (and from) the customer if the load has a low PF. The customer is not billed for the out of phase current and it simply travels from the generator to the customer, charges up, for example, an inductor or capacitor and then, on the opposite cycle, the current is subsequently sent back to the generator. This additional current that has to be carried by the transmission and distribution network results in effects such as increased resistive losses and overloading of transformers. The use of several low PF appliances on the one circuit also tends to increase the circuit’s susceptibility to circuit breaking.19,21 Some electricity rates may include an increased charge for industrial consumers in the form of a ‘low PF penalty’ for large installations.18 Low PF is also an indication of the amount of harmonic distortion caused by the circuits, resulting in decreased equipment lifetime.21 The purpose of this paper is to carry out independent testing to assess what is claimed with regard to power consumption, equivalences to incandescent lamp power ratings, light output, colour rendering, correlated colour and temperature, as well as report the performances, which are important but not subject to claims, in light distribution and power factor. METHODS
Samples Self-ballasted compact fluorescent lamps (CFLs) of several designs (spiral, double, triple and quadruple biaxial), all available brands at the time, and correlated colour temperature ranging from 2700 K to 6400 K were purchased from retailers. They all had bayonet cap (B22) or Edison screw (E27) bases. To contain the cost and time of testing, the sampling was limited to CFLs that claimed to have a light output equivalent to a 75 W incandescent lamp, anecdotally the most commonly pur-
chased incandescent lamp wattage. Where available, three clear and three pearl incandescent lamps from the same brand were bought. The CFLs were operated for a minimum of 30 minutes to allow for warm-up and to attain a stable maximum lumen output. This period exceeds what is reported as the necessary warm-up time to achieve a stable output, which for CFLs is usually less than 30 minutes.2 The laboratories in which testing was carried out had an ambient temperature of 22 ⫾ 2°C. One of each lamp was mounted onto a goniometer (Optronik SMS 10 mm) and centred onto the geometric centre of the lamp to measure the light distribution. The distance between the lamp and the detector was 8.00 m, which ensures that errors due to distance, source size and detector size are negligible (less than one in 10,000). The lamp was powered using a stabilised power supply (240.0 ⫾ 0.1 VAC), with voltage and current at the lamp terminals being measured using multimeters (Hewlett Packard HP 34401A Multimeter and a Fluke 45 Dual Display Multimeter, respectively, with current calibration certificates from a laboratory accredited by the National Association of Testing Authorities of Australia [NATA]). A portable wattmeter (EMU 1, EMU Elektronik) was also used to determine the real power. These data were used to calculate the power factor of the lamp. The photometer was calibrated using three reference lamps operating at 2856 K, for which calibration data had been obtained from the Australian National Measurement Institute. The Optics and Radiometry Laboratory (ORLAB) of the School of Optometry and Vision Science, University of New South Wales is accredited by NATA for luminous flux measurements with a least uncertainty of 2.5 per cent. The angular luminous intensity distributions of the lamps were determined by taking readings at 10 degree intervals of elevation and azimuth using an Optronik SMS 10 mm goniophotometer. The luminous intensity distribution of each lamp (brand, design and colour) was split into quadrants for easy analysis and determined once only. Polar diagrams for each © 2010 The Authors
Journal compilation © 2010 Optometrists Association Australia
Performance of compact fluorescent lamps Yuen, Sproul and Dain
1100
1100
1050
Brand 0 Brand 3
1000
Brand 4
Measured lumens (lm)
Luminous flux (lm)
1000
Brand and claim
900
2 940 lm 3 5
800
11 13 925 lm
Brand 5 Brand 9 Brand 10
950
Brand 13 Brand 14
900
850
13 930 lm
800
700 750
600
700
66
68
70
72
74
76
78
80
82
500
Measured power (W)
700
800
900
1000
1100
Claimed lumens (lm)
Figure 1. Luminous flux and power for incandescent lamps
design were constructed for comparison. Following Method B in AS/NZS 4847.1(Int): 2008, a total luminous flux was determined using the luminous intensity values.23,24 The luminous flux of the two remaining samples of each brand, design and colour were determined by comparison with the first sample using Method B14 through the use of an integrating sphere of one metre diameter (Alexr Wright & Co, Westminster, UK repainted in 2008 with a latex based barium sulphate paint custom manufactured by Dulux Australia) and a detector (IL1700 Radiometer, International Light). The procedure also used the auxiliary lamp in the sphere to correct for self-absorption in the CFL under test. Again, stabilised power sources were used and voltage and current were monitored. The luminous flux of each lamp type was then compared with the output claimed by the manufacturer, the corresponding luminous flux of an incandescent lamp and the minimum incandescent lamp equivalent required for ‘efficient lighting initiative’ certification.16 The Australian Standard does not contain provi-
600
Figure 2. Measured luminous flux as a function of claimed luminous flux for CFLs. The dashed line represents claimed equal measure, so that points lying above the line represent CFLs that exceeded the claims made and those below fell short of the claimed luminous flux.
sions for comparing luminous flux and does not set out equivalency values. The luminous efficacies of the lamps were determined by dividing the luminous flux of the lamp by the apparent power. These values were compared with the ELI and local requirements stated in AS/NZS 4847.2(Int).14 The power factor was calculated using Equation (1) with the real power being measured by the wattmeter and the apparent power being calculated from the voltage and current values measured by the multimeters. These were compared with the local requirements set out in AS/NZS 4847.2(Int).14 The correlated colour temperature (of the incandescent lamps), CIE General Colour Rendering Index (CRI)25 and correlated colour temperature (CCT) were measured using a spectroradiometer (Topcon SR-3) and a reference white surface in a dark room. The spectroradiometer has been wavelength calibrated using mercury and neon discharge lamps (wavelength accuracy ⫾0.6 nm and spectral half band width 6.0 nm) and the response calibrated using three spectral
© 2010 The Authors Journal compilation © 2010 Optometrists Association Australia
irradiance reference lamps, for which spectral irradiance calibration data have been obtained from the Australian National Measurement Institute and a reference white surface for which spectral reflectance data (0/45°) has been obtained from the National Research Council of Canada. ORLAB is accredited by NATA for such measurements with a least uncertainty in absolute spectral irradiance of 2.0 per cent at 380 nm, falling to 1.0 per cent at 555 nm and then rising to 2.4 per cent at 830 nm, measurements of colour rendering index with least uncertainties of ⫾1.0 and correlated colour temperature with least uncertainties of ⫾15 K in the region of 2856 K rising to ⫾50 K at 5500 K and ⫾200 K at 12,000 K. The lamps were allowed to warm up for a minimum of 30 minutes. The chromaticity co-ordinates and CCT were obtained directly using the spectroradiometer software (having been previously validated by calculation using the spectral radiance values), while the CRI was calculated using the spectral radiance data from the spectroradiometer in accordance with CIE Clinical and Experimental Optometry 93.2 March 2010
69
Performance of compact fluorescent lamps Yuen, Sproul and Dain
18
1400
17
1300
16
Brand 0 Brand 1 Brand 2
15
Luminous flux (lm)
Measured power (W)
1200
14 13 12 11
Brand 3
1100
Brand 4 Brand 5
1000
Brand 6
900
Brand 7 Brand 8
800
Brand 9 Brand 10
700
Brand 11
600
10
Brand 12 Brand 13
500
9
Brand 14
400 8 12
14
16
18
Claimed power (W)
RESULTS Three samples of each CFL model were tested in case there were variations in the data obtained, arising from variability in the manufacturing process. A total of 11 incandescent lamps were measured for comparison. As there was little variation in compact fluorescent lamp performance within the same model, the considerations in the remainder of this paper will deal with make and model rather than individual samples. While the distributions of all lamps bought and tested were investigated, it was clear that all lamps of the same shape had the same luminous intensity distribution and only the luminous flux changed significantly between brands As a consequence, some data reduction has been achieved by reporting data Clinical and Experimental Optometry 93.2 March 2010
70
10
12
14
16
18
20
Measured power (W)
Figure 3. Measured power as a function of claimed power for CFLs. The dashed line represents claimed = measured so that points lying above the line represent CFLs that are a higher power than claimed and those below are a lower power than claimed.
13.3.23 These values were compared with the local requirements set out in AS/NZS 4847.2(Int).14
8
Figure 4. Measured luminous flux as a function of measured power for CFLs
relating only to differences rather than similarities. The first issue in assessment of the equivalency claims is to establish the actual luminous flux of 75 W incandescent lamps. Of the seven brands measured, three made claims of 925 lm, 930 lm and 940 lm. The measured luminous fluxes as a function of ‘real power’ are plotted in Figure 1. They range from 716 lm to 1069 lm over the range of powers investigated. The brands complied with the luminous flux claim, where it was made. In the ELI17 system, 940 lm is taken as the nominal 75 W incandescent output for comparison purposes. Figure 2 shows the comparison of claimed (where a claim was made) and measured luminous flux for incandescent lamps and CFLs. With two exceptions, the performance is close to or greater than the claims. Similarly, Figure 3 shows measured versus claimed power. While the CFLs rated as 14 W have measured power
relatively close to the rated value, the CFLs rated as 15 W vary widely, typically consuming less power than the rated value. For all the CFLs, Figure 4 shows the luminous flux as a function of measured power. Much of the variation of luminous flux may be seen as a variation of actual (as against rated) power and not luminous efficacy. To represent the issues of luminous intensity distribution, Figure 5 shows several comparisons between representative CFL types. In Figure 5A the luminous intensity distributions of clear and pearl incandescent (IC) lamps are shown. Figure 5B shows the luminous intensity distribution of the CFLs that use biaxial tubes in their design. Figure 5C shows the globe CFL and spiral CFL luminous intensity distributions. In Figure 6, the power factors of the lamp types tested are compared. It is noted that there is little variation © 2010 The Authors
Journal compilation © 2010 Optometrists Association Australia
Performance of compact fluorescent lamps Yuen, Sproul and Dain
180 160
170 100
-170
1.0 -160
90
150
-150
80
140
-140
70
130
-130
60
120
Clear IC
-120
50 40
110
0.9
-110
30 20
100
Pearl IC
-100
10 0
90
-90 -80
70
Power factor
80
-70 60
-60 50
-50 40
-40 30
-30 20
10
-10
0.8
Brand 0
Brand 1
Brand 2
Brand 3
Brand 4
Brand 5
Brand 6
Brand 7
Brand 8
Brand 9
Brand 10
Brand 11
Brand 12
Brand 13
Brand 14
0.7
-20
0
180 160
170 120
150
-170
0.6
-160 -150
100
140
-140
80
130 120
Triple -130
110
Double
-120
60
MEPS requirement
Quad
0.5
-110
40
100
20
-100
90
0
-90
80
Spiral
Double
-80
70
Quad
Globe
CFL
-70 60
Triple
-60 50
Figure 6. ‘Power factor’ (see text for definition) as a function of CFL type. All the samples meet the MEPS minimum requirement of 0.55. CFL design has no influence on ‘power factor’.
-50 40
-40 30
-30 20
10
-10
-20
0
180 160
170 100
-170
-160
90
150
-150
80
140
-140
70
130 120
-120
50 40
110
Globe
-130
60
20
100
Spiral
-110
30
-100
10 90
-90
0
80
Brand 0
1400
-80
Brand 1 70
1300
-70 60
-60
40
-40 30
-30 20
10
-10
-20
0
Figure 5. Polar plots of luminous intensity for incandescent sources (clear and pearl) in the top panel. Three types of biaxial CFLs are in the middle panel and the spiral and globe types in the lower panel. For the biaxial CFLs the main light distribution is to the sides, while the globe and spiral much more closely resemble the distribution of the incandescent sources. The zero degree point on the polar axis represents the base of the lamp. The representation is of a lamp in a pendant fitting.
© 2010 The Authors Journal compilation © 2010 Optometrists Association Australia
Brand 2
1200
-50
Luminous flux (lm)
50
Brand 3 Brand 4
1100 ELI 75W equivalent threshold
1000
Brand 5 Brand 6
900
Brand 7
800
Brand 8 Brand 9
700
Brand 10
600
Brand 11
500
Brand 12 Brand 13
400
Brand 14
Spiral
Double
Triple
Quad
Globe
CFL design
Figure 7. Luminous flux as a function of CFL design. Most samples do not meet the ELI requirement of 940 lm to qualify as a 75 W incandescent equivalent. CFL design appears to have little influence.
Clinical and Experimental Optometry 93.2 March 2010
71
Performance of compact fluorescent lamps Yuen, Sproul and Dain
Luminous Efficacy (lm.W-1)
85
CCT 4500K
Brand 1 Brand 2
1400
75
Brand 3
-1
Luminous Efficacy (lm.W )
80
Brand 4
Brand 0
70
1200
Brand 5
Brand 3
Brand 6
Total lumen output (lm)
65
Brand 7 Brand 8
60 Brand 10
55
Brand 12 Brand 14
50
ELI: 15-24W
CFLs < 15W
Brand 7
1000
Brand 8 Brand 9
800
Brand 1 0 Brand 1 1 Brand 1 2
600
Brand 1 3 Brand 1 4
400
Brand 2 IC Brand 3 IC
45
PASS
AS/NZS 4847.2 (CFL)
200
ELI: 9-14W
Brand 5 IC Brand 11 IC Brand 1 3 IC
40 Spiral
Double
Triple
0
Quad
CFL Design
40
60
80
Figure 9. Graphical representations of the varying efficacy levels of different brands and designs of CFLs compared with the local luminous efficacy requirements set in AS/NZS 4847.2. All CFLs represented by points to the right of the dashed line meet the standard. Two CFLs lie just to the left of the requirement line and the remaining lamps well to the left are incandescent.
Minimum luminous efficacy levels for CFLs (lm/W) Colour temperature 4500 K
9–14
55
52
15–24
60
57
Table 1. Recommended luminous efficacy level specifications for selfballasted CFLs13
Clinical and Experimental Optometry 93.2 March 2010
72
20
Lumen efficacy (lm/W)
Figure 8. Luminous efficacy as a function of CFL design. The ELI requirements differ depending on nominal power. The symbols surrounded by a grey square represent the performance of CFLs, for which the ELI 9–14 W requirements apply. For the remainder the 15–24 W requirements apply. The majority but not all the samples meet the ELI requirements.
Rating (Watts)
Brand 1 4 IC
0
between the power factors of CFLs of different designs and different manufacturers with one exception. Barely passing the requirement of PF greater than 0.55 seems to be the normal expectancy of performance. In Figure 7, the luminous fluxes of CFLs are compared with their incandescent lamp equivalents, providing the basis by which to judge the claimed equivalencies. CFL design is seen to have no influence on the luminous flux. Figure 8 shows the luminous efficacy values graphed against the ELI thresholds of Table 1 so that compliance with the © 2010 The Authors Journal compilation © 2010 Optometrists Association Australia
Performance of compact fluorescent lamps Yuen, Sproul and Dain
300 Brand 1 Brand 2 Brand 3
88
100
Brand 1 Brand 4 Brand 5
0
Brand 6
-100
Brand 7 Brand 8
-200 Brand 9
-300
Brand 1 1 Brand 1 2
-400
Brand 1 3 > 5 SDCM (limit set in AS/NZS4847.2
-500 2500
3000
3500
4000
4500
5000
5500
6000
6500
CIE Colour rendering index Ra
Measured - claimed CCT (K)
200
Brand 2
87
Brand 3
86
Brand 4 Brand 5
85
Brand 6 Brand 7
84
Brand 8 Brand 9
83
Brand 10
82
Brand 11 Brand 12
81
Brand 13 Brand 14
80 2500
3000
Figure 10. Representation of the difference between the measured ‘correlated colour temperature’ (CCT) and the nominated CCT of the CFLs as a function of the measured CCT. The points surrounded by a circle are more than five standard deviations of colour measurement away from the nominated CCT and, therefore, fail the AS/NZS 4847.2 criteria for colour appearance.
DISCUSSION The selection of a 75 W equivalent was made because we understood that this was the most popular wattage, however, this choice proved to be a complicating factor because the CFLs that claim equivalence
4000
4500
5000
5500
6000
6500
Measured CCT (K)
Claimed CCT (K)
requirements may be assessed. Again CFL design has no influence on the luminous efficacy. Figure 9 summarises the data for CFLs and incandescents to show whether compliance with AS/NZS4847.2 is achieved. Figure 10 shows the difference between the measured CCT and the nominated CCT of the CFLs, with the lamps that failed the AS/NZS 4847.2(Int) criteria for CCT (local and ELI criteria are the same) being surrounded by a circle. The CFL CRI values showed a small range from 80.4 to 87.7. There were no statistically significant trends in CRI with nominated or actual CCT, design or brand. These data are represented in Figure 11.
3500
Figure 11. CIE General Colour Rendering Index (CRI) as a function of measured ‘correlated colour temperature’. Minimum CRI is universally set as 80 for CFLs, so all samples comply with this requirement.
with a 75 W incandescent lamp proved to be, generally, 14 or 15 W. The ELI17 requirements are set in two ranges, 9 to 14 W and 15 to 24 W and the requirements differ slightly in these two groupings. The higher wattage CFLs are required to be more efficient.
Specifications and measurements The claimed luminous flux is reasonably matched by the measured luminous flux (Figure 2) but there is a huge variation in what manufacturers think is the equivalent of a 75 W incandescent. The claims range from 600 lm to 950 lm. The ELI17 requirements set 940 lm as the 75 W equivalent. There is no equivalent set down in the Australian Standard.13–14 The ELI minimum for a 75 W equivalent seems entirely reasonable given the data reported in this study in Figure 1, except that Brand 11 is low both in power and luminous flux, yet only about half the CFLs meet the requirement. Figure 3 seems to identify the problem as an inability to set the power accurately and Figure 4
© 2010 The Authors Journal compilation © 2010 Optometrists Association Australia
shows that much of the variation in luminous flux is due to the inexact setting of power.
Light distribution From Figure 5, it may be seen that the clear incandescent lamp luminous intensity distribution is not as smooth as the pearl due to the filament being visible and the effects of self-shading and irregularities in the envelope glass, however, the distribution is essentially the same, with light being emitted both downwards and sideways in similar proportion. From Figure 1B, it may be noted that, regardless of how many biaxial sets there are in the CFL design, the light distribution remains essentially the same. Most of the light is emitted sideways with relatively little light being directed downwards. In Figure 1C, it may be observed that the luminous intensity distributions of the globe and spiral CFLs are very similar to one another and to the incandescents. The globe and spiral CFL designs are, as a consequence, the most appropriate for Clinical and Experimental Optometry 93.2 March 2010
73
Performance of compact fluorescent lamps Yuen, Sproul and Dain
unshielded or unshaded incandescent lamp substitution. As there are many types and arrangements of domestic light fittings that shield and shade in different ways, the end user should choose the type of CFL carefully to provide the most appropriate distribution for the application and may need sound advice in this choice.
Power factor As mentioned previously, the power factors of incandescent lamps are mostly close to 1 (measured as 0.96 to 0.97) as the lamps are composed of essentially resistive components.18 In Figure 6 it can be seen that there is little variance between the power factors of CFLs of different designs. This is expected, as the less than unity power factor of CFLs is caused by the electronic ballast of the lamp, which is either similar or the same for all lamps and has little to do with the shape of the CFL. It was verified that CFLs have a power factor of around 0.6, showing that they have the potential to cause electrical disturbance if used in large quantities.18,19,21,22 There would also be significant transmission and generation losses for utility companies. Using the local requirement stated in AS/NZS 4847.2(Int), the minimum power factor for lamps is 0.55 and 0.9 for high efficiency ballasts.14 Looking at Figure 3, it is noted that the current CFLs, except for one brand, have PFs just above this requirement.
Comparison of lumen output Before comparing the lumen output of CFLs with incandescent lamps, we must first distinguish between pearl and clear incandescent lamps. The experimental data show that the only difference between the two finishes is light distribution (Figure 5). The luminous flux remains much the same within brands. The data of Figure 7 should be viewed in the light of the ELI Certification scheme that requires that the total lumen output of a CFL equivalent to a 75 W incandescent lamp should be greater than 940 lm.17 This threshold is shown as a dotted line on the graph. Our own data on incandescent lamps shows that this is a Clinical and Experimental Optometry 93.2 March 2010
74
reasonable requirement, with the averaged claimed lumen output for incandescent lamps being 930 lm and the average measured lumen output being 987 lm (disregarding Brand 11 incandescent lamp as an outlier, averaging approximately 740 lm). There are differences in CFL light output within the same brand for different designs and colour temperatures. This may be due to the irregularity of phosphor coatings and the varying available surface area of the lamps due to design. The recommended criteria also imply that lamps of different correlated colour temperatures are expected to perform differently. As reported, there are many currently available CFLs that do meet the incandescent lamp equivalency required by ELI.17 These lie below the horizontal dashed line in Figure 7. This means that only a few CFLs in the market meet the ELI standards with regard to the claimed equivalence to a 75 W incandescent lamp. The Australian/New Zealand standard lacks any regulation of this claim, so these CFLs do not fail the Australian/New Zealand requirements.
Comparison of luminous efficacy Luminous efficacy is the ratio of a lamp’s luminous output in lumens to its electrical power consumption in Watts. The CFLs tested here were rated at either 14 or 15 W and were of either Warm White (CCT 2700 to 3000 K), Cool White (CCT 3500 to 4100 K) or Daylight (CCT greater than 5000 K) colours.10,23 Some values of luminous efficacy are quoted from the ELI equivalency values17 in Table 1. It may be seen in Figure 9 that most lamps are above the requirement for their rated wattage, with the 14 W lamps often well exceeding the requirements, however, there are a few lamps that do fall below the threshold. As these lamps were purchased before the release of the Australian standard, the next few years should see efficacy levels and power factors improving. The local standard requires the use of an equation based on the initial luminous flux of the lamp in lumens. It has the same requirements regardless of the colour
temperature of the lamp. While in this study the lamps were not aged for the required 100 hours, the luminous flux was measured after a minimum of half an hour of stabilisation. Ageing for 100 hours will inevitably reduce the luminous flux, so that this study may represent a slightly optimistic view of the actual situation. As can be seen in Figure 6, the majority of CFLs pass the local criteria for luminous efficacy. The two lamps that failed the lumen efficacy standard also failed the ELI threshold requirements (Figure 9).
CCT and CRI The correlated colour temperatures of all the incandescent lamps were between 2569 and 2760 K, with the two lowest readings coming from the Brand 11 incandescent lamps, which failed the colour appearance requirements. It is noted that the total lumen outputs of the Brand 11 incandescent lamps were also well below the threshold required of 75 W equivalent CFLs (Figure 4). The colour rendering indices of all the incandescent lamps were between 99.2 and 99.9. The minimum CRI required for passing the standard is 80. The colour rendering indices of the compact fluorescent lamps ranged from 80.4 to 87.7 and therefore, all complied. There were no correlations between nominated colour temperature, design or brand. The packaging of CFLs often includes a colour appearance (for example, ‘warm white’), a colour temperature (for example 2700 K) or both. For lamps for which only a worded colour appearance description was used, the corresponding colour temperature was chosen accordingly. From Figure 11, it can be seen that most of the lamps tested easily pass the required accuracy of claimed correlated colour temperature. The few that failed may need to check their phosphor coatings or the manufacturing reproducibility.
Relationship with cost The retail cost of the CFLs purchased ranged from AU$3 to AU$16. There were no statistically significant relationships of cost with luminous flux, luminous efficacy or CRI. © 2010 The Authors
Journal compilation © 2010 Optometrists Association Australia
Performance of compact fluorescent lamps Yuen, Sproul and Dain
CONCLUSIONS Recent concerns over the state of the world environment have spurred new policies regarding the replacement of incandescent lamps with the more energy efficient compact fluorescent lamps. Australia’s new policy may include a national endorsement label and Minimum Energy Performance Standards (MEPS) as stated in the new interim standards, AS/NZS 4847.1 and 4847.2.13,14 Combined with consumer surveys, this has brought attention to some concerns about the light quality, lumen output, colour, size, cost and energy efficiency of the CFLs. This paper addresses some of these issues by testing available brands of CFLs of varying designs and colours and comparing them with their claimed incandescent counterparts. Not all the criteria stated in the standard were assessed in this study. As this study was commenced before the standard was published, the methods used also differ somewhat from the standard. Only three samples of each type of lamp were tested and no sample ageing was applied beyond an initial 30-minute stabilisation period. The testing required in the standard is exceptionally lengthy and therefore, costly. Without government support, this tends to put full testing beyond the resources of all but the manufacturers. This may have the consequence of precluding widespread independent assessment of products. While the biaxial CFLs generally emit more light horizontally (to the side) than downwards, the spiral and globe CFL distributions were very similar to pearl incandescent lamps that radiate similarly in both horizontal and vertical (downward) directions. Where the lamp is shaded, this difference may be insignificant but where direct illumination from the lamp is used, the type of fitting and where light is needed most, should be carefully considered before choosing a CFL type. Users may need good advice on this choice from lighting designers, illuminating engineers and optometrists. The luminous fluxes of the lamps were compared with both the claimed amount
and the measured amount for incandescent lamps. Generally, most brands make valid claims regarding the CFL luminous flux, however, the luminous output of a CFL is not always equivalent to the luminous output of a 75 W incandescent lamp. The international efficient lighting initiative certification states that to qualify for equivalency, a total lumen output of 940 lm must be obtained. Only nine of the 32 CFL models tested (28 per cent) passed this criterion. This seems to be an unfortunate omission from the Australian standard. As a consequence, the claimed equivalent wattage is misleading. The authors are of the opinion that such a requirement must be part of the Australian/New Zealand Standard. Efficiency was assessed by measuring luminous efficacy (lm.W-1) and power factors. The standard gives minimum luminous efficacy values for CFLs as a function of colour temperatures and rated wattage. Most of the tested CFLs meet this requirement easily. The more difficult requirement was the power factor. The standard states that a minimum power factor for CFLs will be 0.55. For consumers, a significant aspect of CFLs is the cost. We also compared the maximum (purchase) cost and the efficacy levels of the lamps. Price appears to make little difference in any aspect of the lamp performance.
SUMMARY While the vast majority of CFLs complied with the Australian Standard requirements for luminous efficacy, correlated colour temperature and colour rendering index, the fact remains that the vast majority of them emit less light that their claimed 75 W incandescent equivalent. Elsewhere in the world, there is a minimum luminous flux set for equivalence claims. Only 28 per cent of the lamps studied here comply. The public seems to be disappointed when they replace an incandescent lamp with a CFL claimed to be equivalent. This study shows that they are entirely justified in their disappointment.
© 2010 The Authors Journal compilation © 2010 Optometrists Association Australia
In the long term, the use of incandescent lamp equivalent wattages will, undoubtedly, disappear. In the meantime, it seems that the public relies on the claim and to continue to omit regulation of such claims from the Australian/New Zealand Standards is a public disservice. As a minimum, the packaging of CFLs should include the information of claimed incandescent lamp power equivalent, power consumption, luminous flux, luminous efficacy and colour category. This would assist in appropriate selection and comparison of CFL products in the retail environment Public education on selection of CFL type for specific uses is essential for the residential applications (retailers and consumers) and commercial/industrial applications (electrical wholesalers and contractors). The setting of a minimum power factor seems to have stifled any effort to improve power factors, since all brands comply but only one seems to have a high power factor option. At present, the manufacturers simply say that this is not an issue in domestic lighting26 but expanding use of lower power factor CFLs must inevitably lead to some action from the power generating and transmitting authorities, which currently bear the cost of reactive power. REFERENCES 1. Turnbull M. Phase-Out of Inefficient Light Bulbs. Australian Greenhouse Office, Department of the Environment and Water Resources, Australian Government. 2007. 2. Roisin B, Bodart M, Deneyer A. On the substitution of incandescent lamps by compact fluorescent lamps: switch on behaviour and photometric distribution. XII National Conference on Lighting. Varna, Bulgaria, 2007. 3. Finn DW, Ouellette MJ. Compact fluorescent lamps: what you should know. Progressive Architecture 1992, Aug, 89–92 Accessed as http://irc.nrc-cnrc.gc.ca/ pubs/cp/lig3_e.html on 7/7/09. 4. Lighting Applications Overview. Hearst Semiconductor Applications. 2007. 5. Minimum Energy Performance Standards: Compact Fluorescent Lamps. National Appliance and Equipment Energy Efficiency Program, 2005. 6. Winton L. Final Report on a Consumer Research Study about Compact Fluorescent Lamps (CFLs). Artcraft Research for Aus-
Clinical and Experimental Optometry 93.2 March 2010
75
Performance of compact fluorescent lamps Yuen, Sproul and Dain
7.
8. 9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
tralian Greenhouse Office, Department of the Environment and Heritage, Australian Government, 2005. Masamitsu E. The compact fluorescent light bulb revolution. Popular Mechanics 2007. Accessed as http://www.popular mechanics.com/home_journal/home_ improvement/4215199.html?page=1 on 7/7/09. CFLs: The future of indian energy scene? Consumer Voice 2006: 9–13. Watson NR. A Comparison of the Philips Tornado Compact Fluorescent Lamp with the New High Power-Factor Eco-Bulb, 2005. Ecobulb. Figueiro M, Fullam J, O’Rourke C, Overington M, Rea M, Taylor J. Increasing Market Acceptance of Compact Fluorescent Lamps (CFLs). New York: Lighting Research Center, Rensselaer Polytechnic Institute, 2003. International CFL Harmonisation Initiative. 2007 http://www.apec-esis.org/www/ cfl/. Accessed 23/11/07. Second Proposal for Tiers of Performance and Marking System for CFLs. International CFl Harmonisation Initiative, 2006. AS/NZS 4847.1:2008—Self-Ballasted Lamps for General Lighting Services: Part 1. Test Methods—Energy Performance. Interim Australian/New Zealand Standard. Standards Australia, Standards New Zealand, 2008. AS/NZS 4847.2:2008—Self-Ballasted Lamps for General Lighting Services: Part 1. Minimum Energy Performance Standards (MEPS) requirements. Interim Australian/New Zealand Standard. Standards Australia, Standards New Zealand, 2008. The Australia Greenhouse Office. PhaseOut of Inefficient Incandescent Lamps and Standards for Compact Fluorescent Lamps. 2007. ELI. Efficient Lighting Initiative: A Commitment to Quality. 2008. http://www. efficientlighting.net/. Accessed 7/7/09. ELI Voluntary Technical Specification for Self-Ballasted Compact Fluorescent Lamps (CFLs), in ELI Technical Documents for Certification. ELI Quality Certification Institute: Beijing, China, 2006. Hambley AR. Electrical Engineering, Principles and Applications. Upper Saddle River, New Jersey: Pearson Prentice Hall, 2008. Wuidart L. Understanding Power Factor, in Application Note: Power Supply and Power Management. Geneva: STMicroelectronics, 2003. Cividino L. Power Factor, Harmonic Distortion; Causes, Effects and Considerations. in 14th International Conference Telecommunications Energy. Washington: 1992.
Clinical and Experimental Optometry 93.2 March 2010
76
21. Total Harmonic Distortion, in Datasheets: Integra Digital Metering Systems. Crompton Instruments, 2007. 22. Grady WM, Gilleskie RJ. Harmonics and how they relate to power factor. In: Proceedings of the EPRI Power Quality Issues and Opportunities Conference (PQA ’93). San Diego, CA, 1993. 23. CIE 13.3-1995. Method of measuring and specifying colour rendering of light sources. Commission Internationale de l’Éclairage Central Bureau. Vienna, 1995. 24. CIE 70-1987. The measurement of absolute luminous intensity distributions. Commission Internationale de l’Éclairage Central Bureau. Vienna, 1987. 25. CIE 84-1989. The measurement of luminous flux. Commission Internationale de l’Éclairage Central Bureau. Vienna, 1989. 26. Jacob B. Lamp and gear technology. Symposium ‘Good lighting with less energy: Possibilities for the future’. London June 18 2009. Society of Light and Lighting/ Lighting Research and Technology/ National Physical Laboratory.
Corresponding author: Professor SJ Dain Optometry and Vision Science University of New South Wales Kensington NSW 2052 AUSTRALIA E-mail:
[email protected]
© 2010 The Authors Journal compilation © 2010 Optometrists Association Australia
View publication stats
