"Metrology for Solid State Lighting" [1] has developed procedures for reliable and traceable measurement of the electrical parameters of SSL sources.
Traceable measurement of the electrical parameters of solid-state lighting products D. Zhao1, G. Rietveld1, J.-P. Braun2, F. Overney2, T. Lippert3 and A. Christensen3 1
VSL, Dutch Metrology Institute, Thijsseweg 11, 2629 JA Delft, The Netherlands Email: 2 Federal Institute of Metrology METAS, Lindenweg 50, 3003 Bern-Wabern, Switzerland 3 Trescal A/S, Mads Clausens Vej 12, 8600 Silkeborg, Denmark Abstract — In order to perform traceable measurements of the electrical parameters of solid-state lightning (SSL) products, it is essential to technically adequately define the measurement procedure and to identify the relevant uncertainty sources. This paper fills the related gaps in the present written standard for testing SSL products. New uncertainty sources with respect to conventional lighting sources are determined and their effects quantified. For power measurements on SSL products the main uncertainty sources are temperature dependence, power supply THD, and stabilization of the SSL product. For current rms measurements the influence of bandwidth, shunt resistor, power supply source impedance, and ac flatness error are significant as well. Index Terms — solid-state lighting, electrical parameters, measurement, traceability, measurement uncertainty.
I. INTRODUCTION In many aspects, Solid-State Lighting (SSL) has proved itself as the energy-efficient alternative of fluorescent lamps and incandescent lamps. Market acceptance of this new technology is presently hampered by ambiguous product performance data. Therefore, the European joint research project on "Metrology for Solid State Lighting" [1] has developed procedures for reliable and traceable measurement of the electrical parameters of SSL sources. The authors are frequently asked by non-electrical experts questions like: “Why our electrical measurement results show such a large discrepancy with results obtained by other laboratories?”, “Is the present power meter in our test setup good enough, or do I need a meter with wider frequency range and higher resolution?”, “How to calculate the uncertainty of the result? Is using the power meter specifications sufficient?” This paper shows that creating an uncertainty analysis of electrical measurements for SSL is not a murky task anymore. The present standard for SSL testing [2] specifies procedures for measuring total luminous flux, electrical power, luminous efficacy, and chromaticity of SSL luminaries and replacement lamp products. The electrical measurements follow the standards of previous kinds of lighting products. However, to achieve traceable and accurate values, these procedures require significant adjustment taking into account the specific characteristic of SSL products. This paper firstly describes the characteristics of SSL sources. Subsequently, the uncertainty sources in electrical testing of SSL are identified and extensively discussed.
978-1-4799-2479-0/14/$31.00 ©2014 IEEE
II. CHARACTERISTIC OF SSL PRODUCTS First of all, SSL products are nonlinear electrical loads: with a pure sinusoidal supply voltage, the current drawn is not sinusoidal. To achieve better efficiency and lower losses, higher switching frequencies are used in SSL. Consequently, the SSL current signal is rich in high frequency harmonic components far beyond 2 kHz, even beyond 100 kHz. For some compact SSL products without power factor correction (PFC) circuit, the current waveform is heavily distorted resulting in significant low frequency harmonic components (100 Hz – 2 kHz). A special feature of such SSL products is its high crest factor (the ratio between the peak current and the root mean square current). SSL products are used in households as well as in public or industry applications. They have a wide power range. The root mean square (RMS) current can be as low as several mA or as high as several amperes. To achieve the best accuracy over this wide power range, it is crucial to choose the most suitable measurement configuration in each situation. III. SOURCE OF UNCERTAINTY A. Temperature inconsistence The temperature dependence of SSL products depends on the circuit design. Most, but not all, SSL products include a temperature compensation circuit. The voltage drop over the LED changes around 0.08 % in 1 ○C. If we assume the driver circuit is ideal, the power consumption can be estimated as 0.08 % change within 1 ○C change of ambient temperature. B. Stabilization of the SSL product Long-term fluctuations are not included as uncertainty here, but short-term fluctuations are a source of uncertainty because we never know at which moment the measurement is done. According to the criterion [2] that stability is reached when the variation the electrical power over a period of 30 min is less than 0.5 %, each moment in this 30 minutes period satisfies the criterion. This 0.5 % fluctuation needs to be included as uncertainty source. This requirement can be set as stringent as 0.2 % to reduce the uncertainty if the condition allows for it. For most SSL products, it is advised to wait 10 minutes longer before starting a measurement.
650
C. Influence of the distorted voltage signal of power supply
G. Influence of power supply source impedance
Voltage sources in SSL testing are never ideal. The source voltage harmonics can produce power with the current harmonics of the same frequency if they are not orthogonal to each other. This high frequency power is measured by a power meter, and becomes an error with respect to the requirements [2] where a perfect 50 Hz voltage source is assumed. In order to estimate the size of this effect, amplitudes of each voltage and current are derived from voltage source total harmonics distortion (THD) specifications and current spectra from typical SSL lamps. Taking into account the phase between voltage and current as well, calculations show that a linear power supply with 3 % THD can result in a 5.8 % discrepancy caused by the high-frequency power. A linear power supply with 0.1 % THD will reduce the error to a negligible level. Thus, the present criterion (
