introduction
Although LED technology has a history of more than 40 years, there are still phenomena such as poor reproducibility of LED optical parameters in the industry, large measurement uncertainty, and poor consistency of test results between different test devices [1~3]. The reason is as the domestic long-term management of the LED industry, the same internationally: the International Semiconductor Equipment and Materials Organization (SEMI), the International Electrotechnical Commission (IEC) and the International Commission on Illumination (CIE) all involve LEDs in different degrees, especially It is the latter two committees. It is precisely because the relevant measurement standards for LEDs are formulated by different international standardization organizations, and there is no systematic planning for various international organizations as a whole, and there is no sufficient consultation between relevant organizations. Therefore, there are different quality evaluation systems, and the documents issued are The technical content is also different. Founded in 1906, the IEC treats LEDs as a display semiconductor device, focusing on its physical properties. Founded in 1913, CIE treats LEDs more as a light source device, resulting in minor differences between their respective LED measurement standards. This article attempts to find out the differences between the two by comparing their respective criteria in order to provide some reference for the finalization of the LED test method standard.
1 CIE and ICE different representations of the same event
1.1 Luminescence (radiation) definition of performance
First, the misunderstanding of the term must be corrected. The luminosity efficiency is not appropriate in this paper because efficiency refers to a dimensionless physical quantity, which is dimensioned here. Therefore, the correct name is “emission (efficiency)â€.
Definition of luminescence (radiation) performance:
CIE Definition: The ratio of the luminous flux (radiant flux) emitted by an LED to the power consumed.
IEC definition: The ratio of the luminous flux (radiant flux) emitted by the LED to the forward current [4].
Comment: The luminous performance of CIE is measured in three physical quantities: total luminous flux, forward current, forward voltage (or internal resistance). The IEC only needs to measure two physical quantities: total luminous flux, forward current. It is wise to choose a forward voltage because the forward voltage will drop as the die temperature increases. The author believes that the definition of IEC is not rigorous enough. Because there is a slight difference in the internal resistance and terminal voltage of the die for the same batch of dies and packages, it is only suitable for situations where the luminescence (radiation) performance is allowed to vary. This test time will be shortened as only two physical quantities need to be measured.
1.2 Measurement distance of luminous intensity
CIE[5] stipulates that there are two kinds of measurement distances for luminous intensity: far field (condition A) is 316mm, corresponding solid angle is 01001Sr; near field (condition B) is 100mm, corresponding solid angle is 0101Sr: both The values ​​can be converted to each other, and the far field measurement result is multiplied by 10 to obtain the near field measurement result.
The measurement distance specified by IEC [6] is only the near field (condition B), the solid angle is < 0101Sr; (why not written equal to 0101Sr ?).
For measuring distances, the CIE specifies the sensitive surface from the top of the LED housing to the photodetector. The IEC regulations are rather vague.
Comment: On the surface, far-field measurements are sometimes less uncertain than near-field measurements because the requirements for measuring distance, current, and stray light can be relatively low under these conditions. At the same time, however, the effect of the LED mounting angle is relatively large, which is a large source of error. In general, far-field conditions can be applied to strong lighting and packaged products, while near-field conditions apply to weak lighting and chips, indicator lights, and backlights. Therefore, there is a need for existence. Second, CIE is not logically rigorous. First of all, it uses a lot of space to explain that the LED is not a point source, so the inverse square law is not true, and the concept of luminous intensity is not applicable. In addition, in practice, it actually measures the illuminance of the LED on the sensitive surface of the photodetector, and then multiplies the square of the distance to obtain the illuminance. Thus, the inverse square law of the distance that was once denied is actually still used.
1.3 External conditions of measurement
The CIE specifies that the measured ambient temperature is 25 °C, but no mention of atmospheric conditions.
The IEC also mentions the appropriate atmospheric conditions while referring to the ambient temperature in general terms.
Comment: According to the author's long-term experience in silicon photodiodes, LED parameters are positively related to atmospheric conditions such as atmospheric pressure, humidity, cleanliness, etc., but they have not been verified by experiments in this area.
2 IEC mentions what was missed by CIE
2.1 CIE does not mention the measurement of LED radiation intensity, which is actually a very important physical quantity.
The IEC refers to this in a larger space. First, it is pointed out that the measurement of the radiation intensity should be the direction of the mechanical axis (normal radiation intensity); second, the use of non-spectral selective detectors such as thermocouples and other thermoelectric detectors as standard detectors: , mentioned the use of near field measurements. Radiation intensity is especially important for infrared light-emitting diodes (IRED).
2.2 CIE does not mention pulse measurement
Since there are "flash LEDs" on the market, the flash frequency is about 113-512 Hz, so there must be a corresponding measurement method.
The IEC stipulates that for the measurement of flash LEDs, the rise time of the photodetector should be small enough and the peak value of the pulse should be readable.
The IEC specifies that in order to measure the bandwidth of the peak wavelength, the wavelength resolution and bandwidth of the monochromatic LED should be adjustable so that the measurement has sufficient accuracy. The spectral response of the radiometer is corrected, and it is assumed that the peak wavelength is 100% for normalization. The IEC states that if the transmission factor of the monochromator and the sensitivity of the radiometer are not constant over the measured wavelength range, the measured values ​​are corrected accordingly.
2.3 CIE does not mention the measurement of LED dark current
The IEC emphasizes that the measurement of LED dark current is very sensitive to temperature, so the measurement accuracy is greatly dependent on the ambient temperature. In addition, stray light is also a factor, and measurement is best performed in an all black environment. The operating current must also be adjusted slowly from zero.
2.4 CIE does not mention the frequency of total capacitance measurement
An optocoupler consisting of an LED and a photo transistor in which the LED acts as a switch. Its anti-interference ability is stronger than the general method. The minimum resolution of the capacitance meter is PC1PC, and the test frequency of the capacitor is specified as 1MHz, which may be useful for the measurement of the total capacitance of the LED.
The current standard only refers to the term "prescribed frequency" in general terms. There is no regulation on how much frequency is used to measure the total capacitance.
2.5 CIE does not mention the switching time of the LED
The IEC explicitly mentions the switching time of the optocoupler. Such as rise time tr, turn-on time ton, turn-off delay time td, fall time tf, turn-off time toff, and so on.
3 CIE mentions what was missed by the IEC
Compared with the previous section, there is too much content in this area, so it can only be reached.
3.1 The nature of the LED
3.1.1 Optical properties of LEDs
• The spatial distribution of luminous intensity emphasizes that the packaging of the chip often changes its original spectral and spatial distribution.
• The spectral distribution of the LED is neither monochromatic (as compared to the laser beam) nor broadband (as compared to incandescent lamps) relative to other sources.
◠The light-emitting area of ​​the LED has different sizes and shapes due to the difference in packaging and lenses, mirrors, and the like. Compared to the measured distance, the LED's illumination area is so large that it is no longer a point source.
3.1.2 Electrical properties of LED
â— Forward voltage: It is best to use a 4-wire measurement method for accurate measurement, and the forward voltage will decrease as the die temperature rises. Due to the heating effect of the LED dies, the radiant power rises faster than the electric power when it is turned on. During steady operation, the LED's optical radiation is strongly dependent on the forward current.
â— Due to the complex relationship between temperature and voltage, it is not recommended to use stable electric power as a prerequisite for stable LED light output power.
The forward voltage is related to the semiconductor material constituting the LED: under the premise of a forward current of 20 mA, the terminal voltage of the infrared tube is 112 V, and the blue tube is 615 V, and other tubes are within this range.
◠CIE stipulates that the ambient temperature during measurement is 25 °C, but the contact point of the LED power supply, the length and resistance of the lead between the die and the heat sink will significantly affect the measurement results.
• CIE also mentions that radiation efficiency is a function of current. Because the increase in operating current will not only increase the output of the LED optical power, but also increase the chip temperature, which in turn will affect the optical power output of the LED. If the operating current is modulated, the chip temperature will fluctuate, resulting in an average optical power output that is not equal to the optical power output of the LED at steady state at the same current.
â— Constant current and a temperature stable voltage will cause the LED to have a stable electrical power consumption. It should be noted, however, that stabilizing the electrical power without controlling the temperature rise will result in a very different LED operating condition. The relative power distribution of the radiation will affect the following two aspects: one is that it will slightly change the shape of the light intensity distribution; the other is that when the temperature rises, the entire distribution will move significantly toward the long wavelength (but the blue LED is generally short wavelength) Move in direction).
3.1.3 Effect of temperature on radiation
â— The surface appears to have a constant current and voltage that seems to provide a constant electrical power to the LED. However, it is impossible to obtain a stable optical power output if the temperature is not controlled. The reason is: the relative power distribution of the LED will change slightly with temperature on the one hand, and the whole distribution will drift toward the long wavelength (when the blue LED is drifting to the short-wave direction) when the temperature rises.
◠As long as there is a change in temperature, the radiant flux emitted by the LED is always in a state of change, that is, the efficiency is changing. For a green monochromatic tube, the change is smaller because it is near the peak of the ∨(λ) curve. For red and blue monochromatic tubes, the change is larger because it is at the end of the ∨(λ) curve.
• The rate at which the die temperature rises depends on the input electrical power and the heat capacity of the LED package product. The temperature of the die after thermal balancing is primarily dependent on the heat dissipation of the die through the pins to the environment. Therefore, the thermal performance of the LED structure, lead length and heat sink determine the die temperature.
â— At high ambient temperatures, the forward voltage drops when the current is fixed. Adjusting this to stabilize the LED's electrical power consumption affects the temperature of the chip, which affects the terminal voltage of the LED. Therefore, the stability of electric power cannot be used as a means to improve the output of LED light power.
3.2 Causes of error
Because the mechanical and optical axes of the LEDs rarely overlap, and there is always a difference in the area, shape, size, and structure of the LEDs, it is difficult to determine the center of illumination of the LEDs, which makes it difficult to align the angle and position. The uncertainty of the measurement.
3.3 The nature of photodetectors
CIE recommends the use of "silicon photodiodes". There are two main selection criteria: one is the uniformity of surface response, that is, the test signals output at each point on the sensitive surface are the same; the other is that the response is independent of the incident angle of the measured light.
3.4 LED luminous intensity measurement
â—LED is not a point source, and the inverse square law is not applicable.
â— Because the LED does not have a true luminous intensity, the "average luminous intensity" is put forward under the compromise.
◠The far field (condition A) is 316 mm, and the corresponding solid angle is 0.001 Sr. The corresponding plane angle is 2°. The near field (condition B) is specified to be 100 mm, and the corresponding solid angle is 0.01 Sr. The corresponding plane angle is 615°. The relationship between the two can be related to each other by a factor of 10.
3.5 Measurement of luminous flux
â—Variable angle photometer method: the top of the LED is the imaginary sphere, the distance between the photodetector and the LED is the imaginary sphere radius, and then the imaginary sphere is divided into N parts. If the illuminance of each part is the same, then the integral or accumulate It is easy to get the total luminous flux. This is the most accurate absolute measurement method, and the disadvantage is that the measurement time is long.
â—Integral ball method: This is a relative measuring instrument, so it must be scaled beforehand. In order to eliminate the LED self-absorption error, CIE recommends inserting an auxiliary LED into the integrating sphere for simultaneous measurement.
3.6 Measurement of spectral quantity
â— Unlike other standards, CIE adds the concept of "central wavelength": the wavelength corresponding to the midpoint of the half-width of the waveform. This solves the confusion of the wavelength description of the light distribution curve of many monochromatic tubes recessed in the normal direction. Others such as the main wavelength, the peak wavelength, and the centroid wavelength are mentioned in various standards, and this is omitted.
â— Pay special attention to the centroid wavelength: It is strongly defined by the shape and distribution of the LED spectral distribution. Even with very small changes in the spectral shape (especially in the UV and IR), the position of the centroid wavelength can be significantly changed.
4 Some suggestions for CIE and IEC
4.1 About the area of ​​the sensitive surface of the silicon photodiode
The CIE states that the area of ​​the sensitive surface is 100 mm2 (corresponding diameter is Φ 1.13 mm) and the sensitive surface must be circular. Domestic silicon photodiodes are currently not suitable for quality reasons. What about foreign products? The Hamamatsu S1337 windowless series of silicon photodiodes have a sensitive surface size of 10 × 10 (mm2 ) and silicon made by UDT and EG&G. The photodiode is also roughly the above size. If 10mm is selected as the diameter of the sensitive surface of the photodetector, the corresponding sensitive surface area is only 7815mm2, which is 2.15mm2 smaller than the CIE specification, and the area is reduced by more than 20%, causing confusion. If the S6337 windowless series is used, its sensitive surface area is 18 × 18 (mm2), and taking 1.13mm is too wasteful. And the price of S1337 is 1 000 yuan / only, while the price of S6337 is 4 500 yuan / only. Therefore, it is strongly recommended that CIE change the diameter of the sensitive surface of the photodetector to 10mm.
4.2 About the spectral sensitivity of the photodetector and the matching accuracy of the ∨(λ) filter
The matching result of the two should conform to the clear visual function curve promulgated by CIE1924, and the deviation is generally expressed by f1. CIE requires f1 <1.5%, which is too demanding. Because the world's best ∨ (λ) filter is currently done in Germany, f1 is 1.5%. At present, the domestic level of ∨(λ) filters is subject to materials (colored neutral glass) and process conditions, which can only be 4% to 5%. According to the national conditions, we think that f1 is set at 2.2%.
4.3 About the bandwidth of the spectral distribution of the monochromatic tube luminous intensity
Bandwidth is an important factor in measuring the radiant intensity of an LED, and the area under the corresponding waveform is the total luminous flux of a monochromatic tube. But mentioning the spectral distribution must involve a monochromator. Essentially a monochromator is a variable wavelength filter. According to the filtering theory, the output signal is a convolution of the input signal and the transfer function of the instrument itself. Only when this convolution is removed can the correct waveform be obtained. Therefore, when a monochrome LED passes through a monochromator, the bandwidth is inevitably increased, commonly known as "instrument widening." This phenomenon can only be handled correctly by the slit function [7]. As for the LED colorimeter of the grating plus linear array CCD, the instrument widening effect is more obvious and will not be discussed here. In addition, it is recommended that the CIE specify the width of the monochromator into the P exit slit because the bandwidth of the waveform is also closely related to this.
4.4 Color temperature correction
Considering the national benchmark for luminous intensity and the national sub-reference for total luminous flux, most of the 2 856K A light source is used as the standard light, and the color temperature of the LED is currently generally between 4 000 and 7 000 K. Therefore, the color temperature correction must be performed, and the correction amount is generally It is about 3%. The prerequisite for color temperature correction is that the spectral response curve of the photodetector used must be known.
5 Conclusion
As the development of standards can not keep up with the pace of development of the LED industry, the measurement standards (draft) at home and abroad can not meet the requirements of the industry. Limited to the level, although there are many errors, the author still hopes that this article will contribute to the formulation of formal standards.
(Edit: Xiaotang)
Incremental Encoder,Linear Scale Encoder,Dual Concentric Rotary Encoder,Dc Servo Motor Encoder
Yuheng Optics Co., Ltd.(Changchun) , https://www.yuhengcoder.com