How to choose the right LED light source?

LED luminaire options include appearance, heat dissipation, light distribution, glare, and installation. We don't talk about the parameters of the luminaire today, only the light source: Do you really pick a good LED light source? The parameters of the light source are: current, power, luminous flux, light decay, light color, color rendering. Our focus today is on the last two items. Let’s briefly talk about the first four items.

First of all, we often say: "How many watts of light do I have?" This habit is to continue the traditional light source. At that time, there were only a few fixed wattages of light source, which could only be selected within those wattages and could not be adjusted freely. And now the LED, the drive current is slightly changed, the power will change immediately! Are you still asking for power? Beware! The same LED light source, when driven with excessive current, the power is up, but it will cause the light effect to decrease and the light decay to increase. Please see the picture below:

In general, redundancy = waste, but for LED operating current, it is a savings. With the drive current reaching the rated allowable maximum, a 1/3 drive current is reduced, and the sacrificed luminous flux is very limited. However, the benefits are huge:

Light attenuation is greatly reduced;

Life expectancy has been greatly extended;

Significant improvement in reliability;

Higher power utilization.

Therefore, for a good LED light source, the drive current should use a maximum rated current of about 70%.

In this case, the designer should directly request the luminous flux. As for the wattage used, it should be decided by the manufacturer himself. This can promote manufacturers to pursue efficiency and stability, rather than blindly pushing up the wattage of the light source, and sacrificing efficiency and longevity.

The above mentioned parameters include: current, power, luminous flux, and light decay. They are closely related, and you should pay attention to the use: Which one is what you really need?

Light color

In the era of traditional light sources, when it comes to color temperature, everyone only cares about "yellow light, white light" and does not care much about the problem of light color deviation. Anyway, the color temperature of the traditional light source is just a few. If you choose one, you will not go anywhere. In the era of LEDs, we found that the color of LEDs is different. Even the same batch of lamp beads, there may be deviations to strange, red, green and green lively ~~

Said that LED is good, energy saving and environmentally friendly. But are you making it? There are really a lot of companies that rot LEDs! The following is a real-life view of a famous domestic brand LED lighting application issued by V friends. This is the light distribution of this family, the consistency of this color temperature, and the faint blue light...

In view of this chaos, there is a conscience LED lighting factory promises to customers: "Our lamps, color temperature deviation within ±150K!", there are also design companies in the product selection, the specification notes: "Bead color temperature requirements deviation Within ±150K".

This 150K is based on the conclusion of the traditional literature: "The color temperature deviation is within ±150K, which is difficult for the human eye to detect." They believe that by setting a color temperature of "±150K or less", it is possible to avoid red, green and green things. In fact, it’s really “not so simple”...

First, a chestnut, in the aging room of this factory, I saw two sets of light bars with distinct light colors, one group was normally warm white, and the other group was obviously greenish. As shown in the figure, can you find the difference between the two light bars? One red and one green, is there wood? According to the above statement, the human eye sees a clear difference, and the color temperature difference is definitely higher than 150K, right?

I tested them separately. The result is this...

did you see? The two light sources that look completely different from the human eye, the "correlated color temperature" is only 20K apart! Do you really think that... the color on the right is 20K higher, is it seen by you?

Isn't the conclusion that the color temperature deviation is within ±150K and the human eye is difficult to detect? Do not worry, listen to me slowly explain: first talk about the two concepts of color temperature (CT) vs correlated color temperature (CCT). We usually say that the "color temperature" of the light source is actually quoted in the column "Related Color Temperature" on the test report. The definition of these two parameters in "Architectural Lighting Design Standard GB50034-2013":

Color Temperature

When the chromaticity of the light source is the same as the chromaticity of the black body at a certain temperature, the absolute temperature of the black body is the color temperature of the light source. Also known as "chroma". The unit is open (K).

Correlated Colour Temperature

When the chromaticity point of the light source is not on the black body trajectory, and the chromaticity of the light source is closest to the chromaticity of the black body at a certain temperature, the absolute temperature of the black body is the correlated color temperature of the light source, referred to as the correlated color temperature. The unit is open (K).

First, give an easy-to-understand chestnut to explain that the horse is "related." Assume that our country has only one railway: the Beijing-Guangzhou line. In the early years, he was only able to travel by train. The wife called to check the post: "Where are you now?" I happened to be in Zhengzhou, and I replied: I am in Zhengzhou. Tomorrow she asked again, I said: now Wuhan. So, we are used to using Beijing, Zhengzhou, Wuhan, Guangzhou to mark their location.

It is also assumed that one day there will be money, buy a car to drive, leave the Beijing-Guangzhou line, and go to Xi'an. He was also investigated, his wife had no culture, only knew the Beijing-Guangzhou line, and he did not know where Xi'an was. I can only find a place that is "closest" from the Beijing-Guangzhou line and tell her: I am in Zhengzhou. You said that I crap: "Xi'an is far from Zhengzhou, you are too embarrassed to say?" But, yes! I am talking about "related location"! If you are in Nanjing, you can say: "I am in Wuhan!"

The above is purely a paragraph, indicating the relationship between "location" and "related position". Let's go back to the color temperature.

The latitude and longitude on the map indicates the location of the city, and the (x, y) coordinate values ​​on the "color map" indicate the position of a certain light color. Looking at the picture below, (0.1, 0.8) is pure green, and (0.7, 0.25) is pure red. The middle part is basically white light. This "white degree" cannot be described by words, and there is a concept of "color temperature".

The light emitted by the tungsten bulb at different temperatures is represented as a line on the color coordinate map, called "black body track", abbreviated as BBL, also called "Planck curve". The color of the light emitted by the black body radiation appears to us as "normal white light." Once the color coordinates of the light source deviate from this curve, we think that he is "color cast".

Our earliest tungsten light bulb, no matter how it is done, his light color can only fall on this line of cold and warm white light (the thick black line in the picture), we call the light color at different positions on this line. Color temperature." Now that technology is developed, the white light we have made, the light color does not necessarily fall on this line. We can only find a "recent" point and read the color temperature of this point, which is called his "related color temperature." Now you know it? Don't say the deviation is ±150K. Even if the two light sources are the same, the light color may be very different. Like the "Xi'an vs Zhengzhou" mentioned above...

Zoom in on the 3000K "isotherm":

what? You still don't understand? Then I can't help it... In short, I believe Xugong said: "It is not enough to say that color temperature is good. Even if everyone is 3000K, there will be a possibility of reddish or greenish." When he was in Zhengzhou, he said that he might be in Xi'an, hehe~

Then... what should I look at?

Let's learn a new indicator: SDCM.

Still using the above chestnuts, these two sets of light bars, their "related color temperature" only differs by 20K! It can be said that it is almost identical. But in fact, they are obviously different colors. where is the problem?

The truth is this: let's take a look at their SDCM diagram:

The picture on the left is the warm white 3265K on the left. Please pay attention to the small yellow dot on the right side of the green ellipse, which is the position of the light source on the chromaticity diagram. The picture on the right is green on the right, and his position goes outside the red ellipse.

can not read it? We still use the chestnuts of the last Beijing-Guangzhou line.

Last time I was in Xi'an, I told my wife to be in Zhengzhou. Later, she was smart, learned to look at the map, and knew that there was a vast territory in addition to the Beijing-Guangzhou line. Draw a few circles on the map: yellow, blue, green, and red, respectively, 50, 100, 150, and 200 kilometers away from the "Zhengzhou City Center." Regulation: In addition to the city, the distance from the city center should be reported in the future. If it is more than 200 kilometers away, it cannot be said to be "in Zhengzhou."

This "distance circle" is used in the chromaticity coordinate diagram, which is the concept of SDCM.

First look at the position of the two light sources on the chromaticity diagram. The values ​​closest to the blackbody curve are 3265K and 3282K respectively. It seems that they only differ by 20K, but in fact they can be far away~ and Xi'an is away from Zhengzhou. Almost ~

There is no 3200K line in the test software, only 3500K, we draw 3200K circles ourselves:

The four circles of yellow, blue, green and red respectively represent the distances of "perfect light color" from 1, 3, 5, and 7 "steps".

When the difference in light color is within 5 steps, the human eye is basically different, which is enough. The new national standard also stipulates: "The color tolerance of the same type of light source should not be greater than 5 SDCM."

Let's see: The following point, the distance from the "perfect" light is just within 5 steps, we think it is a pretty light color. The point above, has already ran out of 7step, and the human eye can clearly find his color cast.

We will use SDCM to evaluate the color of light, so how to measure this parameter? I suggest you bring a spectrometer with you...not a joke, portable spectrometer! For outdoor lighting, the accuracy of light color is especially important. Reddish green is ugly. Therefore, we must keep our eyes open, always ready to use the eyes of the fire to catch unqualified products.

Color rendering index

Outdoor lighting that requires high color rendering index is the lighting of buildings, such as wall washers for building surface lighting and floodlights for landscape lighting. A low color rendering index can seriously damage the aesthetics of the illuminated building or landscape.

For indoor applications, the importance of the color rendering index is particularly evident in residential, retail, and hotel lighting applications. For the office environment, the color rendering characteristics are not that important, because office lighting is designed to provide optimal illumination for the execution of the work, not for aesthetics.

Color rendering is an important aspect to evaluate the quality of lighting. Color Rendering Index is an important method to evaluate the color rendering of light sources. It is an important parameter to measure the color characteristics of artificial light sources. It is widely used to evaluate artificial lighting sources. Different flower effects under Ra:

Generally speaking, the higher the color rendering index, the better the color rendering of the light source and the stronger the color reproduction ability of the object. However, this is only usual. Is this really the case? Is it absolutely reliable to evaluate the color reproduction power of a light source with a color rendering index? Under what circumstances will there be an exception?

In order to clarify these issues, we must first understand what the color rendering index refers to and how it is derived. Just like the company in order to better assess your work, it will set a series of KPI indicators for you, and establish a specification according to a reasonable model, and then score the same. CIE also well defines a set of evaluation light source color rendering. The method uses 14 test color samples, and a series of spectral brightness values ​​are obtained by standard light source test, and its color rendering index is specified to be 100. The color rendering index of the evaluated light source is scored according to a set of calculation methods compared to the standard light source. The 14 experimental color samples are as follows:

Among them, No. 1-8 was used for the evaluation of the general color rendering index Ra, and eight representative color tones having medium saturation were selected. In addition to the eight standard color samples for calculating the general color rendering index, CIE also provides six standard color samples for calculating the color rendering index of the special color, which are selected for some special color rendering performance of the test source, respectively. Higher red, yellow, green, blue, European and American skin color and leaf green (No. 9-14). China's light source color rendering index calculation method also adds a color sample R15 representing the skin color of Asian women.

The problem is: usually we say that the color rendering index value Ra is based on the color rendering of the light source for eight standard color samples. The eight color samples have medium chroma and lightness, all of which are unsaturated colors. They are used for It is a good result to measure the color rendering of a light source with a continuous spectrum and a wide frequency band, and it is problematic for evaluating a light source with a steep waveform and a narrow frequency band. Take the KPI as an example. The company evaluates what it wants based on what it wants. However, there are countermeasures under the policy, and employees will deliberately show what they are doing because of the company’s assessment. So, if the KPI score is high, is this employee really good? Is the color rendering index Ra high, and the color rendering property is good?

Li Li: The following two pictures, the first line in each picture is the performance of the standard light source for various color samples, and the second line is the performance of the tested LED light source for various color samples.

The color rendering index of these two LED sources is calculated according to the standard test method:

The top one is Ra=80 and the bottom one is Ra=67. Accident? the reason? I have already mentioned it above.

So, is there a new and better way to evaluate the color rendering of a light source? The answer to NIST is CQS (Color Quality Index Method). Similar to the color rendering index (CRI), CQS also uses the test color method, but CQS selects 15 saturated colors, which are evenly distributed throughout the visible spectrum, as shown in the following figure.

Either way, there is a possibility that it does not apply. So, if we have a space that is very strict with color, what method should we use to judge whether a certain light source is suitable for use? My approach may be a bit more complicated: look at the source spectrum.

The following are the spectral distributions of several typical light sources, namely daylight (Ra100), incandescent (Ra100), fluorescent (Ra80), a certain brand of LED (Ra93), metal halide lamp (Ra90).

From the spectral distribution, we can easily analyze that although the color rendering index Ra of the test is greater than 85, it is excellent, but in fact, their true color reproduction ability is different.

Therefore, when focusing on the color rendering of a light source, we first look at Ra. For most LED sources, it is usually discontinuous and the band is narrow, depending on the spectral distribution. In places with rich colors and high requirements, such as museums, supermarkets, shops, and restaurants, we should choose a light source with a relatively continuous spectrum and a wide frequency band, so as to truly restore the true, beautiful and comfortable colors of the objects, and help the items. Display and sales and the creation of an environmental atmosphere.

Therefore, LEDs only have chromaticity coordinates and SDCM is not enough. The chromaticity coordinate value can only quantitatively characterize the color performance of the LED when it is viewed directly, but it does not reflect the color performance of the object when it is illuminated by the LED.

As shown, the blue and red curves are the spectral power distributions of incandescent and LED lights, respectively, with a color temperature of 3000K and a CIE coordinate of x=0.437.y=0.404.

Both sources have exactly the same chromaticity coordinates, and their Spectral power distributions (SPD) are very different. We look directly at the light emitted by the two, and the colors look exactly the same. But if you use them to illuminate other objects with color, the illuminated objects will look different colors. The greater the difference in spectral power distribution, the greater the difference in color that can be exhibited when the same object is illuminated by them.

The reason is that the color exhibited by the object is not determined solely by its nature. Color is the product of the complex interaction between the spectral properties of the source, the spectral reflectance of the illuminated object, and the spectral sensitivity of the human eye. The human visual system further adjusts the signal from the retina to produce the final color perception. When we talk about color in everyday life, it seems that this is a property of various objects.

For example, apples are red and bananas are yellow. In fact, the inherent characteristics of making apples appear red, stemming from the fact that apples reflect more long-wavelength light, while the mid-reflection and short-wavelength rays are poor. The reflected light wavelength of bananas is concentrated around 580-590 nm. If we use a red LED to illuminate the banana, the banana will turn red. As shown in the figure, the spectral power distribution of the light wave emitted by the bulb is modified by the spectral reflection characteristics of the light blue balloon color to form a new spectral power distribution and enter the human eye.

When an object is successively illuminated by two light sources, the color it exhibits may change dramatically. Even if the two light sources have exactly the same chromaticity coordinates, as long as their spectral power distributions are inconsistent, the spectral power distribution of the reflected light will basically not be mutually metameric, so it will also show color inconsistency, if not Obviously, at least slight changes will occur.

Color reproduction is especially important for indoor LED lighting. The spectral power distribution of a typical white LED is quite different from the usual indoor lighting (whether incandescent or fluorescent). Consumers may be disappointed to find that when they install new LED bulbs at home or in the office, the colors of the things they are familiar with can change dramatically.

In the past, the lighting industry has faced this problem and proposed the chrominance index (CRI) metric, which is characterized by incandescent light sources (low color temperature range, <5000K) and daylight (high color temperature range, >5000K). For comparison, quantitative analysis of the color reproduction characteristics of the light source. Under the light source under consideration, the color rendering or color reproduction is exactly the same as that under daylight or incandescent light, and the color rendering index is 100. The color rendering index is not the perfect predictor of the color reproduction performance of the light source, but it does practice a reasonable job.

A total of 14 color patches were used for the color rendering index measurement. Of these, 8 are used to obtain actual color rendering index values, and the remaining 6 are used to provide specific measurements for specific colors. The choice of color chips is a typical common material. Their reflection characteristics are shown in the figure.

The color rendering index Ra uses only eight spectral reflection distributions (color patches) to represent an infinite number of spectral reflection distributions that real objects that may be encountered in practical applications may have. This is a big flaw.

The other six color patches included in the color rendering index standard can be used to some extent to compensate for the lack of a few spectral reflectance distributions used in the standard. R9 values ​​for saturated red are often very valuable for white LEDs because white LEDs lack long wavelength light and often do not restore pure red well. While comparing the color rendering index values ​​of two LEDs, it is meaningful to compare their R9 values, especially for applications where red reproduction is important.

The new national standard stipulates: “The color temperature should not be higher than 4000K for a room or place where work or stays for a long time, and the special color rendering index R9 should be greater than zero”.

As shown in the figure: the color rendering of fluorescent lamps with different color rendering indexes, in which the yellow color is basically no difference, but the red (R9) is very different.

Red is especially important for us. The ability of the light source to red color directly affects the color rendering ability of the light source for our human skin color, and we especially care about the red object and the color of our skin color under the light.

The figure below shows the effect of the same food under different R9 color rendering index sources.

R9 is even more important for LEDs. With the rapid development of LEDs, we are able to adjust the spectrum of white LEDs more and more conveniently. In order to improve the light efficiency, everyone has reduced the energy in the longer wavelength region, that is, the energy in the red wavelength region, and put more energy at around 555 nm. In this way, the color rendering ability of the red color for red, especially for the higher chroma, will decrease, and this drop cannot be expressed by CRI, which is why R9 is especially important for LEDs.

The picture shows the different effects of mahogany of the same texture color under different light sources. In the middle and right, the wood color looks so much different! Just because of the color temperature of 2700K and 4000K, will it cause such a big difference? In fact, the spectrum is at play. Look at their spectral differences: the one on the right, the red part, basically no more...

Above we talked about a few small points in terms of light color, in fact, there are many factors that need attention, such as color saturation, performance on white objects. When choosing a light source, we can't easily choose it with a few inherent parameters, such as "color temperature deviation within ±150K", which is the sword and the rubber string.

As a professional lighting designer, it is necessary to select the appropriate light source for lighting according to the characteristics of the design site and the characteristics of the object to be illuminated.