A Look at the Past and Future of LED Binning

Fig.1 CIE 1931 Colour Space
Fig.1 CIE 1931 Colour Space

 

by Dave Grassi, Lumenpulse Optical Engineering Manager

In the lighting industry, the act of "binning" of LEDs is the process of sorting LEDs by certain characteristics, such as colour, voltage, and brightness. The American National Standards Institute (ANSI) standard C78.377-2017 provides the dimensions and coordinates of the bins that are currently used to standardise the colour points of all white light LEDs. These dimensions are calculated from the Correlated Colour Temperature (CCT) and the distance from the Planckian locus (Duv).  The International Commission on Illumination (CIE) first developed the now well-known 1931 XYZ colour space almost 90 years ago, associating the full gamut of visible colours with how these colours blend together mathematically in illumination applications.  The colour space shown here is 2 dimensions (xy), the true colour space is 3 dimensions, and includes Brightness (xyY) . On the outside of the colour space are saturated colours, ROYGBIV, then towards the middle is the pastel region, and at the centre is the white space.  The curved line in the middle is the Planckian locus, which represents the chromaticity that a blackbody radiator (eg the Sun) demonstrates as its temperature changes.

The blackbody emitter emits a specific spectrum as a function of its temperature, this spectrum is mapped to a chromaticity and then correlated back to the temperate, giving us the Correlated Colour Temperature. These physical temperatures emit a specific spectrum, and as the temperature changes, so does the spectrum.  These spectrums are measured to the closest colour match on the CIE 1931 colour space, and then referenced to a specific CCT. This is what forms the basis for all standard CCTs that we use today in general lighting.  An important part of ensuring quality illumination is ensuring consistency in terms of chromaticity.  Two fixtures may have the same measured CCT but will not appear to be identical colours.  Additionally, just because two colours do not have the exact same X and Y coordinates on this diagram does not mean that they won't appear to be the same.  Dr. David MacAdam conducted many experiments to determine what exactly is a Just Noticeable Difference (JND) between two colours by having his test subjects compare different colour samples until both colours appeared to be identical. The definition of a JND is when a 50/50 split in opinion of all test subjects occurs on whether or not a difference between two colour samples can be discerned.  MacAdam found that in all tests, the matching colours would fall within an ellipse, the size and orientation of which varying depending on where on the colour space it was located. This ellipse forms the basis of all LED colour binning today, which defines the C the max deviation from a target colour point, this is commonly measured in steps. For example, a 1-step ellipse represents one standard deviation from this colour target, and at the extents of this ellipse, a significant portion of the test subjects should not be able to notice a difference.  A 3-step ellipse represents three standard deviations from this colour target, and at the extents of this ellipse, almost everyone (with normal eyesight) would be able to discern the difference.

This knowledge was eventually integrated into ensuring consistent chromaticity for legacy white light sources such as incandescent and fluorescent.  Manufacturers would ensure that their light sources would fall within a 4-step ellipse for a specific CCT. This was deemed acceptable by the market, even though most observers would be able to see the difference in colour from fixture to fixture if the fixtures were within the same line of sight.  When white light LEDs were first introduced, this industry norm was written in a way to ensure a maximum possible yield for LED manufacturers, and as such, utilised 7-step quadrangles instead of MacAdam ellipses.  The logic behind this was that although we use MacAdam ellipses to determine what is an acceptable variation in colour, quadrangles would be able to fill in the gaps left by ellipses. These bins would not necessarily be useful by themselves to create standard colour temperatures, but they would allow LED manufacturers to increase overall yield and would give them the flexibility of creating customer specific chromaticity regions and greater overall possibilities in terms of blending various bins together to create standard CCTs. 

 

To Bin or Not to Bin that is the Question

Fig. 2 – Demonstration of standard binning structure at 3000K
Fig. 2 – Demonstration of standard binning structure at 3000K

From the advent of white light LED manufacturing, every manufacturer has put their own spin on binning with the goal of differentiating themselves from their competitors. The most commonly used structure in North America is known as the 1/16 ANSI C78.377 (super catchy).  This utilises the standard 7-step quadrangle to set the outside limits, and then breaks it down into 16 quadrangles.

This structure gives 16 unique colour points within a single CCT quadrangle, which allows for maximised production yields since almost any LED manufactured will fall into one of these bins.  

 

Fig. 3 – When sourcing “pre-binned” LEDs, in the LEDs will vary in colour point within the outlined ellipses (2-step and 7-step in this figure)
Fig. 3 – When sourcing “pre-binned” LEDs, in the LEDs will vary in colour point within the outlined ellipses (2-step and 7-step in this figure)

The advantages of this structure allow for flexibility within the binning process, in that you can achieve intermediate colour points (3250K) with pinpoint accuracy.  However, successfully managing this flexibility requires overcoming challenges in inventory management. To avoid these challenges, some manufacturers will instead rely on sourcing only such bins that fall within industry standard MacAdam ellipses from the target CCT centerpoint ( between 2 and 5 steps, depending on the CCT) .   The disadvantages of this are the additional costs associated with sourcing these bins, as well as a limited supply of available LEDs, which can result in long lead times from the manufacturer. The bigger problem is that 2-step binning can actually result in fixtures that are 4-step MacAdam ellipses from each other.  For example, in figure 3, bins Seven and Ten are each 2 steps from the CCT centerpoint, but 4-steps from each other.  When you are sourcing these sort of "pre-binned" LEDs, you don't have any specific information on where these LEDs are located related to the target CCT, so if you have fixtures on a project built from different batches of LEDs, each fixture might be only 2 steps from the target CCT, but in different locations, and thus obviously different.  Maintaining complete control over the binning process ensures the sort of batch to batch consistency required to ensure no visible differences in fixture appearance.

 

The Island of Misfit Bins

Most LED products utilise multiple LEDs within a single fixture, so it's possible to mix multiple LEDs from different bins and to integrate the result over the beam distribution.  If a manufacturer can successfully implement a system to manage the mixing of these different bins, the possibilities are endless.   As touched in the introduction, the general population can detect colour differences of 3-steps - and even 2-step differences can be detected if fixtures are in close proximity to each other - so why is this the industry norm?  The simple answer is that most manufacturers lack the inventory management or technical knowhow to reliably deliver products that are quickly and accurately binned while taking the application into account.  Software solutions can be developed that take full accounting of each LED's chromaticity, mix the LEDs together on a circuit board to create less than 2-step deviation from the ideal colour point, create custom colour points, and even take into account the fixture application.

 

Why does application matter for binning?

This fixture has been binned to within 2-step Macadam ellipses on the overall fixture, but in the application, there is not enough distance between the fixture and the target surface to allow for the individual LEDs to integrate, and so each LED bin used here is visible.   The linear fixture in this case is being used to graze, utilising a specific optical distribution, the manufacturer should source only a certain subset of bins to ensure that on an LED level, there are no visible differences.

Lumenpulse has been at the forefront of LED binning since our inception, and with the experience gained, we take all the above factors into account when developing the end product. By being able to accept the full yield from our approved LED manufacturers we can ensure the fastest possible lead times of our fixtures while utilising potentially any LED manufacturer and accepting the widest subset of bins available allows us to deliver best-in-class colour consistency from fixture to fixture.  By being able to reliably and accurately bin these LEDs, we can deliver best-in-class colour consistency from fixture to fixture, while taking application into account, therefore guaranteeing our product against noticeable colour variation.