Ordering a white LED by kelvin alone is the fastest way to get a part that measures right and looks wrong. Color lives on a two dimensional map, and the correlated color temperature printed on a datasheet locates only one axis of it. The two numbers that finish the specification are Duv and the chromaticity bin, and both are cheaper to agree on before sampling than to argue about after.
CCT is the temperature of the point on the Planckian locus closest to the chromaticity of the part, computed in CIE 1960 UCS space. It says how warm or cool the light reads against a blackbody radiator. A 3000K part sits near the locus at the 3000K point; the part itself can sit on the locus, above it, or below it, and the CCT number does not say which.
Duv: The Distance from the Locus
Duv is the perpendicular distance from the Planckian locus in CIE 1960 UCS space. A positive Duv places the tint on the green side of the locus; a negative Duv places it on the pink side. Two 4000K emitters with the same luminous flux can differ visibly when one measures +0.006 and the other -0.004: over a white wall the first reads slightly green, on skin the second reads pink. Cinema and museum lighting write Duv tolerances of 0.002 or less for exactly this reason.
ANSI C78.377 draws nominal chromaticity quadrangles around the locus at each standard CCT. Inside one quadrangle the tint range is wide, which is why binning exists on top of it.
The Bin Map
| Term | What it fixes | Typical range |
|---|---|---|
| CCT | Warm or cool reading | 2200K to 6500K nominals |
| Duv | Green-pink offset from the locus | +/- 0.006 standard, 0.003 tight |
| Chromaticity bin | Position inside the quadrangle | 4-step or 2-step MacAdam ellipse |
A chromaticity bin fixes where the part sits inside the quadrangle. A 4-step bin means the whole bin fits inside a set of MacAdam ellipses centered on the locus; a 2-step order halves the spread. The 2-step version costs more, because yield inside a smaller boundary is lower, and the premium is usually smaller than the cost of mixing mismatched reels on site.
Where the Color Drifts
Three physical effects move a part off its bin point. Phosphor load tolerance shifts the starting point across production, and one reel can sit near an edge of the bin. Junction temperature shifts the tint at power: a part measured at 25C reads a different Duv at 85C, commonly by 0.001 to 0.002 for phosphor-converted parts. Phosphor behavior at temperature also trims the deep red first, which pulls the tint warmer and greener at once. Drive current adds a smaller shift, and reflow can move the dome geometry enough to read on a spectrometer.
Writing the Spec
- State CCT, Duv and the bin name together. “3000K, Duv 0.000 +/- 0.002, bin per map” is a spec; “3000K” is a wish.
- Ask for the bin map, not only the bin name, so incoming inspection can plot samples against it.
- Agree on the test current and soak time, because tint at 25C is not tint at operating temperature.
- For mixed batches, order a single bin run. Parts from one production lot match better than parts drawn from two.
- Keep the thermal path short on the board. A starved pad runs the junction hot and pushes the tint green.
- CCT alone fixes one axis; Duv and the bin fix the rest of the color call.
- ANSI C78.377 quadrangles are wide; bins exist because a quadrangle is not a usable tolerance.
- Tint moves with junction temperature and phosphor behavior, so the test condition belongs in the spec.
- One bin run per project beats sorting reels after delivery.
