High-brightness LEDs (HBLEDs) are increasingly becoming
the light source of choice in both general and specialty
lighting applications. Advances in LED technology have led
to higher lumens per watt. Improvements are also being made in
package size, color options, color rendering index (CRI) ratings,
binning, and temperature stability. LEDs bring flexibility, efficiency,
and intelligence to any lighting application.
A typical application of such a system is general household lighting,
where users can create multiple shades of white or colored
light with just one fixture. Such flexible lighting fixtures enrich the
end-user experience by providing control over light that consumers
have never experienced before.
Two very important aspects enable the flexibility of these features:
the possible color gamut of the fixture and the number of
unique mixed colors the fixture can create, which is known as color
resolution. For some applications, users desire as large a potential
color gamut as possible. This enables the fixture to create more
vibrant colors that don’t appear to be washed out.
Color gamut and color resolution can vary for any given system
(Fig. 1). The black triangle and “+” symbols represent a lighting fixture
with a limited gamut and smaller set of unique mixable colors.
The white shape and “+” symbols represent a fixture with a larger
color gamut and a higher color resolution. Fixtures with a larger
color gamut and higher color resolution are more desirable.
One other key aspect of a white light fixture is its CRI rating,
which grades how well colors and objects look when they’re illuminated
with the fixture. CRI is better when a light fixture emanates
more unique wavelengths of light for a given mixed color.
Each of the three aspects—gamut, color resolution, and CRI—
can be optimized by a single design factor: the number of uniquely
colored LED channels in the system. Larger numbers of LED color
channels increase the possible color resolution exponentially with
each channel added. They also increase the color gamut, since
the different LED colors will cover a greater area of the color space.
Finally, using more wavelengths of light to mix colors boosts the
CRI rating of the fixture.
Four-channel color combinations often work well for LED color
mixing fixtures. Two common combinations are RGBA and RGBW.
(“A” is for amber, and “W” is for white.) The RGBA combination
provides a larger gamut than RGB or RGBW, and it generally creates
light with a good CRI. RGBW doesn’t have a larger color gamut,
but it has a good CRI and more of the primary mixed color that’s
desired—white. The number of channels needn’t stop at four. LED
lighting fixtures with five, six, or even seven channels are sometimes
necessary for very high-performance or high-end systems.
While having more independent LED channels
presents clear advantages, there are
drawbacks, too. These include the obvious
need for more hardware (LEDs and drivers)
and an embedded microcontroller with more
complex firmware. In such a system, the controller
continuously calculates the appropriate
dimming levels needed for each LED color
channel. The output for each channel must be
finely adjusted to mix to the proper color.
In a multichannel system, the processor in
the embedded microcontroller receives the
request for a particular mixed color from a
data network or some other interface (Fig. 2).
The processor must then calculate the dimming
values needed for each LED channel to
create the mixed color. This calculation process
should only be as complex as needed so
it doesn’t burden the processor.
The system requirements will determine
the optimal calculation process. One question
needs to be answered—how many unique
colors must the fixture be able to create? Put
another way, can any color in the gamut be
requested, or only a small subset of colors?
For instance, a lighting fixture that just creates
various shades of white may only need to create
100 or so unique colors ranging from warm white to cool white.
In this case, it’s advantageous to design the microcontroller’s
firmware to calculate the dimming values with a lookup table
(LUT). For each unique mixed color, a set of dimming values is predefined
and stored in flash memory. When the processor receives
a mixed color input, it looks up the appropriate values in the LUT.
This fast and simple method allows very complex calculations to
be preprocessed so the microcontroller doesn’t have to make timeconsuming
calculations on the fly.
A LUT method is best if it’s feasible. It
becomes unfeasible when many unpredictable,
unique mixed colors are required.
For example, a color mixing system that
has four LED channels with an 8-bit dimming
resolution for each can create more
than 4 billion unique colors. For each
of these unique colors, the LUT would
need to store one 8-bit dimming value for
every dimming channel, requiring more
than 16 Gbytes of memory. In general, a
fixture with firmware using a LUT for color
mixing can only support as many unique
mixed colors as the memory size allows.
With color-mixing algorithms, the
firmware can generate a large number
of unique color outputs without using a
gigantic LUT. An algorithm can take any
color request input and calculate dimming
values that will create that mixed color. A
general algorithm doesn’t use an excessive
amount of memory, but it’s more complex
to develop and takes a longer amount of
CPU processing time to generate dimming
value outputs.
For an embedded color-mixing system,
choosing the right microcontroller is
essential. It’s best to use more LED color channels in a colormixing
system when light quality or flexibility is important. When
designing the firmware for such a system, choose a LUT method or
a dimming value calculation algorithm appropriate to the requirements
of the fixture.