Using a small MCU to achieve color control of LED lighting

There are still many unsolved problems in LED technology, one of which is how to produce high-quality white light. White light LEDs are composed of blue LEDs and a phosphor that shifts the light output to other parts of the spectrum. Many white light LEDs do not produce a high Color Rendering Index (CRI), which is a parameter used to measure the ability of a light source to reproduce colors faithfully.

Higher quality white light systems can be obtained by mixing two or more colors of LED light. In these multi-color systems, the light output of each color source will drift over time and temperature. Light sensors and small microcontrollers ( MCUs ) can be used to maintain specific colors and related color temperatures (CCT).

There are many small and affordable light sensors on the market that can provide information to the MCU for processing. Typically, the sensor has some optional color filters to measure red, green, blue , or white light (no filter). The light sensor output can be connected to the MCU through a variety of methods. Light-to-voltage sensors are connected to analog-to-digital converters (ADCs) through the output voltage. Light-to-frequency sensors provide a variable frequency output that is proportional to the amount of light. The pulsed outputs of these sensors can be accumulated in an MCU timer to determine light levels. Light-to-digital sensors typically have a serial digital interface, such as I2C. Each type of sensor interface has unique benefits and requires different MCU resources. The system block diagram in Figure 1 shows a variety of MCU peripherals that are useful in color-tunable LED lighting designs.

Figure 1 MCU peripherals that can be used in LED color control systems

In a complete closed-loop color control system, the MCU must read the color content from the light sensor, calibrate the light sensor output, and adjust the output of each LED driver to achieve the desired color. The LEDs require a constant current driver to maintain consistent light output. This can be achieved using a variety of driver technologies, including linear and switch mode solutions. The final choice depends on factors such as efficiency requirements, input voltage range, and number of LEDs used.

The driver output can be controlled using different methods. First, the MCU can generate an analog reference voltage through a digital-to-analog converter ( DAC ) or a digital potentiometer. The reference voltage can vary the driver output between zero and maximum current. The MCU can also provide a PWM signal that modulates the driver output. The PWM signal can be used to enable/disable the driver itself or to control a switch that disconnects the LED from the driver output. If PWM control is used, the PWM frequency should be selected to be high enough so that the human eye cannot detect any flicker.

The designer must determine what control resolution the color control system requires in order to select an MCU with the appropriate peripherals. For light-to-voltage sensors, the measurement resolution of the ADC on the MCU is important. Light-to-frequency sensors require an MCU time base that is incremented by an external clock . Light-to-digital sensors require the appropriate serial communication interface peripherals.

MCUs with multiple PWM peripherals can be used to control individual LED drivers. In high-resolution color control systems, PWM peripherals with 16-bit or higher control resolution are preferred. Serial communication peripherals such as UART, SPI, I2C, LIN, and USB support input/output control and display functions. MCU devices such as the PIC24FJ16GA002 are excellent choices

for color control systems . The PIC24 device has a small 28-pin package, a program memory range of 16 to 64 KB, and provides a serial communication interface, a 10-bit ADC, and 5 PWM channels in a single device. The 16-bit MCU core can easily handle the arithmetic operations related to sensor calibration and color control. The sensor data output must be calibrated to a reference voltage to provide consistent results. The calibration process uses a colorimeter to mathematically relate the output of different color LEDs to the spectral response and sensitivity of the light sensor in a standard chromaticity coordinate system. The calibration process generates a coefficient matrix that must be stored in nonvolatile memory with the lighting system and used to determine the difference between the correlation and the desired output at each control of the control system. Once the calibration is completed, the MCU can compare the sensor data with the ideal CIE (International Commission on Illumination) chromaticity diagram coordinates and adjust the output channels until the ideal CCT is achieved. The PID control algorithm for each output channel adjusts the sensor data using the calibration value, finds the difference from the target set point, and then adjusts the output channel. The PID is run continuously to reduce the error until the output CCT matches the set point CCT. The PID coefficients can be fine-tuned to maximize the system response, but how quickly the PID algorithm converges to the target CCT is also a function of the efficiency of the MCU processing arithmetic operations. Some color control systems may require faster processing speed and response than others. For example, general lighting systems have lower requirements than local dimming systems for HDTV panels . Systems with dimmable light sources or high CRI have a range of user control requirements. Medical devices with graphic LCD displays may have dimmable LED backlights (which require the MCU to communicate with the LCD via SPI) and a touch screen interface for adjusting CCT and brightness . General lighting for commercial display equipment may require control from a central panel or computer to automatically adjust brightness, CCT, and on/off according to the time of day. Communication between these devices may be achieved using hardwired serial bus protocols such as DALI or DMX512, while some other devices may require custom interfaces implemented over USB or Ethernet . In completed buildings, installing a hardwired infrastructure may not be feasible and control may need to be done via wireless communications and protocols such as ZigBee. For such lighting applications, MCUs with flexible peripherals are ideal for implementing communications and user interfaces.