Automatic progressive dimming using multiple LED controllers

Utilizing two or more independent LEDs, today’s drivers can control trendy decorative lights that can be used in portable systems. Not only is the ILED peak current fully programmable, each LED can be dimmed to any value between 0 and 100% brightness. In addition, embedded progressive dimming functions that work in both the upward and downward directions provide a special lighting sequence that the end customer requires. This article describes the features of this driver and focuses on progressive dimming based on a typical application. In addition, the associated software is discussed as a typical example.

Basic simulation operations

Typically, an LED driver provides a constant current to bias the LED under the proper conditions. If we consider a portable system, the power source is a battery with an output voltage range of 2.8 to 4.2V (assuming a standard Li-Ion battery). Since the forward voltage of today's low-power LEDs can vary between 2.8 and 3.5V depending on the bias current and room temperature, an interface is needed to ensure that the LED is properly biased during normal operation. This is the purpose of the driver IC, and the first block to consider is the voltage range of the current control system.

Figure 1 Basic LED wiring

Figure 2 Comparison of advantages and disadvantages of series-parallel connection

At this point, we can consider connecting LEDs in series or in parallel: both connection methods have their own advantages and disadvantages, see Figure 2.

The key point is the ability to adjust the brightness of each LED independently and dynamically in color applications. Although it is possible to use a boost structure and use switches to connect between each LED to control them, the series arrangement is not the preferred solution, and the parallel structure is the easiest to implement.

The charge pump is the most suitable DC-DC converter to generate low voltage and minimize EMI problems. On the other hand, the use of multiple operating modes (1X, 1.5X, 2X) purely improves efficiency, allowing the system to save as much energy as possible when running in portable devices.

Besides the DC-DC converter, the second critical parameter is the current matching between LEDs belonging to a common array: the RGB structure cannot accommodate the bias current differences between LEDs, because these differences will translate into color performance in video and image displays. This problem is solved by using a set of accurate current mirrors as shown in Figure 3.

To achieve an accurate and stable forward bias condition in the LED, a reference current is generated by an external resistor related to the constant voltage provided by the bandgap reference. Transistor Q1 associated with operational amplifier U1 generates a constant output voltage at the Vref pin. The external resistor connected between Vref and ground generates a constant current flowing through Q1 and Q2. This current is now mirrored and amplified by the series of transistors Q3 to Q7, each of which is connected to switches S1 to S5 and summed by transistor Q8. Finally, transistor Q9 copies the reference current to LED1. This structure is replicated for each LED, and the layout of the chip is carefully analyzed to optimize the matching between each LED.

Figure 3 Basic current mirror structure

Figure 4 Typical independent PWM control

In this way, each LED shares the same I-LED peak current and no additional electronic circuit is needed to control the brightness of each LED individually. This functionality is achieved by using an independent PWM modulation for each LED (see Figure 4): switches S6~S8 controlled by digital signals PWM1 to PWM3 enable/disable the associated current mirror, thus generating brightness control of the associated LED. The net advantage is that the LED peak current is constant, ensuring that the color performance is not weakened by the brightness control and the operating point of the LED remains at the reference color defined by the color map, see Figure 5.

The waveforms in Figure 5 are derived from an industrial application and show the characteristics of the three PWMs embedded in the selected device. The three LEDs are controlled by a common low-frequency clock with a duty cycle setting to suit the given application. It is obvious that each PWM can be weakened or strengthened independently, ranging from 0 to 100% duty cycle, while the ILED peak current is constant.

Figure 5 Typical industrial PWM operation

A more complex circuit design can be used to obtain fully independent control of the LEDs: both the I-LED peak current and the PWM can be digitally programmed, producing an almost unlimited color range and brightness as the I-LED peak current moves through the color map. The basic model description is shown in Figure 6.

Figure 6 Comparison of LED and CCFL color gamut with NTSC standard

Digital Control

The standard I2C port is used to handle I-LED and PWM, and the built-in functions of the controller are set by software. To better illustrate the gradual dimming, we will use the NCP5623 controller as a reference to describe the operation of this function.

Before PWM can occur, the ILED peak current is set by sending the appropriate code to the chip as defined in the NCP5623 datasheet. To create a smooth increase in brightness, the software sends the driver the total number of steps available: in this case, we have 31. A simple loop could be implemented in the microcontroller (MCU) to handle this, but the brightness ramp-up process could be disrupted due to priority interrupt issues associated with real-time systems. The NCP5623 has a built-in sequence that avoids the need for real-time MCU operations: both increasing and decreasing brightness, gradual dimming can be achieved with a very limited number of software steps and without the impact of high-priority interrupt events.

Basically, the two built-in registers will be pre-adjusted as follows.

Goals and directions of progressive dimming:

-Brightness Enhancement=%101xxxxx

→ The last bits [B4:B0] contain the final ILED target for the enhancement

-Brightness reduction=%110xxxxx

→ The last bit [B5:B0] contains the final ILED target attenuation.

Timing and start conditions:

GRAD=%111xxxxx

→The last bit [B5:B0] contains the timing of each stage

The ILED current will increase smoothly from 0 to 5.5mA, with the total sequence timing equal to the contents of the GRAD register bits [B5:B0] multiplied by the number of levels defined in the UPWARD register. In this example:

T=GRAD[B5:B0]*UPWARD[B5:B0]

T=64*26=1664ms

Figure 7 Typical NCP5623 automatic upward gradual dimming process (8ms per level)

The waveforms shown in Figure 7 demonstrate upward gradual dimming; downward dimming operation is achieved by appropriately programming the DWNWRD register.

As we can observe, the ILED current increases in a quasi-exponential curve, which is good enough to compensate for the sensitivity of the human eye.

The reverse direction is easily accomplished by using the appropriate code in the upper three bits of the data register, with the rest of the sequence being the same.

Built-in registers make it possible to dynamically control the gradual dimming, allowing the simulation of different visual effects. For example, we can repeat a sequence created by digital modulation of the up or down period, or we can combine a set of gradual dimming with a sudden change to create a sawtooth-like waveform on the opposite side of the waveform.

Finally, it is possible to combine progressive dimming with the PWM embedded in the chip to create quite complex lighting sequences by modulating the ILED peak current via the IREF pin: a decorative light system can be constructed with a minimum number of passive components around the main controller.