[For a copy of this article in PDF format, which displays figures and equations,
Requires Adobe Acrobat Reader,
.]
The shift from 2G to 3G has turned cell phones into multimedia devices with built-in cameras, Web browsers, games and personal digital assistants (PDAs). This, in turn, has introduced color displays and lighting features, such as cover lights, color indicator lights, camera flash, auxiliary lights and ambient light sensing around phones. As a result, there is a need for more sophisticated cell phone lighting management solutions.
Thus, a huge market for white LEDs, display modules and RGB lighting has been created. The display has become a MCU-controlled stand-alone module, with all of the related electronics integrated into the design. The methods for driving and controlling white LEDs also has evolved over the last few years.
White LEDs for Backlighting
Today, almost all cell phones have one or two color displays. Though requirements for backlighting vary, depending on the number of white LEDs used, most phones now have a main display with up to four white LEDs. In addition, clamp cell phones have a secondary display (with one or two backlight white LEDs) for displaying the time, caller ID, remaining battery life or screen savers. Typically, the maximum current specifications for backlighting are around 15mA/LED to 20mA/LED. Taken as a whole, this amounts to 100mA to 120mA, representing a significant percentage of the phone's total power consumption.
Since the typical forward voltage (VF =3.2V to 4V) of white LEDs is higher than the typical minimum battery voltage (2.8V to 3V), white LEDs can't be driven directly from an Li-Ion battery. As a result, a dc-dc boost converter is required to step up the battery voltage. This boost converter can be inductive (magnetic boost) or capacitive (charge pump), depending on the required maximum load current, efficiency levels, external component limitations and noise specifications.
Currently, white LED backlight drivers are linked using serial or parallel connections. The serial connection requires a much higher output voltage, requiring a magnetic boost converter. Parallel connections of LEDs, however, can be driven from much lower voltages (4.5V˜5V) and these can be generated with either a magnetic boost or a charge pump. In both systems, the adjustment of LED current is possible using an external PWM signal where the duty cycle determines the output current -I(out) = Duty * I(max).
RGB Lighting
While LEDs are commonly used to backlight phone displays, they're also useful in several other applications. New multipurpose RGB LEDs (components with red, green and blue LEDs in one package), traditionally used as single-indication lights, are now smaller, brighter and cheaper, allowing for more widespread adoption by handset manufacturers. These new RGB LEDs have enabled manufacturers to design handsets with covers that change color or that feature brighter color indications and warning lights, blinking and changing visual game effects, and ring tone or music-controlled color effects. Some of the new RGB LEDs have even been used to generate bright white light by blending the red, green and blue lights. This white light feature has been used as a source for a camera-enabled phone's flash or auxiliary light.
How the RGB LEDs are controlled affects the visual effects that can be obtained. RGB LED's internal LEDs have different forward voltages. To drive blue (Vf˜3.5V) and green (Vf˜3V) LEDs, a dc-dc boost is again essential. In some cases, red LEDs (Vf˜1.7V) can be driven directly from the battery. RGB-LED control is normally done with PWM-controlled switches, together with LED current-limiting (ballast) resistors. The intensity of each color is determined by the duty cycle of the PWM control signal to each sub-LED. One way to control the intensity is to use an external FET for each LED and the baseband processor to generate PWM signals. However, the baseband processor must simultaneously control several processes, thus limiting the ability of the processor to effectively drive PWM signals. Consequently, programming a stand-alone RGB controller to a certain color sequence is a better solution. The stand-alone controller integrates the switching FETs and, due to its optimized logic and clock, is able to generate programmed PWM sequences for LEDs. Fig. 1 shows the RGB color sequencing examples generated with LP3936. Different colors and intensities are obtained by controlling separately the brightness of each basic color — red, green and blue.
Optimizing Lighting
Adding lighting features to the phones is more complex than it sounds. By defining lighting requirements at the outset, you can avoid any setbacks along the way. The following are some questions to consider:
-
How many displays do I need to drive?
-
How many white LEDs have to be on at the same time?
-
How much current do I need to drive per LED → total IWLED?
-
What kind of control and interface do I need to adjust the backlight intensity?
-
Is white LED current matching important?
-
What driver do I need to drive the LEDs?
-
How many RGB LEDs do I need to drive?
-
How much current per LED I need to drive → total IRGB?
-
Are RGB LEDs and backlight white LEDs on at the same time?
-
What kind of control do I need for RGB and how am I going to do it?
-
If RGB will be used for flash, how am I going to drive the high current?
-
What kind of display do I have? Can I switch the display backlight off in sunlight condition?
-
Do I need to take care of ambient light sensing?
-
What is the space requirement for lighting ICs and external components?
-
Do I need extra current for some other LEDs, such as keypad LEDs?
-
How many driver ICs should I have?
-
How much will this cost?
The answers to these questions can help you weigh the cost effects that result from various component requirements, the complexities of lighting controls, efficiency considerations and the control demands of the RGB LED.
Saving Power
One area that demands attention is power savings. When designing new phones, engineers try to save power wherever possible. This also holds true when it comes to handset lighting. With respect to lighting, the type of dc-dc converter topology used plays an important role in affecting overall efficiency. Both a magnetic boost converter and a switched capacitor charge pump can achieve high efficiency when designed correctly. The use of magnetic boosters has become more widespread on portable applications because inductors are now available in desirable sizes (for example, 10-uH shielded coil — 3mm × 3mm × 1.2mm) and at attractive prices from coil manufacturers.
Power savings also are achieved by controlling backlight intensity with adjustable white-LED current drivers. New semitransparent color displays require different backlight brightness based on ambient light conditions. In practice, this means that in dark or sunny environments, backlight current can be reduced by 50% to 100%. In dark environments, the human eye needs less backlight; in sunny environments, the reflective material at the bottom of the backlight module reflects enough light for proper backlighting. Ambient light monitoring requires a small correctly placed photodiode and an analog-to-digital converter.
Complete Solution for Handheld Lighting
For modern handheld lighting, let's examine an example specification of the phone lights. The table specifies the required phone light functions, including display backlights, RGB LED, camera flash, key pad light, dc-dc converter (driver) and ambient light sensing. The values shown on the table are typical numbers for today's phones.
Fig. 2 shows the discrete solution. It contains three separate LED drivers: one for main backlight (magnetic boost, series connection to minimize mismatch of LED current), another for sub-backlight and keypads (switch cap charge pump with 70mA max load), and a third for flash and RGB LEDs (magnetic boost for high current loads). The problem of having multiple LED drivers is the high number of ICs, external components and large PCB footprint area. A large part of the cost comes from the coils and Shottky diode. Also, note that each function requires its own enable pin or PWM-type control signal from baseband processor (BBP). Also, the RGB PWM controls, keypad enable and flash require external FETs.
Fig. 3 illustrates the lighting management unit (LMU)-based solution. The required drive and control functions are integrated into one LMU chip. The magnetic boost with high-output current capability has been integrated with display backlight drivers (programmable constant current sink), a stand-alone RGB LED driver (integrated switches, individual control of brightness, color and blinking cycle), and an ambient light sensor interface (8-bit Adc). The programmable output voltage at the boost output also allows driving LEDs with optimized efficiency. Plus, the control interface has been minimized to 2-wire I2C or 4-wire SPI/MicroWire. Because the boost is capable of driving high current loads, it also can drive the flash and keypad LEDs. One practical way to control the flash and keypad LED is to use an external FET. Additionally, the baseband processor's control for these lights is a straightforward on/off type.
Conclusion
Compared to the discrete solution, the LMU-based lighting approach offers many advantages. First, the overall efficiency is improved because the drivers for white and RGB LEDs are integrated on-chip. Second, when compared to typical switch capacitor charge pump converters, the magnetic converter is 15% to 20% better across the input voltage range. With magnetic boost, the user can switch off the boost and drive white LEDs directly from battery, assuming that battery voltage is high enough (bypass). Third, the standby mode total power consumption is reduced because only one IC is on standby, compared to three in the discrete approach. Although in most cases the difference is only a few microamperes, in the standby mode, these are critical numbers, as it affects the phone standby time. Fourth, the PCB area can be improved by reducing external component count. The LMU solution requires only slightly larger area than one conventional magnetic boost driver. The discrete solution, having large amounts of components and wiring, isn't as space efficient as the LMU approach. Specifically, this is important when using integrated display modules, where all electronics must be inside the module on a small flexprint. Finally, the LMU-type approach allows lower system price, easy control interface, better system reliability, faster software development, direct ambient light control and flexible use for most mobile phone manufacturers.
For more information on this article, CIRCLE 341 on Reader Service Card