2012年5月27日 星期日

How to design a power efficient LED driver

In the latest generation of smartphones, which feature large displays suitable for showing web pages and video, the LED backlight system is one of the biggest power loads in the device.

With larger displays, the number of LEDs in the backlighting system grows, putting even more pressure on the power budget.

There are three building blocks in a backlighting system: a battery to store energy, the LED driver, and the LEDs. To increase run-time, extra power, or power savings, need to be found in one or more of these three blocks.

The LEDs form part of the display assembly, placed on a flexible PCB, and emitting light from the edge into a diffuser behind the LCD panel.

The number of LEDs depends on the display size, available space and required resolution. The forward voltage of the LED is normally in the range 3.0-3.4V. Mobile phones can contain as many as 12 LEDs.

The arrangement of the LEDs can vary. For instance, all the LEDs could be connected together at their anode, with an output pin at each cathode. This connection scheme can be driven by a capacitive backlight driver, and is widely used in small displays.

A typcial approach in large-screen smartphones is to connect all the LEDs in series as a single string. The advantage of this topology is that it needs only two connection pins to the display.

An inductive DC-DC converter is more suitable for driving single or multiple strings of LEDs than a capacitive driver is. And the more LEDs are connected in series, the lower the efficiency of the inductive DC-DC converter falls.

The problem is the result of both switching losses and core losses in the inductor; these losses rise higher the higher the voltage at the converter.

So why does output voltage increase as the number of LEDs in a string rises? The reason is that the output voltage required from the DC-DC converter is calculated by multiplying the forward voltage of the LEDs by the number of LEDs, plus the compliance voltage of the current sink.

For example, in a single string of 12 LEDs with a forward voltage of 3.2V, the DC-DC converter needs to boost the output voltage to 12 x 3.2 + 0.5V = 38.9V. (0.5V is the compliance voltage of the current sink.) By comparison, in a multi-string configuration of, for example, 4S x 3, the maximum voltage would be 4 x 3.2 + 0.5 = 13.3V.

In the multi-string configuration, the lower voltage of the DC-DC converter, 13.3V, results in lower switching losses and core losses than in a converter operating at 38.9V. This results in better efficiency, and also allows the use of smaller external components.

Drawing on the phenomenon shown in Figure 2, the improved efficiency of the multi-string topology can thus be seen to be attributable to the reduced output voltage of the converter.

System designers considering whether to adopt a multi-string topology will regret losing the simplicity of the two-pin connection to the display module which the single string topology affords them.

沒有留言:

張貼留言