LED Journal – Online Source for Light Emitting Diode Technology

LED Journal

Click here for reuse options!

How to Dim a WLED Backlight Driver
By Jeff Falin, Factory Applications Engineer • Texas Instruments

White light-emitting diodes (WLEDs) have many advantages over fluorescent lights. These include being a solid-state device and a directional light source. They operate at much lower voltages and are easier to dim over a wider brightness range. They also have more linear brightness changes when dimming.

Many electronics with display now use WLED-based backlights. Selecting the correct method for dimming LEDs is challenging without a proper understanding of how each method is implemented and the benefits and limitations of each method. After a brief review of how to power LEDs, this article summarizes two LED dimming methods with the pros and cons of each. Using this information, one can easily choose the appropriate dimming method, and therefore LED driver IC, for the application.

Configuring a Voltage-Regulating Converter to Power a WLED
A WLED’s brightness varies linearly with the current passing through it. For the best WLED current accuracy and uniform WLED brightness per string, a LED driver should regulate current through, not voltage across a LED. Figure 1 shows how any adjustable output, voltage-regulating DC/DC converter can be easily re-configured as a constant current source to drive multiple WLEDs in series, as long as its output is greater than the sum of the LEDs forward voltage (VLED) drops.

By regulating VSENSE the voltage across the current sense resistor (RSENSE), and not the output voltage (VO), the driver is essentially a constant current source. This leaves VO free to self-adjust for changes in SVLED. WLEDs have voltage drops ranging from 3.0 V to 4.0 V. The drop varies directly with the LED current and inversely with temperature. More recent, low-power drivers replace the external sense resistor with one or more current sinks, essentially a FET, as shown in Figure 2.

The driver performs two functions. It adjusts the sink FET’s drive voltage to achieve the appropriatecurrent through the sink FET relative to a bias current. It alsoadjusts the output power of the integrated DC/DC converter, typically a boost converter, so that the FET has the minimum drain-to-source voltage necessary for that current. An example of such a driver with an integrated boost converter and eight integrated current sinks is the TPS61195.

Pulse Width Modulation Dimming
To enhance viewing as well as optimize LED driver efficiency across different ambient light levels, newer LED backlit electronics boast a wide dimming range. Two methods are used to dim LEDs: Pulse width modulation (PWM) dimming, and analog dimming. Figure 3 illustrates how the LED’s current, and therefore its brightness, varies when using analog and PWM dimming.

To implement PWM dimming, a digital signal processor (DSP) or microcontroller sends a PWM signal at various duty cycles (D) to enable and disable either the WLED driver’s converter for a driver as shown in Figure 1, or the current sink for a driver as shown in Figure 2. Therefore, the average current through the WLED string is the duty cycle times the maximum current, i.e., ILEDavg = D X ILEDmax.

Because the maximum current through the LEDs is the same, PWM dimming results is a very linear change in brightness. Also, since an LED’s emitted light spectrum varies with its voltage drop, and the voltage drop varies with ILED, which remains at the maximum value, the LED backlight’s chromaticity (its colorfulness and hue or how “white” it really is), is excellent when using PWM dimming.

The key disadvantage with PWM dimming is audible noise. If the PWM signal is used to enable and disable the converter, the driver’s maximum dimming ratio is limited by the time it takes the converter to startup, recharge the output capacitor, and settle to their respective maximum currents. Even though WLED drivers may have converters operating with 1 MHz + switching frequencies, the converters have control loop response times and/or startup times in the hundreds of microseconds to a few milliseconds range. Therefore, to allow time for the driver to settle at its maximum current, the PWM dimming frequency can be only a few hundred Hertz.

Ceramic output capacitor’s piezoelectric properties cause the capacitor to charge and discharge at PWM signal frequencies in the audible range (20 Hz to 20 kHz). It vibrates and the human ear hears the capacitor’s and printed circuit board’s movement as ringing or buzzing. This vibration is directly proportional to the amplitude of the voltage change and ceramic capacitor package size. Reducing capacitor package size reduces the ringing.

Drivers as shown in Figure 2 implement PWM dimming by turning the current sink on and off. Alternatively, drivers like the TPS61093 have a FET in series with the LEDs. The FET is turned on and off quickly, removing the LEDs from the driver’s output. In both cases, a second voltage feedback loop provides protection from over-voltage and maintains the voltage on the output capacitor when the LEDs are off. Because the output capacitor’s voltage change is minimized, its vibration and ringing is reduced.

Analog Dimming
The term analog dimming describes when the DC current through the LEDs themselves changes relative to duty cycle, D. To implement analog dimming for a driver as shown in Figure 1, the DSP or microcontroller must provide an external DC voltage (or low-pass filtered PWM signal) that is higher than the converter’s regulation voltage.

Some drivers with a current sink take the input PWM signal, filter it and apply a level-shifted version to drive the current sink. Other driver’s, like the TPS6116x family, use the input PWM signal to apply the duty cycle, D, to the bandgap reference voltage so VREF = D * VREF(MAX). Because the ILED DC level current changes slowly, the output capacitor voltage does not have ripple. Thus, the capacitor does not vibrate as with PWM dimming.

Another benefit of analog dimming versus PWM dimming is higher power efficiency and electrical-to-optical efficiency. Specifically, the boost converter output voltage = SVLEDs lowers as ILED lowers. Hence, the converter’s output power is slightly lower when using analog versus PWM dimming.

Since the boost converter needs to provide a lower output voltage, its input power requirement drops and its efficiency increases. Figure 4 compares the driver’s efficiency when using mixed-mode and PWM dimming at the same input voltage, and with the same LED. In mixed-mode dimming, the driver performs analog dimming down to D=6.25 percent, then converts to PWM dimming for improved brightness linearity.

Furthermore, the driver has higher optical-to-electrical efficiency, meaning more lumens for the same power consumed. However, analog dimming has some current accuracy problems when deep dimming because either the feedback regulation voltage or the current sink voltage becomes too small to accurately control. This is due to the offset voltage of the error amplifier. The brightness linearity and chromaticity are not as good as can be achieved with PWM dimming, especially when deep dimming. Figure 5 compares the brightness of a string of dimming LEDs when using analog and PWM dimming.

Realistically, the human eye can rarely discern the difference in chromaticity or linearity, unless two identical displays are compared side by side.

Conclusion
If the application’s lighting needs the best linearity and chromaticity, then a driver capable of true PWM dimming may be the best option. If the application is noise sensitive or needs highest efficiency, then a driver capable of analog dimming may be required. Finding a driver that performs PWM dimming and a second feedback loop to lower output voltage ripple, ringing may be unavoidable. Drivers capable of switching between dimming methods to achieve the best features of each, such as with the TPS61195, are just now becoming available. Regardless, once you have a better understanding of the LED dimming options along with their pros and cons, selecting your LED driver should be greatly simplified.

Jeff Falin is a Factory Applications engineer with the High-Performance Analog Portable Power Applications group at Texas Instruments. Jeff provides customer applications support for both linear regulators and high-efficiency switching power ICs used primarily in consumer electronics ranging from cell phones to LCD TVs. Jeff received his MSEE from the University of Tennessee with a concentration in IC design. He can be reached at ti_jfalin@list.ti.com.