Power Electronics

Optimizing Power Designs for Digital Light Projectors

For the PDF version of this article including diagrams and/or equations, click here.

The digital light projector (DLP) display technology developed by Texas Instruments uses an optical semiconductor to manipulate light digitally. An all-digital device, the DLP chip delivers a high-quality picture across a broad range of products, including large-screen digital HDTVs; projectors for business, home and professional venues; and digital cinema. These applications present unique power design challenges because of the power requirements posed by the light source, which may be a high-intensity discharge (HID) lamp or an LED array, as well as standards-based requirements for power factor correction (PFC) and energy efficiency.

To satisfy these power requirements, designers need to understand the basic operation of the DLP chips and the options for providing power in the application.

DLP Operation

The DLP chip is a sophisticated light switch that contains a rectangular array of up to 2 million hinge-mounted microscopic mirrors. Each of these micromirrors measures 16 microns × 16 microns. When a DLP chip is coordinated with a digital video or graphic signal, a light source and a projection lens, its mirrors can reflect an all-digital image onto a screen or other surface.

A DLP chip's micromirrors are mounted on tiny hinges that enable them to tilt either toward the light source in a DLP projection system (when it is turned on) or away from it (when it is turned off) — creating a light or dark pixel on the projection surface. The bit-streamed image code entering the semiconductor directs each mirror to switch on and off up to several thousand times per second. When a mirror is switched on more frequently than off, it reflects a light gray pixel; a mirror that's switched off more frequently reflects a darker gray pixel. In this way, the mirrors in a DLP projection system can reflect pixels in up to 1,024 shades of gray to convert the video or graphic signal entering the DLP chip into a highly detailed gray-scale image.

The white light generated by the lamp in a DLP projection system passes through a red, green and blue color filter as it travels to the surface of the DLP chip. After passing through this filter, the colored light then falls sequentially onto the DLP chip to create an image with up to 16.7 million colors. Some DLP projection systems include a three-chip architecture capable of producing up to 35 trillion colors.

The on and off states of each micromirror are coordinated with these three basic building blocks of color. For example, a mirror responsible for projecting a purple pixel will only reflect red and blue light to the projection surface. Our eyes then blend these rapidly alternating flashes to see the intended hue in a projected image (Fig. 1).

Powering the DLP System

Fig. 2 presents a block diagram of a typical DLP HDTV power system. Total power supplied is on the order of 200 W. Because these products are supplied to the European market, a PFC circuit is often provided to meet their harmonic requirements. The PFC circuit provides a regulated 400 V, which feeds the lamp power supply and low-voltage logic and analog circuits. In addition, there is a standby power supply that powers a small sustaining load during the off condition. Typically, this standby supply will need to be energy efficient, or green. To be Energy Star compliant, this supply must consume less than 0.5 W of input power at no load.

Using LEDs as a light source is another trend that directly impacts the power-supply design of DLP products. In addition to eliminating the need for a ballast, LEDs also provide much longer lamp life, are more light efficient and eliminate the color filter. LEDs offer a whole new range of possibilities for generating superb images. Color segments are no longer tied to the color filter design and speed of rotation, allowing more mixing options, faster on/off switching speeds and intensity control through current-level management. The smaller sizes of LED light engines offer a distinct advantage in portable products.

Fig. 3 shows a block diagram of a power supply for a LED projector. It is very similar to a DLP LED HDTV in that it provides a standby supply, a PFC circuit, a main power supply and a supply for the LEDs. In this block diagram, the LEDs are driven from one of the main power supply outputs. Alternate configurations supply the LED drivers from the 400-V output of the PFC. Although these supplies seem very simple from the block diagrams, each has its own design challenges.

Transition or Continuous-Conduction Mode?

In DLP applications that employ an HID lamp and ballast, a decision must be made whether to use a transition-mode or continuous-conduction-mode (CCM) PFC. Both topologies are nonisolated boost converters that generate a regulated 400-Vdc output from the full-wave rectified ac line input. In addition to generating a dc voltage, a PFC also forces the line current, which is also the current in the PFC boost inductor, to follow the input voltage in shape and phase. This reduces line-frequency harmonics and improves the power factor.

The differences between CCM and transition-mode control are shown in Fig. 4. A PFC that employs CCM uses a fixed-frequency PWM to regulate the average current in the inductor. As a result, the PFC MOSFET must turn on while current is still flowing through the inductor and diode. This can lead to high switching and reverse-recovery losses. Ultrafast diodes, which add cost, are typically used in CCM PFCs to lower the reverse-recovery losses.

By contrast, a transition-mode PFC regulates the peak inductor current and waits until the inductor current returns to 0 A before beginning the next pulse. This substantially reduces the reverse-recovery and turn-on losses, but also leads to much higher peak currents. High peak currents can result in proximity losses in the PFC inductor and a substantially larger EMI filter. In addition, the transition-mode switching frequency is variable, which further complicates the EMI filter design.

Transition-mode controllers tend to be simpler and less expensive than CCM controllers. As illustrated in the table, the typical rule of thumb is to use transition mode for output powers less than 200 W, and to use CCM for output powers greater than 200 W.

Powering the Ballast

When the light in the TV comes from a HID lamp, an electronic ballast is needed to control the lamp. The HID lamp is comprised of two opposing electrodes in a high-pressure gas-filled bulb. The high-pressure gas must be broken down for current to flow in the lamp, and consequently, a high-voltage circuit is used to generate a 30-kV impulse to create an arc within the bulb's gas.

After the gap has been broken down, it has a nearly constant-voltage characteristic of around 40 V. The voltage changes in the short term as the gas in the bulb heats up, thereby increasing the pressure. It also has a long-term variation as the tips of the electrodes erode and the gap length increases. The electronic ballast must regulate the bulb's power to keep the lamp output constant over time.

Several protection features must be built into the HID ballast supply as illustrated in Fig. 5. Once the igniter has been fired, a decision is made as to whether the lamp sustained an arc. If it has not, a digital counter is incremented and a decision is made as to whether to retry ignition. If there is a sustained arc, the power from the ballast supply is limited and the output voltage is then monitored. If the voltage monitor senses an overvoltage condition caused by an aging lamp or open circuit, the supply is disabled. Finally, there is quite a bit of housekeeping required; the warming up and cooling down of the bulb is controlled, and if the supply enters a standby mode, the PFC must be disabled. With all this overhead required, a microcontroller makes the best choice for overall power-supply control and fault monitoring, and it has become feasible for the PWM portion of the supply.

Energy Conservation

Standby power-consumption requirements worldwide vary from 1 W to 15 W, depending on the energy-conservation program and type of television. For example, to receive the EPA's Energy Star certification, a digital television must consume less than 3 W while in standby mode.

One obvious way to reduce standby power is to minimize the power required by the system during standby. Unfortunately, this is typically out of the hands of the power-supply designer, and the power-supply designer is saddled with having to deliver around 300 mW from a limited input power budget. While this may seem easy enough to accomplish, the PFC and 250-W main power supply typically draw more than enough power at no-load operation to push the losses well above the acceptable limits. As a result, it is usually necessary to disable all unused supplies, including the PFC, during standby. Typically, this is accomplished by gating the bias power to the supply controllers.

Fortunately, IC manufacturers have taken notice of the need for efficient light-load controllers and now offer controllers targeted for these applications. An example of a PFC and green-mode flyback converter standby supply is shown in Fig. 6. This circuit uses an energy-efficient UCC28600 to minimize power loss while in standby mode. The UCC28600 enters a burst-mode operation at light loads and provides a signal to disable the bias power to the PFC controller.

The circuit shown in Fig. 6 is adequate for reducing the standby power consumption to below 3 W, but may not be enough to achieve less than 1 W of input power. PFC controllers require resistor dividers to sense the ac line voltage and the PFC output voltage. These resistors can easily dissipate more than 200 mW. In addition, the leakage current of the PFC output capacitor can lead to another 200 mW of unwanted loss. Combined, these losses can push the standby losses above the acceptable limit. In these situations, it may be necessary to use a relay to disconnect the ac power to the PFC and all downstream converters. This relay may be used in conjunction with a dedicated standby supply. In addition, this relay may be a solid-state type as long as it does not require significant bias power while the system is in standby mode.

Lower Power with LEDs

The most basic LED light engine consists of red, blue and green LEDs that are pulsed on and off at duty cycles and frequencies that mimic the rotation of the color filter. Separate on/off signals for each color are sent from a microprocessor to the LED drivers. The intensity of each color is fed back to the microprocessor through an optical sensor. To achieve the proper color balance, the microprocessor sends a signal to the LED drivers to adjust the current in each LED.

An example of an LED driver circuit for a portable DLP projector is shown in Fig. 7. In this circuit, a TPS40071 controller is used to control a synchronous buck power stage that is operated as a current source while the LED is on, and as a voltage source while the LED is off. The LED on/off signal, sent from the microprocessor, turns the LED on by turning on FET Q1 and moving switch S1 to the down position, which provides the current feedback signal. When the LED is turned off, S1 is returned to the up position, which allows the TPS40071 to regulate the driver output voltage. The LED current is controlled by varying the pulse width of a 10-kHz digital pulse train sent from the microprocessor. This PWM signal is filtered and summed into the feedback pin of the TPS40071.

In Fig. 7, the resistor divider of R1 and R2 is designed so that the regulated voltage in the LED off state closely matches the forward drop of the LED during the on state. This keeps the output of the TPS40071 internal error amplifier at nearly the same level during the two states and minimizes the rise time of the LED current when it is switched on. This is significant because a fast current rise time provides more flexibility in the digital control of the projected light.

The waveforms of Fig. 8 show the output voltage and LED current during this transition. For these waveforms, the LED driver is powering two 1-A green LEDs connected in series. The wide bandwidth of the synchronous buck-current source, about 100 kHz, helps to minimize the current rise time.

Benefits Beyond DLP

As with any unique product, DLP technology has brought new problems for the designer to solve. But these problems have inspired component developments, which may benefit other applications as well as satisfy DLP system requirements.

Controllers like the ones described previously have been developed to power HID lamps. Because the controller required green-mode operation, it was necessary to develop variable frequency control that shuts portions of the power system off.

Meanwhile, LEDs have brought improved reliability to DLP projectors, while advancing the development of rapid slew rate power supplies. Finally, the need for transition-mode control in the PFC stage has led to cost and size reductions in this circuitry.

Transition mode offers advantages at lower powers.
CCM Transition mode
Peak current 1.2 × ILINE 2 × ILINE
Diode 15 ns 50 ns
Inductor Conventional Litz wire
FET High PSW and PRR Low PSW and PRR
EMI filter 1 to 2 stages 2 to 3 stages
Typical power range > 200W < 200W
Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.