LEDs may be garnering the most headlines these days. But high-intensity discharge (HID) lighting is as popular as ever, too, thanks to its high brightness, excellent color temperature, high lumens/watt, and long lifetime.
As the older copper/iron magnetic ballasts used to drive HID lamps become obsolete due to poor efficiency, newer and more efficient electronic ballasts continue to emerge on the market. However, complex lamp requirements make electronic HID ballast design very challenging. One solution, though, offers a way to overcome these potential thorny issues.
Today’s lighting technologies include incandescent, fluorescent, halogen, HID, and LED. Each of these light sources is unique in how it produces light, as well as in terms efficacy, lifetime, and types of applications (see the table).
Incandescent lamps contain a tungsten filament resistor that’s connected directly across the ac line. As current flows through the filament, the filament heats up to 2200°C, causing the metal atoms in the filament to release light. Less than 10% of the total energy consumed by the bulb actually produces light, while the rest is wasted as heat.
Fluorescent lamps consist of a glass tube filled with argon gas and a small amount of mercury and filaments located at each end. As electrons flow across the tube from one filament to the other, they collide with mercury atoms. The excited mercury atoms give off ultraviolet (UV) light, which is then converted into visible light as it passes through the phosphor coating on the inside of the tube.
Halogen lamps use a tungsten filament encased inside a small quartz envelope. Similar to an incandescent lamp, the electrical current heats the tungsten filament above 2500°C, causing the filament to get “white hot” and release light. The halogen gas inside the envelope combines with tungsten atoms as they evaporate and re-deposits them back on the filament.
This recycling process results in a much longer lasting filament versus incandescent. In addition, because the filament is running hotter, more light per unit energy is achieved, making halogen lamps ideal for “spot” lighting applications.
HID lamps produce light using a technique similar to that used in fluorescent lamps. Unlike fluorescent lamps, though, HID lamps operate at a high temperature and high pressure, the arc length is very short, and visible light is produced directly without the need for a phosphor.
LEDs work exactly the same as a standard p-n junction diode, except the semiconductor material used has a higher band-gap energy level. As current flows through the LED, electrons jump across the wide band-gap junction between the N-type and the P-type materials to recombine with holes.
Continue on next page
The energy lost by each electron during the jump is emitted as a photon of light. The photons produced in the junction that successfully pass through the layers of the device (and the package) appear as the total light emitted from the LED chip. The actual performance of LEDs depends heavily on their operating temperature.
DRIVING HID LAMPS
HID lamps require a high voltage for ignition (3 to 4 kV typical, or greater than 20 kV if the lamp is hot), current limitation during warmup, and constant power control during running. It’s important to have a tight regulation of lamp power to minimize lamp-to-lamp color and brightness variations.
Also, HID lamps are driven with a low-frequency ac voltage (less than 200 Hz typical) to avoid mercury migration and prevent damage of the lamp due to acoustic resonance. A typical metal halide 250-W HID lamp requires a 250-W nominal wattage, 100-VRMS nominal voltage, 2.5-ARMS nominal current, 2.0-minute warmup time, and 4000-VP (volts peak) ignition voltage.
HID lamps exhibit a characteristic startup profile (Fig. 1). Before ignition, an HID lamp is open circuit. After the lamp ignites, the lamp voltage drops quickly from the open-circuit voltage to a very low value (20 V typical) due to the low resistance of the lamp. This causes the lamp current to increase to a very high value and should therefore be limited to a safe maximum level.
As the lamp warms up, the current decreases as the voltage and power increase. Eventually the lamp voltage reaches its nominal value (100 V typical) and the power is regulated to the correct level. Satisfying the lamp requirements and different operating modes requires an electronic ballast circuit topology that efficiently converts the ac mains voltage to the desired ac lamp voltage, ignites the lamp, and regulates lamp power.
HID BALLAST CIRCUIT TOPOLOGY
The ballast circuitry for an HID lamp (Fig. 2) is complex. The boost power-factor correction (PFC) stage runs in critical-conduction mode. During this mode, the boost stage operates with a constant on-time and variable off-time, resulting in a free-running frequency across each rectified half-wave of the ac line cycle. Frequency range is typically from 200 kHz near the valley of the half-wave to 50 kHz at the peak.
The on-time regulates the dc bus to a constant level, and the off-time is the time it takes for the inductor current to reach zero each switching cycle. The electromagnetic interference (EMI) filter filters the triangular-shaped inductor current to produce a sinusoidal input current at the ac mains input for high power factor and low harmonic distortion.
The ballast’s main control circuit, the buck control circuit, controls the lamp current. The buck stage is necessary to step down the constant dc bus voltage from the boost stage to the lower lamp voltage across the full-bridge stage. This particular buck circuit can run in continuous or critical-conduction operating modes, depending on the condition of the load.
The lamp voltage and current are measured and multiplied together to produce a lamp power measurement, which is fed back to control the buck on-time. During the lamp warmup period (after ignition) when the lamp voltage is very low and the lamp current is very high, the lamp current feedback will determine the buck on-time to limit the maximum lamp current.
Continue on next page
During lamp steady-state running conditions, the power feedback will then determine the buck on-time to control the lamp power. The continuous-conduction mode allows the buck circuit to supply more current to the lamp during warmup without saturating the buck inductor.
The full-bridge stage is necessary to produce an ac lamp current and voltage while running. The full bridge typically operates at 200 Hz with a 50% duty cycle. It also contains a pulse transformer circuit for producing 4-kV pulses across the lamp necessary for ignition.
The HID control IC implements a state machine to ignite and run the lamp, as well as shutdown when ballast or line fault conditions occur (Fig. 3). The IC initially starts in undervoltage lockout (UVLO) mode when the supply voltage to the IC is below the turn-on threshold. When VCC increases high enough, the IC exits UVLO mode and enters ignition mode, and the on/off ignition timer is activated to deliver high-voltage pulses to the lamp for ignition.
If the lamp ignites successfully, the IC transitions into run mode and the lamp is regulated to a constant power level. If fault conditions occur, such as open/short circuit, the lamp fails to ignite or warm up, or there’s lamp end-of-life (EOL) or arc instability, then the IC will enter fault mode and shut down safely before any damage occurs to the ballast.
An appropriate buck and full-bridge control circuit can be built around the IRS2573D HID control IC (Fig. 4). The IC includes an integrated high-side driver for the buck gate drive (BUCK pin) and high-side buck, cycle-by-cycle, overcurrent protection (CS pin).
The lamp power-control loop (PCOMP pin) or lamp current-limitation loop (ICOMP pin) controls the on-time of the buck switch. The off-time of the buck switch is controlled by the inductor current zero-crossing detection input (ZX pin) during critical-conduction mode or by the off-time timing input (TOFF pin) for continuous-conduction mode.
The IC also includes a fully integrated, 600-V, high-side and low-side full-bridge driver. An external timing pin (CT pin) controls the operating frequency of the full bridge.
The IC provides lamp power control by sensing the lamp voltage and current (VSENSE and ISENSE pins) and then multiplying them together internally to generate the lamp power measurement. Ignition control is performed using an ignition timing output (IGN pin) that drives an external ignition MOSFET (MIGN) on and off to enable the lamp’s ignition circuit (DIGN, CIGN, TIGN). The ignition timer is programmed externally (TIGN pin) to set the ignition circuit on-times and off-times.
Finally, the IC includes a programmable fault timer (TCLK pin) for programming the allowable fault duration times before shutting the IC off safely. Such fault conditions include failure of the lamp to ignite, failure of the lamp to warm up, lamp EOL, and open/short circuit of the output.
Some empirical results from the circuit illustrate operational behavior (Fig. 5). The buck is working in critical-conduction mode during running conditions, and the constant-power feedback loop controls the on time.