Designers of LED lighting have quickly become familiar with applicable safety standards, such as IEC 62560 for general-purpose and domestic LED lamps, IEC 62031 for LED arrays and modules, and IEC 61347 for drivers and power supplies. Exceptional hazards include high-energy surges on input power lines, which can be caused by events such as nearby lightning strikes. IEC 61000-4-5 describes surge testing using a standard 8 x 20 μs waveform, and specifies levels as high as 10 kV/5 kA for outdoor lighting applications in Europe.
Devices such as inline fuses, Metal-Oxide Varistors (MOV) and Transient-Voltage Suppression (TVS) diodes connected in parallel can be used throughout power-supply and driver circuitry. Suppliers such as Littelfuse provide extensive guidance on how to select and position devices to absorb and divert the energy from potentially damaging transients.
Figure 1 provides an overview of surge-protection devices used in a generic LED-lighting solution. MOVs placed across line to neutral, neutral to ground, and line to ground, as shown, provide high surge-withstand capabilities, such as the Littelfuse V300SM7. Excessive voltage across the MOV causes the device to create a conductive path, thereby diverting surge energy. TVS diodes may be devices such as the Littelfuse P6KE300, and protect circuit components by dissipating transient energy. The chosen device must be capable of withstanding the maximum impulse current due to the applied transient voltage.
Figure 1: Littelfuse guidance for designing surge-protection devices into an LED-lighting application.
Protecting against line-voltage fluctuations
The devices illustrated are effective in protecting circuitry against short-term high-energy pulses. However, fluctuations with slower time constants can also pose a threat. It is well known that utilities are under pressure to maintain grid stability as end-user demands increase, infrastructure ages, and traditional generating capacity based on fossil fuels transitions to a greener model more heavily reliant on distributed generation from renewable sources. Under these conditions, under and overvoltage fluctuations can arise that may reduce the reliability and lifetime of components in some types of circuits.
As one example, LED replacements for familiar lighting products such as MR16 or GU10 bulbs face both tight limits on cost and size. To help overcome these pressures, the Texas Instruments TPS92210 LED-driver controller features an internal MOSFET that is designed to be connected in a cascode configuration with an external high-voltage MOSFET. This simplifies startup, allows use without an external current-sense resistor, and reduces primary-side switching losses. By supporting discontinuous conduction mode (DCM) operation, it also minimizes the output rectifier diode reverse-recovery losses. As a result, the TPS92210 helps increase efficiency and reliability while also reducing system cost compared to conventional flyback architecture. Figure 2 shows the schematic for a typical application. Note that the external MOSFET connected to the DRN pin (pin 6) connects to the drain of the TPS92210 internal driver MOSFET to form the cascode circuit.
Figure 2: LED driver circuit designed to improve performance over conventional flyback converter.
This driver circuit is designed to deliver constant power to the LED string. If instability in the grid causes the line voltage to reduce, the input current to the driver will increase to maintain constant output power. This increased current can place excessive stress on the driver components. Similarly, a significant increase in the line voltage, combined with ringing due to the transformer primary-side winding inductance, may exceed the ratings of important components such as MOSFETs and capacitors. While standard components such as the MOVs and TVS diodes mentioned previously provide effective protection against short high-energy surges, additional protection may be needed to prevent damage by underlying line instability.
When using a controller like the TPS92210, external circuitry can be designed to take advantage of the IC’s transformer zero-energy detection (TZE) capability to temporarily disable the driver when the AC line input rises or falls below its normal range.
Over/undervoltage protection circuit operation
When the driver is operating in DCM, each successive switching cycle is initiated only when the transformer has been completely reset or when its energy is zero. The resistor-divider connected to the TZE pin enables the transformer zero-energy point to be detected by monitoring the current sourced out of the TZE pin when the primary bias winding goes negative with respect to ground.
Figure 3 shows a protection circuit that stops the driver operating by preventing initiation of the next switching cycle in the event of input under/overvoltage. This is achieved by forcing a DC voltage on the TZE pin to prevent zero-crossing detection. When the input voltage is within safe operating range, the circuit does not output DC voltage to the TZE pin thereby allowing normal zero-crossing detection that enables the controller to coordinate valley switching for optimum efficiency.
Figure 3: Schematic for input undervoltage and overvoltage protection.
The circuit operates by receiving the rectified, unsmoothed line voltage from the bridge rectifier output. This voltage is clamped at 12 V with Zener D2, and further reduced by resistor-dividers. Resistors R3 and R4 are associated with undervoltage protection, while R5 and R6 handle overvoltage protection. The resistor values R3, R4, R5 and R6 are determined to set trip thresholds of 1 V and 2.5 V, respectively.
The 12 V bias is also used to supply the precision quad single-supply micro-power op-amp, U1 (TLC27L4). A micro-power op-amp is chosen for U1 to allow operation directly from the Zener diode without experiencing erratic on/off cycles at low input voltages, which can occur if a device requiring higher supply current is used. U1-A acts as a peak detector generating a DC voltage proportional to Vin(rms) on the capacitor C4. U1-B buffers this DC voltage and U1-C outputs an error signal if the peak detector voltage is below the undervoltage reference VR1. Similarly, U1-D compares the peak detector output with overvoltage reference VR2 to generate an error signal when the RMS input voltage exceeds the overvoltage trigger threshold. The outputs of U1-C and U1-D are clamped at 3.3 V with Zener D5, and then buffered with the transistor Q1 before being fed to the TZE pin. R10 and R12 introduce about 5 V of hysteresis to prevent false triggering at the boundary limits.
Since the TZE input of TPS92210 is continuously being scanned for valley transitions, switching cycles are prevented when the protection circuit forces DC voltage on the pin. Switching is able to resume when the input voltage returns to within normal operating range. The table shows the behavior of the device and driver output status in response to normal and surge conditions at the input.
|Input Range of Operation||AC Input (V)||U1-C Output||U1-D Output||TZE Pin Input||LED Driver Status|
|Undervoltage||< 85||High||Low||DC voltage forced||Off|
|Normal operating range||85-260||Low||Low||Normal operation||On|
|Overvoltage||> 260||Low||High||DC voltage forced||Off|
Table 1: Summary of TPS92210 driver states.
Conventional surge-suppression devices such as fuses, MOVs and TVS diodes are essential to ensure that LED lighting solutions comply with international safety standards. Additional circuitry can provide smart protection against deterioration of AC power-line quality, by preventing potentially damaging currents or overvoltage from reaching driver components or LEDs.