More and more frequently, high power LED's are leaving their niche in small displays to be used in mainstream lighting applications. Some of these, including street lighting and similar applications, require power levels of 100W or greater. LEDs are attractive in such applications due to long life and efficiency in generating light. The ballast designed to drive such a load must have low harmonic line current, high energy efficiency and small dimensions. This article will explore an example ballast circuit with an output power of up to 200W.
This ballast consists primarily of three distinct stages:
- A power factor controller with preceding EMI filter and rectification.
- A DC-DC converter based on LLC topology
- Three switch-mode current sources.
This design, as described below, is able to drive about 105 power LEDs in total with an overall efficiency of 90%. Even more impressively, the efficiency for the PFC and DC-DC stages is close to 95%.
Power Factor Correction
As is typical, the PFC pre-regulator is implemented in a boost topology and uses the FAN7529 PFC controller that operates in critical conduction mode -- also called boundary or transition mode. This mode is considered the most economic solution for a power load of 150 to 200W maximum. In critical mode the peak current through the boost inductor is controlled in such a way that it is proportional to the instantaneous rectified input voltage. During off-time however, the current goes back to zero and this zero crossing (i.e. de-magnetization of the inductor) is detected and initiates the next switching cycle. As you can see, the average inductor current is proportional to the input voltage -- the needed result.
The FAN7529 operates in so called voltage mode, where the conduction time of the MOSFET is kept constant during at least one power line half-cycle. Keeping on-time constant, peak switch current proportional to the input voltage can be easily derived from the basic differential equation dI/dt = V/L. The output voltage of the boost converter is then sensed and regulated by adjusting the MOSFET's on-time. The advantage of voltage mode in contrast to current mode, is that there is no need for sensing the rectified input voltage in order to generate a reference signal. This simplifies the controller itself and reduces component count.
A big advantage of the critical mode is that sensing the de-magnetization of the boost inductor before the next switching cycle starts results in zero current turn-on of the MOSFET. Thus switching losses are quite low and efficiency will be high, especially since reverse recovery of the rectifier diode is not an issue. On the other hand, peak input current is higher than in a continuous conduction mode (CCM) PFC and might make the EMI filtering more complicated.
The schematic of the PFC and the input stages is detailed in Figure 1. When the application is powered up, C96 is charged via R93a and R93b. As soon as the start voltage of IC91 is reached, normal operation starts. When this occurs, the gate of the MOSFET is driven through the network R96, D98 and R99, enabling fast turn-off and slower turn-on of the latter. The boost inductor consists of two distinct inductors since the universal input demands for a high current and high inductance device that can not be implemented with a single low profile core. The current through the MOSFET is monitored at the CS input of the controller in order to achieve pulse by pulse over-current protection.
The output voltage is scaled by the divider R910a & b and R911 and fed to the chip's error amplifier that is frequency compensated by the network connected to the COM pin. Then, the output of the error amplifier sets the on-time of the MOSFET accordingly. De-magnetization of the inductors is detected by monitoring the voltage across a secondary winding of one of the inductors that is fed to the ZCD input. The power supply of the controller during normal operation also comes from this secondary and is rectified and limited by the network R94, C913, ZD91 and D90. Resistor 91 sets the maximum on-time of the MOSFET so that it is less than the time needed at full load and minimum input voltage. The purpose of R92 is to add some modulation of the on-time with the input voltage to improve THD. The PFC pre-regulator as described generates a DC output voltage of 400V that is fed to the DC-DC controller.
Figure 1: Schematic of 200W DCM voltage mode PFC.
(Click this image to view a larger, more detailed version)