|
Low Cost Power Supply for High Power LED Applications
by Dr. Michael Weirich, Fairchild Semiconductor GmbH (LpR Issue 02, p 28-31)  Since high power LEDs are more and more used in general lighting there is an increasing demand for off-line power supplies to drive these. Due to the V-I characteristic of LED such a power supply must have a constant current output. The following article describes a PSU based on a Fairchild Power Switch (FPS™) that realizes constant secondary current with primary-side regulation. The absence of operational amplifiers and an optocoupler for the stabilization of the output current makes this PSU extremely cost effective in case safety isolation is needed. Conventional constant current output PSU
The traditional approach for a constant output current offline power supply (PSU) is shown in Fig. 1. This PSU is based on flyback topology and can deliver a current of 700mA and a maximum output voltage of 5.1V from a input voltage from 85 VRMS to 265 VRMS. As the technical data imply, the PSU is mentioned to drive a 3W high power LED. The circuit is quite straightforward: After rectification (BD1 - BD4) and filtering (C2, C3 and L1) of the mains input, a flyback with the FPS™ FSQ510 follows.
The FSQ510 is the ‘smallest’ member of a family of integrated circuits that contain all necessary functions to implement a state of the art switched mode power supply. While the FSQ510 is a monolithic device with integrated 700V Sense-FET, the higher power members of the family are two chip devices consisting of a controller and a separate Sense-FET with a VDS of 650V. Basic functions and behavior are almost identical for all members of the family, thus the description of operation of the PSU with FSQ510 may stand for the complete family. After connecting the PSU to the mains, operation starts via the internal start-up path of the device i.e. an internal high voltage JFET charges C8 to the start voltage of 13V typical. As soon as this voltage level is reached, the internal Power-MOSFET starts to switch and normal operation of the PSU begins. The internal JFET is switched off now to reduce consumption of the power supply. The power for the FPS™ is now supplied from a separate winding of the transformer, rectified with D2 and filtered by R7 and C8. RS2 and RS3 together with DS1 and C82 form a clamping or ‘snubber’ network that takes the energy stored in the leakage inductance of the transformer. This is necessary to limit the drain voltage to a save level. 
Fig. 1: Conventional constant current output power supply
The transformed voltage is rectified by D1 and filtered by C4 with post filtering by L2 and C5. The output voltage is regulated with the network consisting of R2, R3, R5, R6, U1 and U2. U1 couples the feedback signal to the primary side and C6 and R17 form a frequency compensation network in order to get stable close loop operation.
The actual output current in this application is sensed with the shunt R11|R13|R14 and regulated with the help of Q1 and U1. When the voltage drop across the shunt exceeds the VBE of Q1, a current starts to flow through the LED of U1 and in turn the voltage at the feedback pin of the FPS™ is lowered. In sequence the duty cycle of the Power-MOSFET is reduced and at last the output voltage respective current. Since the VBE of a BJT is strongly temperature dependant, a compensation network consisting of R10 and the NTC THR1 is added. Purpose of R8 and R9 is to disable U2 in order the voltage loop doesn’t spoil operation of the current regulation loop.
The network R12, R15, R16, D4 and C10 form a network that enables Quasi-Resonant-Switching of the Power-MOSFET inside the FPS™. QR-Switching means that the drain voltage is monitored and the actual turn-on of the MOSFET happens when the drain voltage is at minimum. This uses the fact that after the energy stored in the transformer has been transferred completely to the secondary, an oscillation of the drain voltage occurs. This oscillation is due to the resonant network formed by the magnetizing inductance of the transformer and the drain-source capacitance of the MOSFET. By switch-on at minimum drain voltage the switching losses are considerably reduced and EMI performance is improved. The synchronization network is actually not connected to the drain of the MOSFET but to the VCC winding of the transformer, that has identical waveform but lower amplitude.
Primary Side Regulated Constant Current PSU
In a flyback converter it is possible to get a fair regulation of the output voltages without explicitly regulating these. That is due to the fact that – if parasitic effects are neglected - the ratio of two output voltages is equal to the winding ratio of the respective transformer windings. Consequently it is possible to regulate e.g. the voltage of the VCC winding and get quite stable isolated outputs without using an optocoupler. Fig. 2 shows the schematic of a primary side regulated power supply – still without the constant current feature.
Most blocks are identical to the secondary side regulated PSU, but the feedback loop is completely different. As mentioned before, feedback is taken from the same transformer winding that supplies the FPS™. D3 rectifies this voltage and feds the network R2/C7 that generates the VCC of the chip and R4/C4 that filters the feedback voltage. In general the feedback could be taken from C7 as well. But since the FPS™ needs quite a big capacitor to support start-up current consumption it is better to have the additional path with a different time constant. The Zener-Diode D7 delivers base current to Q1 which acts as error amplifier. If VCC and at the same time the output voltages rise, this transistor gets more base current and in turn lowers the potential at the feedback pin of the FPS™, similar to the PSU with optocoupler feedback. ... Full article and access to all LpR back issues - Subscribe HERE |
