www.fairchildsemi.com Application Note AN6016 tm LCD Backlight Inverter Drive IC (FAN7311) 1. Description (PWM) driver stage, soft start, open lamp regulation, and Under Voltage Lockout protection (UVLO). The FAN7311 includes an internal shunt regulator that allows operation with an input voltage from 5V to 25.5V. It supports analog and burst dimming modes of operation. The FAN7311 provides open lamp regulation and protection. Open lamp regulation protects the transformer from over-voltage during start up or when an open lamp occurs. The transformer voltage is regulated by reducing duty cycle when an over-voltage is detected. Open lamp protection can be used to shut down an IC when an open lamp condition continues longer than a specified time. Design goals for a Cold Cathode Fluorescent Lamp (CCFL) inverter for use in a notebook computer or other portable applications include small size, high efficiency, and low cost. The FAN7311 provides the necessary circuit blocks to implement a highly efficient CCFL backlight power supply in 20-SSOP and 20-SOIC packages. The FAN7311 typically consumes less than 4mA of operating current, improving overall system efficiency. External parts count is minimized and system cost is reduced by the integration of features including; feedback-controlled Pulse Width Modulation F1 FUSE C22 220μ 25V CN5 1 2 3 4 5 6 7 8 9 10 IC1 FAN7311 0 12V OLP RT1 OLR OUTB R6 ENA 10k 12505WR-10 0 OUTA 0 0 S_S VIN DIM (0~3.3V) R4 22k 0 R2 56k C21 10n 1μ GND PGND 0 10k SP DP GP DP 0 C27 1μ 0 C7 10μ LTM190EX TX1 FDS8958A 1 M2 2 J1 REF OUTC ADIM OUTD 0 0 R27 C5 220p BDIM CT EA_IN RT 0 10k R527k 0 C4 4.7n 0 EA_OUT BCT SN DN GN DN SP DP GP DP FDS8958A C10 15p OLR FB OLR R26 1k 0 10k R9 9.1k OLP2 C14 10n R13 1k C30 10n 0 0 0 Q1 R22 10k 1 HOT OLP2 R28 10k C12 15p OLP3 OLP3 BAW56 R20 10k OLP4 R21 10k C19 2.2n 0 C20 2.2n D3 BAV70 R12 1k 0 0 R11 1k 0 C15 10n 0 BAV99 0 D9 R18 1k R19 1k BAV99 0 0 0 FB C17 2.2n 0 C29 10n CN4 CCFL OLP4 D8 OLR 0 0 BAW56 C18 2.2n HOT 2 COLD C13 15p BAW56 KST2222 D10 R17 1k 0 CN3 2 COLD CCFL OLP1 D1 0 FB 1 D11 BAV99 0 TX2 0 REF C9 1μ D7 R16 1k BAV99 0 R14 100k 0 CN2 CCFL D6 D4 BAV70 R15 100k 2 COLD C11 15p OLP1 RT R8 R1 330k CN1 CCFL COLD 1 HOT 4.7n OLP HOT C8 10μ 0 C3 0 R25 DN 1μ R70 R3 18k DN GN 0 C2 REF SN C6 0.22μ 0 0 RT OUTA C1 C28 10n 0 0.1μ 82k OUTB R24 ON/OFF C25 1μ M1 C26 0 Figure 1. Application Circuit REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE 2. Block Diagram and Basic Operation 2.1 Block Diagram RT CT max. 2V OUTA OSCILLATOR Output Driver min. 0.5V + + 6μA S_S OUTB Output Control Logic PGND 1mA OUTC Output Driver OUTD MRT1 RT1 Striking Logic OLP S_S 1.4μA EA_OUT UVLO + ADIM Error Amp. Q CLR R - EA_IN + VOLP+α Q SET S UVLO VOLP 2.5V 1.5V 2.5V + Solr BCT min. 0.5V OLR Solr 105μA - Sburst 85μA Va+α max. 2V Voltage Reference & Internal Bias - 2V 2.5VREF REF + Sburst BDIM AGND OLP - + VIN + ENA 1.4V VIN UVLO - UVLO 5V Figure 2. Block Diagram 2 REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE 2.2 Under Voltage Lockout (UVLO) The UVLO circuit guarantees stable operation of the IC’s control circuit by stopping and starting it as a function of the VIN value. The UVLO circuit turns on the control circuit when VIN exceeds 5V. When VIN is lower than 5V, the IC’s standby current is less than 200µA. 2.3 ENA Applying voltage higher than 2V to ENA pin enables the operation of the IC. Applying voltage lower than 0.7V to ENA pin disables the operation of the inverter. Voltage Reference & Internal Bias 2.5VREF REF + VIN UVLO ENA 2.5 Oscillator VIN + - Figure 5. Soft Start During Initial Operation 1.4V UVLO 5V 2.5.1 Main Oscillator Figure 3. Under Voltage Lockout and ENA Circuits 4 Timing capacitor CT is charged by the reference current source. The source is formed by the timing resistor RT whose voltage is regulated at 1.25V. The sawtooth waveform of the main oscillator circuit charges up to 2V, then the capacitor begins discharging down to 0.5V. The capacitor starts charging again and a new switching cycle begins. 3 1.25 4 RT (2.1) I ch arg e = --- ---------- Icc (mA) 2 0 5 10 VIN (V) 15 20 The main frequency can be programmed by adjusting the values of RT and CT. The main frequency can be calculated as shown below. 19 f op = ---------------------32 R T C T Figure 4. Start Voltage and Operating Current (2.2) 2.4 Soft Start The soft-start function is provided by the S_S pin and is connected through a capacitor to GND. A soft-start circuit ensures a gradual increase in the input and output power. The capacitor connected to S_S pin determines the rise rate of the duty ratio. It is charged by a current source of 6µA. Icharge 2V + S SET Q CT R CLR Q 20 x Icharge 0.5V + Figure 6a. Main Oscillator Circuit REV. 1.0.1 4/20/06 3 AN6016 APPLICATION NOTE Figure 6b. Main Oscillator Waveform Figure 7b. Burst Oscillator Waveform 2.5.2 Burst Dimming Oscillator Burst dimming timing capacitor BCT is charged by the reference current source, formed by the timing resistor RT whose voltage is regulated at 1.25V. The sawtooth waveform charges up to 2V. Once reached, the capacitor begins discharging down to 0.5V, then starts charging again and a new switching cycle begins. 3 1.25 I ch arg e = ------ ---------32 R T 2.6 Analog Dimming For analog dimming, the lamp intensity is controlled with the ADIM signal. A 2.5V on ADIM brings full brightness. Analog dimming waveforms are shown in Figures 8 and 9. (2.3) The burst dimming frequency can be programmed by adjusting the values of RT and BCT. The burst dimming frequency can be calculated as below. 3.75 f burst = -------------------------96 R T BC T (2.4) The burst dimming frequency should be greater than 120Hz to avoid visible flicker. To compare the input of BDIM pin with the 0.5~2V triangular wave of burst oscillator makes the PWM pulse for burst dimming. The PWM pulse controls EA_OUT voltage by summing 85µA into the EA_IN pin. Figure 7 shows burst dimming oscillator circuit and waveform. Icharge 2V Figure 8. Analog Dimming at Maximum + S SET Q CT R CLR Q 20 x Icharge 0.5V + Figure 7a. Burst Oscillator Circuit 4 Figure 9. Analog Dimming at Minimum REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE 2.6.1 Setting Lamp Current Sensing Resistors The data to input The calculated data 1) Positive Polarity Analog Dimming CFB VCS RFB CCFL CCFL RCS1 Vsense – VREF + Error Amp. RCS2 Rsense Rsense Figure 10. Calculating Value of the Analog Dimming Circuit Parameter VREF 2.5 Ilamp 6.5 Rsense 1 kΩ Rsense_eff 0.95 kΩ Diode drop voltage 0.3 V Vsense 5.259453252 V RCS1/RCS2 1.103781301 RCS1 10 kΩ RCS2 9.059765727 kΩ Rsense_effective 0.950148969kΩ = Rsense/(RCS1+RCS2) Lamp current is sensed at Rsense and the sensed voltage is divided by RCS1 and RCS2 and is averaged at Error Amp. by RFB and CFB. 2) Negative Polarity Analog Dimming CFB R sense_eq V DF 1 + ------------------------------------------------I lamp ⋅ ( R cs1 + R cs2 ) = ----------------------------------------------------------- ≈ R sense || ( R cs1 + R cs2 ), R RFB + ⎛1 V sense = ⎜ --⎝π π ∫0 ⎞ 2 ⋅ I Lamp ⋅ R sense_eq ⋅ sin θ ⋅ dθ⎟ – V DF ⎠ R cs2 2 V CS = V sense ⋅ ---------------------------- = ⎛ --- ⋅ 2 ⋅ I Lamp ⋅ R sense – V DF⎞ ⎝π ⎠ R cs1 + R cs2 R sense_eq R cs2 V ref = V CS = V sense ⋅ ---------------------------R cs1 + R cs2 R cs1 V sense ---------- = --------------- – 1 R cs2 V ref VREF RCS2 Rsense Rsense Lamp current is sensed at Rsense and the sensed voltage is divided by Rcs1 and Rcs2 and is averaged at Error Amp. by RFB and CFB. (2.5) Equation (2.5) assumes that the error amplifier loop is closed. The relationship between VCS and Vref is given in equation (2.6). R cs2 2 = ⎛ --- ⋅ 2 ⋅ I Lamp ⋅ R sense – V DF⎞ ⋅ ---------------------------⎝π ⎠ R cs1 + R cs2 CCFL Figure 11. Calculating Value of the Analog Dimming Inverting Circuit Parameter 2 = --- ⋅ 2 ⋅ I Lamp ⋅ R sense_eq – V DF ( 1 ) π R cs2 ⋅ ---------------------------R cs1 + R cs2 CCFL – Error Amp. V DF is diode forward voltage VA Vsense VCS R CS1 V DF 1 + ------------------------------------------------I lamp ⋅ ( R cs1 + R cs2 ) = ----------------------------------------------------------- ≈ R sense || ( R cs1 + R cs2 ), R V DF is diode forward voltage ⎛1 V sense = ⎜ --⎝π π ∫0 ⎞ 2 ⋅ I Lamp ⋅ R sense_eq ⋅ sin θ ⋅ dθ⎟ – V DF ⎠ 2 = --- ⋅ 2 ⋅ I Lamp ⋅ R sense_eq – V DF ( 1 ) π (2.6) (2.7) R cs2 2 V CS = V sense ⋅ ---------------------------- = ⎛ --- ⋅ 2 ⋅ I Lamp ⋅ R sense – V DF⎞ ⎝π ⎠ R cs1 + R cs2 R cs2 ⋅ ---------------------------R cs1 + R cs2 (2.8) For example, suppose: V ref = 2.5V, I Lamp = 6.5mA, R sense = 1kΩ, R cs1 = 10kΩ From these values, an approximate value of Rcs2 can be derived. To get a more precise value for RCS2, use an iterative calculation. Use Rsense to calculate RCS2, because the Rsense_eq value is unknown. After finding the value of Rsense_eq, use Rsense_eq to calculate RCS2. Calculate iteratively until the previous Rsense_eq value is almost equal to the current Rsense_eq value. REV. 1.0.1 4/20/06 Equation (2.8) assumes the error amplifier loop is closed. The relationship between VCS and VA (dimming control voltage) is given in equation (2.9). V A ⋅ R FB + V CS ⋅ R A V ref = --------------------------------------------------R FB + R A (2.9) 5 AN6016 APPLICATION NOTE The relationship between dimming control voltage and lamp current can be programmed for the application. For example, suppose: V Amin. = 0, I Lamp.max = 7mA (2.10) V Amax. = 3.3, I Lamp.min = 3mA (2.11) I Lamp.min = α ⋅ I Lamp.max (2.12) Substituting for VA and VCS in equation (2.9) from equation (2.10) results in: V CSmax ⋅ R A V REF = ------------------------------R FB + R A V ref ⋅ ( R FB + R A ) V ref ⋅ ( 1 + β ) ⋅ R FB V CSmax = ------------------------------------------- = ---------------------------------------------RA β ⋅ R FB V ref ⋅ ( 1 + β ) 1 = -------------------------------- = V ref ⋅ ⎛ 1 + ---⎞ ⎝ β β⎠ R cs2 = V sense ⋅ ---------------------------R cs1 + R cs2 R cs2 2 = ⎛ --- ⋅ 2 ⋅ I Lamp ⋅ R sense – V DF⎞ ⋅ ---------------------------⎝π ⎠ R cs1 + R cs2 (2.22) R cs1 V sense ---------- = ----------------------------- – 1 R cs2 1 V ref ⎛ 1 + ---⎞ ⎝ β⎠ (2.13) (2.23) For example: Substituting for VA and VCS in equation (2.9) from equation (2.11) results in: V ref = 2.5V, I lampmax = 6.7mA, I lampmin = 4mA, V Amax ⋅ R FB + V CSmin ⋅ R A V ref = -------------------------------------------------------------------R FB + R A R FB = 100kΩ, R sense = 1.5kΩ , R CSI = 10kΩ V Amax ⋅ R FB + α ⋅ V CSmax ⋅ R A = ----------------------------------------------------------------------------R FB + R A (2.14) Multiplying equation (2.13) by RFB + RA gives: V ref ⋅ R FB + V ref ⋅ R A = V CSmax ⋅ R A (2.15) From these values, it is possible to obtain the value of RCS2. To get more precise value of RCS2, use an iterative calculation. Use Rsense to calculate RCS2, because Rsense_eq is unknown. After the first calculation, Rsense_eq can be resolved. Calculate the RCS2 value using Rsense_eq. Calculate iteratively until the previous Rsense_eq value becomes almost equal to the current Rsense_eq value. The data to input Multiplying equation (2.14) by RFB + RA gives: The calculated data Vref ⋅ R FB + Vref ⋅ R A = VAmax ⋅ R FB + α ⋅ VCSmax ⋅ R A (2.16) ( V ref – V Amax ) ⋅ R FB + V ref ⋅ R A = α ⋅ V CSmax ⋅ R A (2.17) Multiplying equation (2.15) by α gives: α ⋅ V ref ⋅ R FB + α ⋅ V ref ⋅ R A = α ⋅ V CSmax ⋅ R A ( α – 1) ( V Amax + V ref ) ⋅ R FB + V ref ( α – 1) ( α – 1) ) ⋅ R FB = V ref ⋅ RA = 0 (2.19) ( α – 1) (2.20) ⋅ RA Equation (2.20) can be rewritten as: ( α – 1) ( V Amax + V ref ) ⋅ R FB R A = ----------------------------------------------------------------- = β ⋅ R FB ( α – 1) V ref (2.21) RA is calculated by selecting RFB and solving equation (2.21). Substituting for RA in equation (2.13) from equation (2.21) and rewriting gives: 6 VA Ilamp Min. – 0 4 Typ. 2.5 - – Max. – 3.2 6.7 α 0.597014925 6.7 (2.18) Subtracting equation (2.17) from equation (2.18) gives: ( V Amax + V ref VREF Ilamp_max/Ilamp_min RFB 100 kΩ RA 217.6296296 kΩ b 2.176296296 VCSmax 3.64874064 Rsense 1.5 Rsense_eff 1.3861 kΩ Diode drop voltage 0.3 V Avg_maxVsense 8.061120587 V RCS1/RCS2 1.209288459 RCS1 10 kΩ RCS2 8.269325588 kΩ Rsense_effective 1.386187316kΩ =Rsense//(RCS1+RCS2) REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE 2.7 Burst Dimming Lamp intensity is controlled with the BDIM signal. 0V on BDIM commands full brightness. The duty cycle of the burst dimming comparator determines the lamp brightness as a percent of the rated lamp current. Burst dimming is implemented by summing 85µA into the feedback node to turn down the lamp. If there is sufficient voltage for the lamp to strike, the feedback loop controls the lamp at the rated current using a fixed current-sense resistor. When the voltage of EA_IN is brought higher than Vref, EA_OUT becomes low and the MOSFET stops switching. At this time, the resonant tank voltage decays until the lamp extinguishes. CFB is reduced, if possible, to speed up the lamp re-strike. Burst dimming waveforms are shown in Figures 12, 13, and 14. Figure 14. Burst Dimming at 25% 2.8 Open Lamp Regulation and Open Lamp Protection Power stage operation must be suspended if an open lamp occurs, because the power stage is at high gain. When a voltage higher than 2V is applied to the OLR (Open Lamp Regulation) pin, the part enters the regulation mode and controls EA_OUT voltage to limit the lamp voltage by adding 105µA into the feedback node. The OLP (Open Lamp Protection) capacitor, which is connected to the OLP pin, is charged by the 1.4µA internal current source. Figure 12. Burst Dimming at 75% 2.8.1 Open Lamp at Initial Operation OLP voltage starts from 1V. After reaching 2.5V, the IC shuts down when all the output are high. The relationship between the OLP capacitor and the time ΔT before the IC shuts down is calculated using the approximation I = CΔV/ΔT, where I = 1.4μA, ΔV = 1.5V, resulting in ΔT(s) = 1.1C(μF). 2.8.2 Normal Operation and Open Lamp OLP voltage starts at 0V. After reaching 1.5V, the IC shuts down when all the outputs are high. The relationship between the OLP capacitor and the time ΔT before the IC shuts down is calculated using the approximation I = CΔV/ΔT, where I = 1.4μA, ΔV = 1.5V, resulting in ΔT(s) = 1.1C(μF). Figure 13. Burst Dimming at 50% REV. 1.0.1 4/20/06 7 AN6016 APPLICATION NOTE 2.8.3 OLP Operation CCFL1 CCFL2 Lamp current 1 Lamp current 2 R1 R3 FB D1 R4 R2 D2 R5 OLP VREF R7 R8 C1 C2 D3 1V 0 1V 0 R6 Q1 C3 1V 0 1V 0 D4 Normal operation Open lamp Figure 15. Operating OLP In normal operation, the voltage of D3’s cathode is over 1V and D3 is turned off; Q1 is on; and OLP remains low. When open lamp occurs, the voltage of D3’s cathode is under 1V and either D3 is turned on. Then Q1 is turned off and OLP start charging by an internal current source of 1.4µA. If OLP reaches 2V, the IC is shut down. The base current of Q1 should be more than 1.4µA/hfe. R6 is determined by this condition. R4, R5, R7, and R8 are determined so that Q1 and D4 are turned off in the open lamp condition. C1 and C2 are determined so that the voltage of D3’s cathode is over 1V in normal operation. 8 REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE 2.8.4 OLR Operation D3 D4 OLR>2V 105μA Solr - CCFL2 CCFL1 OLR + 2V Lamp current 2 Lamp current 1 R3 R1 D1 + Error Amp. R2 VREF R4 D2 R5 OLR > 2V ↑ → Solr on duty ↑ → EA_IN ↑ → EA_OUT ↑ → Switching duty ↑ → Lamp voltage ↓ → OLR > 2V ↓ → Solr on duty ↓ → EA_IN → EA_OUT ↓ → Switching duty ↑ → Lamp voltage ↑ Figure 16. Operating OLR 105µA + Solr - OLR 2V To EA_IN Figure 17. Open Lamp Regulation Circuit To Cont rol Logic Figure 19. OLR Voltage During Striking Mode 1.4µA UVLO OLP + Q Q SET CLR S R - VOLP 2.5V 1.5V UVLO Figure 18. Open Lamp Protection Circuit REV. 1.0.1 4/20/06 9 AN6016 APPLICATION NOTE 2.9 Output Driver The four output drives are designed so that the two pairs of switches, pair A and B and pair C and D, never turn on simultaneously. The OUTA-OUTB pair is intended to drive Dead Time Dead Time Dead Time Dead Time one half-bridge in the external power stage. The OUTCOUTD pair drives the other half-bridge. The detailed timing relationship is shown below. Dead Time Dead Time Dead Time Dead Time Dead Time EA_OUT CT SYNC T T1 POUT A NOUT B POUT C NOUT D A&D B&C Figure 20. New Phase Shift Control Waveforms 10 REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE 2.10 CCFL Striking Sequence ENA ENA 4V S_S S_S 0.5V 1.5V OLP OLP 1V 0.6V RT1 Normal RT1 Fnor Striking frequency F st = Open lamp 19 . 1 64 (RT // RT1 ) . C T shutdown 19 1 = . 64 (RT // RT1) .CT Normal operation frequency Fnor = 19 . 1 64 (RT // RT1 ) . CT Figure 23. Open Lamp Figure 21. CCFL Ignites ENA S_S 0.5V 2.5V 1V OLP RT1 Striking frequency shutdown 19 1 Fst = 64 (RT // RT1 ) .CT Figure 22. CCFL Does Not Ignite REV. 1.0.1 4/20/06 11 AN6016 APPLICATION NOTE 2.11 PCB Layout Guideline 1. Separating ground for analog and power portions of circuitry is one of the simplest and most effective methods of noise suppression. This is shown in Figure 24. 2. The traces between drive output and the MOSFET gates should be as short as possible and as wide as possible. 3. The traces of RT, CT, and BCT should be kept away from high-current components and traces. FAN7311 OLP RT1 OLR OUTB ENA OUTA S_S VIN GND PGND REF OUTC ADIM OUTD BDIM CT EA_IN RT EA_OUT BCT PGND AGND Figure 24. PCB Layout 12 REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE 3. Power Stage Design 3.1 Resonant Circuit Cc L l1 N p:Ns LM V Cp CCFL A circuit of LCD backlight inverter Cp CCFL A circuit of LCD backlight inverter Ideal VIN 0 (a) -VIN Cc L l1 N p:Ns L l2 LM Vrms Vrms = L l2 2 2 V π IN Ideal (b) Cc /n2 n 2*L l1 Ns /Np =n L l2 n 2*LM nVrms Cp R lamp Ideal transformer is neglected. Cp R lamp Assuming LM is infinite, IM is near zero. LM is neglected. Cp R lamp Primary and secondary side leakage inductances are combined. Cp R lamp DC blocking capacitor is neglected. (c) Cc /n2 n 2*L l1 L l2 nVrms (d) Cc /n2 Ll nVrms (e) Ll nVrms (f) Z Ll Cs Rs nVrms (g) Transform the parallel resonant circuit into the series resonant circuit. Z Figure 25. Resonant Circuit REV. 1.0.1 4/20/06 13 AN6016 APPLICATION NOTE The resonant circuit (f) in Fig. 25 is a second-order low-pass filter and can be described by the following normalized parameters: The loaded quality factors QL and Qr are related by: ωr L 1 Q r = --------- = ------------------ = ω r C p R lamp Rs ωr Cs Rs • The corner frequency: 1 ω 0 = ---------------LI Cp ωρ = Q r ⎛ −−− ⎞ = ⎝ ω 0⎠ (3.1) 2 Q L – 1 , for Q L ≥ 1 (3.12) 3.2 Voltage Transfer Function • The characteristic impedance: v(t) 1 Z 0 = ω 0 L I = ------------- = ω0 Cp LI -----Cp (3.2) VIN 2Dπ 3 π 2 • The loaded quality factor at the corner frequency fo: R lamp R lamp Q L = ω 0 C p R lamp = -------------- = -------------ω0 LI Z0 • The resonant frequency that forms the boundary between capacitive and inductive loads: ωr LI 1 Q r = ----------- = -----------------Rs ωr Cs Rs (3.5) • The input impedance of the resonant circuit (f) in Fig. 25 is: ω 1 ω 1 R lamp 1 – ⎛ ------⎞ + j ------- ⎛ ------⎞ R lamp ⋅ ------------⎝ ω 0⎠ Q L ⎝ ω 0⎠ jωC p Z = jωL + ---------------------------------- = -----------------------------------------------------------------------1 ω R lamp + ------------1 + jQ L ⎛ ------⎞ ⎝ ω 0⎠ jωC p 2 (3.6) = R S + jX s ω 2 2 ⎛ ω ⎞2 1 – ⎛ ------⎞ + -----⎝ ω 0⎠ ⎝ ω 0⎠ ------------------------------------------------------------2 ω 1 + ⎛ Q L ------⎞ ⎝ ω 0⎠ ⎧ ⎫ ω 1 ω ϕ = arc tan ⎨ Q L ⎛ ------⎞ ⎛ ------⎞ + --------2- – 1 ⎬ ⎝ ω 0⎠ ⎝ ω 0⎠ QL ⎩ ⎭ (3.7) As shown in Fig. 26, the input voltage of the resonant circuit v is a square wave of magnitude VIN, given by: 1 v = 0, for 0 < ωt ≤ ⎛ --- – D⎞ π ⎝2 ⎠ 1 1 v = V IN , for ⎛ --- – D⎞ π < ωt ≤ ⎛ --- – D⎞ π ⎝2 ⎠ ⎝2 ⎠ 1 3 v = 0, for ⎛ --- + D⎞ π < ωt ≤ ⎛ --- – D⎞ π ⎝2 ⎠ ⎝2 ⎠ 3 v = 0, for ⎛ --- – D⎞ π < ωt ≤ 2π ⎝2 ⎠ v i1 = V m sin ( ωt ), (3.14) (3.8) (3.9) X S = Z sin ϕ (3.10) • The resonant frequency, fr, is defined as a frequency at which the phase shift is zero. The ratio of fr to the corner frequency, fo, is: (3.11) 4 V m = --- V IN sin Dπ π (3.15) You can obtain the rms value of vi1: Vm 2 2 V rms = -------- = ---------- V IN sin Dπ π 2 (3.16) Which leads to the voltage transfer function from VIN to the fundamental component at the input of the resonant circuit: V rms 2 2 M Vs = ------------ = ---------- sin Dπ V IN π 14 (3.13) in which the amplitude of vi1 can be found from Fourier analysis as: R S = Z cos ϕ 1 1 – --------2- , for Q L ≥ 1 QL -VIN The fundamental component of this voltage is: 2 QL 2 fr ---- = f0 ωt 2π 3 3 v = – V IN , for ⎛ --- – D⎞ π < ωt ≤ ⎛ --- – D⎞ π ⎝2 ⎠ ⎝2 ⎠ where, Z ------ = Z0 2D π (3.4) • The loaded quality factor at the resonant frequency fr: jϕ π Figure 26. Input Voltage of the Resonant Circuit 1 ω r = ---------------LI Cs = Ze π 2 (3.3) (3.17) REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE According to Fig. 25(f), the voltage transfer function of the resonant circuit is: M Vr 2 2⋅n 1 M Vl ( max ) = ------------------- , for 0 ≤ Q L ≤ ------π 2 R lamp -------------jωC p ---------------------------------1 R lamp + ------------V Ri jωC p = --------------------------- = --------------------------------------------------R lamp 2 ⋅ nV rms -------------jωC p jωL + ---------------------------------1 R lamp + ------------jωC p 1 ejϕ = --------------------------------------------------- = M Vr ω 2 1 ω 1 – ⎛ ------⎞ + j ------- ⎛ ------⎞ ⎝ ω 0⎠ Q L ⎝ ω 0⎠ The maximum magnitude of the DC-to-AC voltage transfer function of the LCD backlight inverter without losses is: 2 2 ⋅ nQ L 1 M Vl ( max ) = ----------------------------- , for Q L ≥ ------1 2 π 1 – -----------24Q L 3.2 Design Procedure (3.18) A LCD monitor backlight circuit illustrates a design based on the FAN7311. The inverter is designed to drive two CCFLs with the following specifications. Panel Model where, V Ri 1 M Vr = --------------- = -----------------------------------------------------------------nV rms 1 ω 2 ω ⎞2 2 ⎛ + --------2- ⎛ ------⎞ 1 – -----⎝ ω 0⎠ ⎝ ⎠ QL ω0 1 ω ------- ⎛ ------⎞ Q L ⎝ ω 0⎠ ϕ = -arctan ----------------------2ω 1 – ⎛ ------⎞ ⎝ ω 0⎠ (3.24) (3.19) (3.20) LM151X2(LG.PHILIPS LCD) Input Voltage 9 ~ 15V Striking Voltage 880Vrms Operating Voltage 585Vrms (Typ.) Operating Current 8mArms (Typ.) Operating Frequency 50kHz (Typ.) Rated Power 4.68W/CCFL Efficiency 85% (Typ.) 1) Select Transformer’s Primary Turns The maximum value of MVr is obtained by differentiating the quantity under the square-root sign with respect to f/fo and setting the result equal to zero. Hence, the normalized peak frequency is: f pk 1 ------- = 0, for 0 ≤ Q L ≤ ------f0 2 f pk ------- = f0 (3.21) VIN,min = Minimum input voltage (Volts) ΔB = Core magnetic flux density change (Tesla) resulting in the maximum magnitude of the voltage transfer function of the resonant circuit: 1 M Vr ( max ) = 1, for 0 ≤ Q L ≤ ------2 (3.22) The magnitude of the DC-to-AC voltage transfer function of the LCD backlight inverter without losses is obtained from (3.17) and (3.22): V lamp M VI = -------------- = M Vs ⋅ ( nM Vr ) V IN 2 2 ⋅ n sinDπ = -------------------------------------------------------------------ω 2 1 ω 2 π 1 – ⎛ ------⎞ + --------2- ⎛ ------⎞ ⎝ ω 0⎠ ⎝ ⎠ QL ω0 REV. 1.0.1 4/20/06 V IN, min ⋅ Δt max N p, min = -------------------------------------ΔB ⋅ A e where Np,min = Minimum number of primary turns 1 1 1 – -----------2-, for Q L ≤ ------2 4Q L QL 1 M Vr ( max ) = ------------------------- , for Q L ≤ ------2 1 1 – -----------24Q L The number of primary turns is determined by Faraday’s law. Np,min is fixed by the minimum voltage across the primary and the maximum on time. (3.23) Δtmax = Maximum overlap on-time of diagonal MOSFET switches (us) Ae = Core cross-sectional area (mm2) A transformer used in a full-bridge topology operates in two quadrants of the B-H curve such that the maximum magnetic flux density is Bmax = 0.5ΔB. For most cores, saturation magnetic flux density is about 400mT. Margin considered, determine that the maximum magnetic flux density Bmax = 0.5 Bsat, so the maximum magnetic flux density is Bmax = 200mT. In an example with a minimum voltage of 9V, operating frequency 50KHz, maximum on time of diagonal MOSFET switches of 10µs and a core cross-sectional area (EPC17, EPC19, EFD1820) of 22mm2, the minimum number of primary turns required is: V IN, min ⋅ Δt max 9 ⋅ 10 N p, min = -------------------------------------- = ------------------- ≈ 10T s ΔB ⋅ A e 400 ⋅ 22 15 AN6016 APPLICATION NOTE 2) Select QL and Operation Frequency to Determine the Turns Ratio Select a value of 1 for QL. Assume that fop = fpk = 50kHz based on the LCD panel specification. From (3.21), the corner frequency is: f pk 50 f o = ------------------------- = ------------------------------ = 70.7 ( kHz ) 1 1 1 – -----------21 – ----------------22Q L 2 ⋅ 1.1 From (3.11), the resonant frequency that forms the boundary between capacitive and inductive loads is: 1 1 – --------2QL 1 f r = f o ⋅ 1 – --------2- = f pk ⋅ ------------------------- = 0 ( kHz ) 1 QL 1 – -----------22Q L Therefore, zero-voltage switching (ZVS) can be achieved at any operating frequency. For the reference design, the required secondary lamp voltage is 585V and the minimum voltage is 9V. Therefore, from (3.23), the minimum number of the turns ratio is: ω op 2 2 1 ω op 2 + --------2- ⎛ ---------⎞ 1 – ⎛ ---------⎞ ⎝ ω0 ⎠ ⎝ ⎠ QL ω0 585 n ≥ --------- ⋅ --------------------------------------------------------------------------- ≈ 62.5T s , 9 2 2 sin Dπ π 2 2 ⋅ n sin Dπ 585 ∴M VI = ---------------------------------------------------------------------- ≥ --------9 1 ⎛ ω ⎞2 ω ⎞2 2 ⎛ + --------2- -----π 1 – -----⎝ ω 0⎠ ⎝ ω 0⎠ QL 3) Determine the Required Output Capacitance Using the above specifications, the equivalent resistance of a CCFL is: The corner frequency is 70.7kHz. Assume a parasitic capacitance per lamp of 10pF. Each parasitic capacitance is effectively in parallel with each of the output capacitors. The output capacitor is: QL C out = C p – C para = --------------------- – 10pF ≈ 21pF, ω o R lamp ∴Q L = ω o R lamp C p Using (3.3), the value of the leakage inductance is: 1 ≈ 164.6 ( mH ) L I = --------------2 ωo Cp Note: Considering minimum primary turns, minimum turns ratio, and leakage inductance, determine primary turns, turns ratio, and the gap of core to get the required leakage inductance. For the sample design, the number of primary turns is 30Ts and that of the secondary turns is 2200Ts. Turns ratio is 66.7. 4) Select the Proper Wire Gauges for the Primary and Secondary Transformer Windings The approximate primary winding rms current Ip and approximate secondary winding rms current Is are determined by the following equations. π P lamp I p = ---------- ⋅ -------------, 2 2 ηV IN 2 2 ∴P lamp = ηV IN I IN , I IN = ----------I p π Is = 2 I lamp + [ 2πf op C out V lamp ] 2 V lamp 585 R lamp = -------------- = ------------- ≈ 75 ( kΩ ) I lamp 0.008 16 REV. 1.0.1 4/20/06 AN6016 APPLICATION NOTE Values that must be known or selected initially: Parameter Description Typical Value Units Vlamp Nominal lamp operating voltage 585 V Ilamp Nominal lamp operating current 8 mA fop Operating frequency 50 kHz fpk Peak frequency 50 kHz Vin Input voltage 9 V D Duty ratio at input voltage 50 % QL Loaded factor at the corner frequency 1 Cpara Parasitic capacitance 10 pF Ae Core cross-sectional area 22 mm2 Bsat Saturation magnetic flux density 0.4 T ALleakage AL value of leakage inductance 22 nH/N2 Values that are calculated: Bmax Maximum magnetic flux density 0.2 T ΔB Core magnetic flux density change 0.4 T Δtmax Maximum overlap on-time of diagonal switches 10 µs fo Corner frequency 70.71067812 kHz fr Resonant frequency Rlamp Equivalent resistance of a CCFL Np,min The minimum number of transformer’s primary turns nmin The minimum number of the turns ratio Cout The output capacitor 20.78 pF Ll The leakage inductance of the transformer 164.59 mH Np The number of transformer’s primary turns 31 Turns 0 kHz 73.125 k¾ 10 Turns 62.5 Values that must be selected with more than minimum turn ratio. n The number of the turns ratio 62.5 The values that are calculated: Ns The number of transformer’s secondary turns REV. 1.0.1 4/20/06 1934.1 Turns 17 AN6016 APPLICATION NOTE DISCLAIMER FAIRCHILD SEMICONDUCTOR RESERVES THE RIGHT TO MAKE CHANGES WITHOUT FURTHER NOTICE TO ANY PRODUCTS HEREIN TO IMPROVE RELIABILITY, FUNCTION, OR DESIGN. FAIRCHILD DOES NOT ASSUME ANY LIABILITY ARISING OUT OF THE APPLICATION OR USE OF ANY PRODUCT OR CIRCUIT DESCRIBED HEREIN; NEITHER DOES IT CONVEY ANY LICENSE UNDER ITS PATENT RIGHTS, NOR THE RIGHTS OF OTHERS. THESE SPECIFICATIONS DO NOT EXPAND THE TERMS OF FAIRCHILD’S WORLDWIDE TERMS AND CONDITIONS, SPECIFICALLY THE WARRANTY THEREIN, WHICH COVERS THESE PRODUCTS. LIFE SUPPORT POLICY FAIRCHILD’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT DEVICES OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT OF FAIRCHILD SEMICONDUCTOR CORPORATION. As used herein: 1. Life support devices or systems are devices or systems which, (a) are intended for surgical implant into the body, or (b) support or sustain life, or (c) whose failure to perform when properly used in accordance with instructions for use provided in the labeling, can be reasonably expected to result in significant injury to the user. 2. A critical component is any component of a life support. www.fairchildsemi.com 4/20/06 0.0m 001 © 2006 Fairchild Semiconductor Corporation