LED7706TR

LED7706 6-rows 30 mA LEDs driver with boost regulator for LCD panels backlight Features ■ ■ Boost section – 4.5 V to 36 V input voltage range – Internal power MOSFET – Internal +5 V LDO for device supply – Up to 36 V output voltage – Constant frequency peak current-mode control – 250 kHz to 1 MHz adjustable switching frequency – External synchronization for multi-device application – Pulse-skip power saving mode at light load – Programmable soft-start – Programmable OVP protection – Stable with ceramic output capacitors – Thermal shutdown Backlight driver section – Six rows with 30 mA maximum current capability (adjustable) – Rows disable option – Less than 500 ns minimum dimming ontime (1 % minimum dimming duty-cycle at 20 kHz) – ±2 % current matching between rows – LED failure (open and short-circuit) detection VFQFPN-24 4x4 Description The LED7706 consists of a high efficiency monolithic boost converter and six controlled current generators (rows) specifically designed to supply LEDs arrays used in the backlighting of LCD panels. The device can manage an output voltage up to 36 V (i.e. 10 white LEDs per row). The generators can be externally programmed to sink up to 30 mA and can be dimmed via a PWM signal (1 % dimming duty-cycle at 20 kHz can be managed). The device allows to detect and manage the open and shorted LED faults and to let unused rows floating. Basic protections (output over-voltage, internal MOSFET over-current and thermal shutdown) are provided. Applications ■ LCD monitors and TV panels ■ PDAs panel backlight ■ GPS panel backlight Table 1. Device summary Order codes Package LED7706 Packaging Tube VFQFPN-24 4x4 (exposed pad) LED7706TR April 2009 Tape and reel Rev 2 1/46 www.st.com 46 Contents LED7706 Contents 1 Typical application circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2 Pin settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3 2.1 Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.2 Pin description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Electrical data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3.1 Maximum rating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3.2 Thermal data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 4 Electrical characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 5 Operation description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 5.1 5.2 5.3 2/46 Boost section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 5.1.1 Functional description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 5.1.2 Enable function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 5.1.3 Soft-start . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 5.1.4 Overvoltage protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 5.1.5 Switching frequency selection and synchronization . . . . . . . . . . . . . . . 15 5.1.6 Slope compensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 5.1.7 Boost current limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 5.1.8 Thermal protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Backlight driver section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 5.2.1 Current generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 5.2.2 PWM dimming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Fault management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 5.3.1 FAULT pin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 5.3.2 MODE pin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 5.3.3 Open LED fault . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 5.3.4 Shorted LED fault . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 LED7706 6 Contents Application information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 6.1 System stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 6.1.1 Loop compensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 6.2 Thermal considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 6.3 Component selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 6.4 6.5 6.3.1 Inductor selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 6.3.2 Capacitors selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 6.3.3 Flywheel diode selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Design example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 6.4.1 Switching frequency setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 6.4.2 Row current setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 6.4.3 Inductor choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 6.4.4 Output capacitor choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 6.4.5 Input capacitor choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 6.4.6 Over-voltage protection divider setting . . . . . . . . . . . . . . . . . . . . . . . . . 33 6.4.7 Compensation network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 6.4.8 Boost current limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 6.4.9 Power dissipation estimate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Layout consideration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 7 Electrical characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 8 Package mechanical data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 9 Revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3/46 Typical application circuit 1 LED7706 Typical application circuit Figure 1. Application circuit L VIN VOUT Slope Compensation OVP selection MLCC +5V SWF OVSEL LX LDO5 SLOPE VIN AVCC Switching Frequency selection ROW1 Internal MOS OCP BILIM ROW2 RILIM ROW3 Rows current selection LED7706 SS ROW4 PGND MODE SYNC ROW6 EN SGND FAULT ROW5 DIM COMP Up to 10 WLEDs per row Dimming Fault Enable Faults Management Selection Sync Output AM00594v1 4/46 LED7706 Pin settings 2 Pin settings 2.1 Connections COMP 1 LX DIM EN FAULT SYNC SS Pin connection (through top view) 19 18 24 RILIM BILIM ROW6 LED7706 FSW ROW5 MODE ROW4 13 12 6 ROW3 ROW2 ROW1 SGND SLOPE 7 VIN AVCC OVSEL PGND LDO5 Figure 2. AM00595v1 5/46 Pin settings 2.2 LED7706 Pin description Table 2. 6/46 Pin functions N° Pin Function 1 COMP Error amplifier output. A simple RC series between this pin and ground is needed to compensate the loop of the boost regulator. 2 RILIM Output generators current limit setting. The output current of the rows can be programmed connecting a resistor to SGND. 3 BILIM Boost converter current limit setting. The internal MOSFET current limit can be programmed connecting a resistor to SGND. 4 FSW Switching frequency selection and external sync input. A resistor to SGND is used to set the desired switching frequency. The pin can also be used as external synchronization input. See Section 5.1.5 on page 15 for details. 5 MODE Current generators fault management selector. It allows to detect and manage LEDs failures. See Section 5.3.2 on page 21 for details. 6 AVCC + 5 V analog supply. Connect to LDO5 through a simple RC filter. 7 LDO5 + 5 V LDO output and power section supply. Bypass to SGND with a 1 μF ceramic capacitor. 8 VIN 9 SLOPE Slope compensation setting. A resistor between the output of the boost converter and this pin is needed to avoid sub-harmonic instability. Refer to Section 6.1 on page 24 for details. 10 SGND Signal ground. Supply return for the analog circuitry and the current generators. 11 ROW1 Row driver output #1. 12 ROW2 Row driver output #2. 13 ROW3 Row driver output #3. 14 ROW4 Row driver output #4. 15 ROW5 Row driver output #5. 16 ROW6 Row driver output #6. 17 PGND Power ground. Source of the internal Power MOSFET. 18 OVSEL Over-voltage selection. Used to set the desired OV threshold by an external divider. See Section 5.1.4 on page 14 for details. 19 LX 20 DIM Dimming input. Used to externally set the brightness by using a PWM signal. 21 EN Enable input. When low, the device is turned off. If tied high or left open, the device is turned on and a soft-start sequence takes place. 22 FAULT Fault signal output. Open drain output. The pin goes low when a fault condition is detected (see Section 5.3.1 on page 21 for details). 23 SYNC Synchronization output. Used as external synchronization output. 24 SS Input voltage. Connect to the main supply rail. Switching node. Drain of the internal Power MOSFET. Soft-start. Connect a capacitor to SGND to set the desired soft-start duration. LED7706 Electrical data 3 Electrical data 3.1 Maximum rating Table 3. Absolute maximum ratings (1) Symbol Parameter Value VAVCC AVCC to SGND -0.3 to 6 VLDO5 LDO5 to SGND -0.3 to 6 PGND to SGND -0.3 to 0.3 VIN VIN to PGND -0.3 to 40 VLX LX to SGND -0.3 to 40 LX to PGND -0.3 to 40 RILIM, BILIM, SYNC, OVSEL, SS to SGND V -0.3 to VAVCC + 0.3 EN, DIM, SW, MODE, FAULT to SGND -0.3 to 6 ROWx to PGND/ SGND -0.3 to 40 VIN - 0.3 to VIN + 6 SLOPE to VIN SLOPE to SGND -0.3 to 40 Internal switch maximum RMS current (flowing through LX node) PTOT Unit 2.0 A Power dissipation @ TA = 25 °C 2.3 (2) W Maximum withstanding voltage range test condition: CDF-AEC-Q100-002- “human body model” acceptance criteria: “normal performance” ±1000 V 1. Stresses beyond those listed under “absolute maximum ratings” may cause permanent damage to the device. Exposure to absolute maximum rated conditions for extended periods may affect device reliability. 2. Power dissipation referred to the device mounted on the demonstration board described in section 5.5 3.2 Thermal data Table 4. Symbol Thermal data Parameter Value Unit 42 °C/W RthJA Thermal resistance junction to ambient TSTG Storage temperature range -50 to 150 °C Junction operating temperature range -40 to 150 °C TJ 7/46 Electrical characteristics 4 LED7706 Electrical characteristics VIN = 12 V; TJ = 25 °C and LDO5 connected to AVCC if not otherwise specified (a) Table 5. Electrical characteristics Symbol Parameter Test condition Min. Typ. Max. Unit 36 V Supply section VIN Input voltage range 4.5 VBST Boost section output voltage VLDO5 LDO output and IC supply voltage EN High ILDO5 = 0 mA Operating quiescent current RRILIM = 51 kΩ, RBILIM = 220 kΩ, RSLOPE = 680 kΩ DIM tied to SGND. 1 Operating current in shutdown EN low 20 30 3.8 4.0 VAVCC IIN,Q IIN,SHDN VUVLO,ON LDO5 under voltage lock out upper threshold VUVLO,OFF LDO5 under voltage lock out lower threshold 36 4.4 5 5.5 V mA μA V 3.3 3.6 LDO linear regulator Line regulation 6 V ≤ VIN ≤ 36 V, ILDO5 = 30 mA LDO dropout voltage VIN = 4.3 V, ILDO5 = 10 mA LDO maximum output current VLDO5 > VUVLO,ON 30 25 VLDO5 < VUVLO,OFF 80 120 40 60 20 30 mV mA Boost section tON,min Minimum switching on-time fSW Default switching frequency 200 FSW connected to AVCC 570 Minimum FSW sync frequency 660 750 210 FSW sync input threshold ns kHz 240 FSW sync low level FSW sync input hysteresis 350 mV 270 ns 40 % 20 FSW sync min. ON time SYNC output Duty-Cycle FSW connected to AVCC (Internal oscillator selected) SYNC output high level ISYNC = 10 μA SYNC output low level ISYNC = -10 μA 34 VAVCC -20V mV 20 a. Specification referred to TJ from 0 °C to +85 °C. Specification over the 0 to +85 °C TJ range are assured by design, characterization and statistical correlation. 8/46 LED7706 Table 5. Electrical characteristics Electrical characteristics (continued) Symbol Parameter Test condition Min. Typ. Max. Unit 5·105 6·105 7·105 V 280 500 mΩ 1.234 1.280 Power switch KB RDSon LX current coefficient RBILIM = 600 kΩ Internal MOSFET on-resistance OC and OV protections VTH,OVP Over voltage protection reference threshold (OVSEL) 1.190 VTH,FRD Floating channel detection threshold 1.100 ΔVOVP,FRD V Voltage gap between OVP and FRD thresholds 1.145 1.190 90 mV Soft-start and power management EN, Turn-on threshold 1.6 EN, Turn-off threshold 0.8 DIM, high level threshold 1.3 V DIM, low level threshold 0.8 EN, pull-up current 2.5 SS, charge current SS, end-of-startup threshold SS, reduced switching frequency release threshold 4 5 6 2.0 2.4 2.8 μA V 0.8 Current generators section TDIM- Minimum dimming on-time 500 ns KR Current generators gain 987 V ΔKR(1) Current generators gain accuracy ON,min ±2.0 VIFB Feedback regulation voltage Vrowx, Shorted LED fault detection threshold MODE tied to SGND 3.4 MODE connected to AVCC 6.0 FAULT pin low-level voltage IFAULT,SINK = 4 mA 200 FAULT VFAULT, LOW 400 % mV V 350 mV Thermal shutdown TSHDN Thermal shutdown turn-off temperature 150 Thermal shutdown hysteresis 30 °C 1. IROW = KR / RRILIM, ΔIROW/IROW ≈ ΔKR/KR+ ΔRRILIM/RRILIM 9/46 Operation description 5 LED7706 Operation description The device can be divided into two sections: the boost section and the backlight driver section. These sections are described in the next paragraphs. Figure 3 provides an overview of the internal blocks of the device. Figure 3. Simplified block diagram VIN SLOPE Current Sense +5V LDO + + UVLO Detector UVLO + 0.7V gm _ BILIM Boost Control Logic _ COMP LX ZCD + LDO5 Ramp Generator PGND Boost_EN Current Limit _ OVP SS VROW6 Prot_EN Current Generator 6 CTRL6 ÷2 VROW5 Current Generator 5 CTRL5 Ext Sync Detector VROW4 OSC Current Generator 4 CTRL4 VROW3 FSW Current Generator 3 CTRL3 VROW2 Prot_EN EN MODE Current Generator 2 CTRL2 Boost_EN AVCC OVSEL Min Voltage Selector Soft Start SYNC 1.172V + ROW6 ROW5 ROW4 ROW3 ROW2 UVLO CONTROL LOGIC CTRL6 CTRL5 CTRL4 CTRL3 CTRL2 MODE CTRL1 LOGIC VROW1 OVP FAULT VTH,FLT 4V ROW1 I to V + _ DIM I to V Thermal Shutdown Current Generator 1 1.2V RILIM 10/46 SGND AM00596v1 LED7706 Operation description 5.1 Boost section 5.1.1 Functional description The LED7706 is a monolithic LEDs driver for the backlight of LCD panels and it consists of a boost converter and six PWM-dimmable current generators. The boost section is based on a constant switching frequency, peak current-mode architecture. The boost output voltage is controlled such that the lowest row's voltage, referred to SGND, is equal to an internal reference voltage (400 mV typ. see Figure 5). The input voltage range is from 4.5 V up to 36 V. In addition, the LED7706 has an internal LDO that supplies the internal circuitry of the device and is capable to deliver up to 40 mA. The input of the LDO is the VIN pin. The LDO5 pin is the LDO output and the supply for the power MOSFET driver at the same time. The AVCC pin is the supply for the analog circuitry and should be connected to the LDO output through a simple RC filter in order to improve the noise rejection. Figure 4. AVCC filtering VIN LDO5 LDO Rfilt 4R7 AVCC CLDO5 1uF LED7706 CAVCC 100nF SGND AM00597v1 Two loops are involved in regulating the current sunk by the generators. The main loop is related to the boost regulator and uses a constant frequency peak currentmode architecture to regulate the power rail that supplies the LEDs (Figure 5), while an internal current loop regulates the same current (flowing through the LEDs) at each row according to the set value (RILIM pin). Figure 5. Main loop and current loop diagram VIN ROWx LX Slope SGND PWM COMP E/A Minimum voltage drop selector RILIM 0.4V Error amplifier AM00598v1 11/46 Operation description LED7706 A dedicated circuit automatically selects the lowest voltage drop among all the rows and provides this voltage to the main loop that, in turn, regulates the output voltage. In fact, once the reference generator has been detected, the error amplifier compares its voltage drop to the internal reference voltage and varies the COMP output. The voltage at the COMP pin determines the inductor peak current at each switching cycle. The output voltage of the boost regulator is thus determined by the total forward voltage of the LEDs strings (see Figure 6): Equation 1 NROWS mLEDS VOUT = max ( i=1 Σ VF,j ) + 400mV j=1 where the first term represents the highest total forward voltage drop over N active rows and the second is the voltage drop across the leading generator (400 mV typ.). The device continues to monitor the voltage drop across all the rows and automatically switches to the current generator having the lowest voltage drop. Figure 6. Calculation of the output voltage of the boost regulator Row with the highest voltage drop across LEDs VIN max Σ VF Boost controller VBOOST ILED Current generators section 400 mV Leading generator AM00599v1 8 1 5.1.2 Enable function The LED7706 is enabled by the EN pin. This pin is active high and, when forced to SGND, the device is turned off. This pin is connected to a permanently active 2.5 μA current source; when sudden device turn-on at power-up is required, this pin must be left floating or connected to a delay capacitor. When turned off, the LED7706 quickly discharges the soft-start capacitor and turns off the power MOSFET, the current generators and the LDO. The power consumption is thus reduced to 20 μA only. In applications where the dimming signal is used to turn on and off the device, the EN pin can be connected to the DIM pin as shown in Figure 7. 12/46 LED7706 Operation description Figure 7. Enable pin driven by dimming signal DIM BAS69 EN 220kΩ LED7706 100nF SGND AM00600v1 5.1.3 Soft-start The soft-start function is required to perform a correct start-up of the system, controlling the inrush current required to charge the output capacitor and to avoid output voltage overshoot. The soft-start duration is set connecting an external capacitor between the SS pin and ground. This capacitor is charged with a 5 μA (typ.) constant current, forcing the voltage on the SS pin to ramp up. When this voltage increases from zero to nearly 1.2 V, the current limit of the power MOSFET is proportionally released from zero to its final value. However, because of the limited minimum on-time of the switching section, the inductor might saturate due to current runaway. To solve this problem the switching frequency is reduced to one half of the nominal value at the beginning of the soft-start phase. The nominal switching frequency is restored after the SS pin voltage has crossed 0.8 V. Figure 8. Soft-start sequence waveforms in case of floating rows OVP Floating ROWs detection 95% of OVP Output voltage SS pin voltage AVCC 2.4V 1.2V 0.8V Protections turn active tss 100% Nominal switching frequency release Current limit EN pin voltage t AM00601v1 During the soft-start phase the floating rows detection is also performed. In presence of one or more floating rows, the voltage across the involved current generator drops to zero. This voltage becomes the inverting input of the error amplifier through the minimum voltage drop 13/46 Operation description LED7706 selector (see Figure 5). As a consequence the error amplifier is unbalanced and the loop reacts by increasing the output voltage. When it reaches the floating row detection (FRD) threshold (95% of the OVP threshold), the floating rows are managed according to Table 6 (see Section 5.3 on page 21). After the SS voltage reaches a 2.4 V threshold, the start-up finishes and all the protections turn active. The soft-start capacitor CSS can be calculated according to equation 2. Equation 2 C SS ≅ ISS ⋅ t SS 2 .4 Where ISS = 5 µA and tSS is the desired soft-start duration. 5.1.4 Overvoltage protection An adjustable over-voltage protection is available. It can be set feeding the OVSEL pin with a partition of the output voltage. The voltage of the central tap of the divider is thus compared to a fixed 1.234 V threshold. When the voltage on the OVSEL pin exceeds the OV threshold, the FAULT pin is tied low and the device is turned off; this condition is latched and the LED7706 is restarted by toggling the EN pin or by performing a Power-On Reset (the POR occurs when the LDO output falls below the lower UVLO threshold and subsequently crosses the upper UVLO threshold during the rising phase of the input voltage). Normally, the value of the high-side resistors of the divider must be chosen as high as possible (but lower than 1 MΩ) to reduce the output capacitor discharge when the boost converter is off (during the off phase of the dimming cycle). The R2/R1 ratio is calculated to trigger the OVP circuitry as soon as the output voltage is 2 V higher than the maximum value for a given LED string (see equation 3). Two additional filtering capacitors, C10 and C13, may be required to improve noise rejection at the OVSEL pin, as shown in Figure 9. The typical value for C10 is in the 100 pF-330 pF range, while the C13 value is given by equation 4. Equation 3 R 2 = R1 ⋅ 1.234 V (VOUT,OVP + 2V − 1.234 V) Equation 4 C13 = 2 ⋅ C10 14/46 R2 R1 LED7706 Operation description Figure 9. OVP threshold setting VIN VOUT C13 LX R1 COUT OVSEL LED7706 R2 C10 SGND AM00602v1 5.1.5 Switching frequency selection and synchronization The switching frequency of the boost converter can be set in the 250 kHz-1 MHz range by connecting the FSW pin to ground through a resistor. Calculation of the setting resistor is made using equation 5 and should not exceed the 100 kΩ-400 kΩ range. Equation 5 RFSW = FSW 2.5 In addition, when the FSW pin is tied to AVCC, the LED7706 uses a default 660 kHz fixed switching frequency, allowing to save a resistor in minimum component-count applications. Figure 10. Multiple device synchronization SLAVE MASTER AVCC Sync Out FSW SYNC LED7706 RFSW SGND FSW SYNC SYNC LED7706 SGND AM00603v1 The FSW pin can also be used as synchronization input, allowing the LED7706 to operate both as master or slave device. If a clock signal with a 210 kHz minimum frequency is applied to this pin, the device locks synchronized. The signal provided to the FSW pin must cross the 270 mV threshold in order to be recognized. The minimum pulse width which allows the synchronizing pulses to be detected is 270 ns. An Internal time-out allows synchronization as long as the external clock frequency is greater than 210 kHz. 15/46 Operation description LED7706 Keeping the FSW pin voltage lower than 270 mV for more than 4.8 µs results in a stop of the device switching activity. Normal operation is resumed as soon as FSW rises above the mentioned threshold and the soft-start sequence is repeated. The SYNC pin is a synchronization output and provides a 35 % (typ.) duty-cycle clock when the LED7706 is used as master or a replica of the FSW pin when used as slave. It is used to connect multiple devices in a daisy-chain configuration or to synchronize other switching converters running in the system with the LED7706 (master operation). When an external synchronization clock is applied to the FSW pin, the internal oscillator is over-driven: each switching cycle begins at the rising edge of the clock, while the slope compensation (Figure 11) ramp starts at the falling edge of the same signal. Thus, to prevent subharmonic instability (see Section 5.1.6), the external synchronization clock is required to have a 40 % maximum duty-cycle when the boost converter is working in continuousconduction mode (CCM) in order to assure that the slope compensation is effective (starts with duty-cycle lower than 40%) Figure 11. External sync waveforms 270ns minimum FSW pin voltage (ext. sync) 270mV threshold Slave SYNC pin voltage Slave LX pin voltage AM00604v1 5.1.6 Slope compensation The constant frequency, peak current-mode topology has the advantage of very easy loop compensation with output ceramic caps (reduced cost and size of the application) and fast transient response. In addition, the intrinsic peak-current measurement simplifies the current limit protection, avoiding undesired saturation of the inductor. On the other side, this topology has a drawback: there is an inherent open loop instability when operating with a duty-ratio greater than 0.5. This phenomenon is known as “SubHarmonic Instability” and can be avoided by adding an external ramp to the one coming from the sensed current. This compensating technique, based on the additional ramp, is called “slope compensation”. In Figure 12, where the switching duty-cycle is higher than 0.5, the small perturbation ΔIL dies away in subsequent cycles thanks to the slope compensation and the system reverts to a stable situation. 16/46 LED7706 Operation description The SLOPE pin allows to properly set the amount of slope compensation connecting a simple resistor RSLOPE between the SLOPE pin and the output. The compensation ramp starts at 35% (typ.) of each switching period and its slope is given by the following equation: Equation 6 ⎛V − VIN − VBE SE = K S ⎜⎜ OUT R SLOPE ⎝ ⎞ ⎟ ⎟ ⎠ Where KS = 5.8 ⋅1010 s-1, VBE = 2 V (typ.) and SE is the slope ramp in [A/s]. To avoid sub-harmonic instability, the compensating slope should be at least half the slope of the inductor current during the off-phase when the duty-cycle is greater than 50%. The value of RSLOPE can be calculated according to equation 7. Equation 7 RSLOPE ≤ 2 ⋅ K S ⋅ L ⋅ (VOUT − VIN − VBE ) (VOUT − VIN ) Figure 12. Effect of slope compensation on small inductor current perturbation (D > 0.5) Inductor current (CCM) Programmed inductor peak current with slope compensation (SE) 0.35·TSW IBOOST, PEAK ΔIL Inductor current perturbation TSW t AM00606v1 17/46 Operation description 5.1.7 LED7706 Boost current limit The design of the external components, especially the inductor and the flywheel diode, must be optimized in terms of size relying on the programmable peak current limit. The LED7706 improves the reliability of the final application giving the way to limit the maximum current flowing into the critical components. A simple resistor connected between the BILIM pin and ground sets the desired value. The voltage at the BILIM pin is internally fixed to 1.23 V and the current limit is proportional to the current flowing through the setting resistor, according to the following equation: Equation 8 IBOOST,PEAK = KB R BILIM where K B = 6 ⋅ 10 5 V The maximum allowed current limit is 5 A, resulting in a minimum setting resistor RBILIM > 120 kΩ. The maximum guaranteed RMS current in the power switch is 2 A. In a boost converter the RMS current through the internal MOSFET depends on both the input and output voltages, according to equations 9a (DCM) and 9b (CCM). The current limitation works by clamping the COMP pin voltage proportionally to RBILIM. Peak inductor current is limited to the above threshold decreased by the slope compensation contribution. Equation 9 a IMOS,rms = VIN ⋅ D D FSW ⋅ L 3 Equation 9 b IMOS,rms = IOUT 5.1.8 2 ⎞ ⎛ D ⎞ VOUT 1⎛ ⎜ ⎟ (D(1 − D))3 ⎟ ⎜ + ⎟ ⎟ ⎜ (1 − D)2 12 ⎜ I ⎝ OUT ⋅ fSW ⋅ L ⎠ ⎠ ⎝ Thermal protection In order to avoid damage due to high junction temperature, a thermal shutdown protection is implemented. When the junction temperature rises above 150 °C (typ.), the device turns off both the control logic and the boost converter and holds the FAULT pin low. The LDO is kept alive and normal operation is automatically resumed after the junction temperature has been reduced by 30 °C. 18/46 LED7706 Operation description 5.2 Backlight driver section 5.2.1 Current generators The LED7706 is a LEDs driver with six channels (rows); each row is able to drive multiple LEDs in series (max. 36 V) and to sink up to 30 mA maximum current, allowing to manage different kinds of LEDs. The LEDs current can be set by connecting an external resistor (RRILIM) between the RILIM pin and ground. The voltage across the RILIM pin is internally set to 1.23 V and the rows current is proportional to the RILIM current according to the following equation: Equation 10 IROWx = KR R RILIM Where KR = 987 V. The graph in Figure 13 better shows the relationship between IROW and RRILIM and helps to choose the correct value of the resistor to set the desired row current. Figure 13. Row current vs RRILIM Row current vs RRILIM 35 Row current [mA] 30 25 20 15 10 5 0 0 50 100 150 200 250 RRILIM [kΩ] AM00605v1 The maximum current mismatch between the rows is ± 2 % @ Irowx = 20 mA. 19/46 Operation description 5.2.2 LED7706 PWM dimming The brightness control of the LEDs is performed by a pulse-width modulation of the rows current. When a PWM signal is applied to the DIM pin, the current generators are turned on and off mirroring the DIM pin behavior. Actually, the minimum dimming duty-cycle depends on the dimming frequency. The real limit to the PWM dimming is the minimum on-time that can be managed for the current generators; this minimum on-time is approximately 500 ns. Thus, the minimum dimming duty-cycle depends on the dimming frequency according to the following formula: Equation 11 DDIM,min = 500ns ⋅ fDIM For example, at a dimming frequency of 20 kHz, 1% of dimming duty-cycle can be managed. During the off-phase of the PWM signal the boost converter is paused and the current generators are turned off. The output voltage can be considered almost constant because of the relatively slow discharge of the output capacitor. During the start-up sequence (see Section 5.1.3 on page 13) the dimming duty-cycle is forced to 100% to detect floating rows regardless of the applied dimming signal. Figure 14. PWM dimming waveforms 20/46 LED7706 5.3 Operation description Fault management The main loop keeps the row having the lowest voltage drop regulated to about 400 mV. This value slightly depends on the voltage across the remaining active rows. After the softstart sequence, all protections turn active and the voltage across the active current generators is monitored to detect shorted LEDs. 5.3.1 FAULT pin The FAULT pin is an open-collector output, (with 4 mA current capability) active low, which gives information regarding faulty conditions eventually detected. This pin can be used either to drive a status LED or to warn the host system. The FAULT pin status is strictly related to the MODE pin setting (see Table 6 for details). 5.3.2 MODE pin The MODE pin is a digital input and can be connected to AVCC or SGND in order to choose the desired fault detection and management. The LED7706 can manage a faulty condition in two different ways, according to the application needs. Table 6 summarizes how the device detects and handles the internal protections related to the boost section (overcurrent, over-temperature and over-voltage) and to the current generators section (open and shorted LEDs). Table 6. Faults management summary FAULT MODE to GND MODE to VCC Internal MOSFET over current FAULT pin HIGH Power MOS turned OFF Output over voltage FAULT pin LOW Device turned OFF, latched condition Thermal shutdown FAULT pin LOW. Device turned OFF. Automatic restart after 30 C temperature drop. LED short shorted led FAULT pin LOW Device turned OFF, latched condition (Vth = 3.4 V) FAULT pin LOW Faulty row(s) disconnected Device keeps on working with the remaining row(s) (Vth = 6 V) Open row(s) FAULT pin LOW Device turned OFF at first occurrence, latched condition FAULT pin HIGH Faulty row(s) disconnected. Device keeps on working with the remaining row(s) 21/46 Operation description 5.3.3 LED7706 Open LED fault In case a row is not connected or a LED fails open, the device has two different behaviors according to the MODE pin status. Connecting the MODE pin to SGND, as soon as an open row is detected the FAULT pin is tied low and the device is turned off. The internal logic latches this status: to restore the normal operation, the device must be restarted by toggling the EN pin or performing a Power-On Reset (POR occurs when the voltage at the LDO5 pin falls below the lower UVLO threshold and subsequently rises above the upper one). If the MODE pin is high (i.e. connected to AVCC), the LED7706 behaves in a different manner: the open row is excluded from the control loop and the device continues to work properly with the remaining rows. The FAULT pin is not affected. Thus, if less than six rows are used in the application, the MODE pin must be set high. Figure 15 shows an example of open channel detection in case of MODE connected to AVCC. At the point marked as “1” in Figure 15, the row opens (row current drops to zero). From this point on the output voltage is increased as long as the output voltage reaches the floating row detection threshold (see Section 5.1.3 on page 13). Then (point marked as “2”) the faulty row is disconnected and the device keeps on working only with the remaining rows. Figure 15. Open channel detection (MODE to AVCC)   2 1 22/46 LED7706 5.3.4 Operation description Shorted LED fault When a LED is shorted, the voltage across the related current generator increases of an amount equal to the missing voltage drop of the faulty LED. Since the feedback voltage on each active generator is constantly compared with a fault threshold VTH,FAULT, the device detects the faulty condition and acts according to the MODE pin status. If the MODE pin is low, the fault threshold is VTH,FAULT = 3.4 V. When the voltage across a row is higher than this threshold, the FAULT pin is set low and the device is turned off. The internal logic latches this status until the EN pin is toggled or a POR is performed. In case the MODE pin is connected to AVCC, the fault threshold is set to 6 V. The LED7706 simply disconnects the rows whose voltage is higher than the threshold and the FAULT pin is forced low. This option is also useful to avoid undesired triggering of the shorted-LED protection simply due to the high voltage drop spread across the LEDs. Figure 16 shows an example of shorted LED detection in case MODE is connected to GND. At the point marked as “1” in Figure 16 one LED fails becoming a short-circuit. The voltage across the current generator of the channel where the failed LED is connected increases by an amount equal to the forward voltage of the faulty LED. Since the voltage across the current generator is above the threshold (3.4 V in this case), the device is turned off and the fault pin is set low (point “2”). Figure 16. Shorted LED detection (MODE to GND)   1 2 23/46 Application information LED7706 6 Application information 6.1 System stability The boost section of the LED7706 is a fixed frequency, current-mode converter. During normal operation, a minimum voltage selection circuit compares all the voltage drops across the active current generators and provides the minimum one to the error amplifier. The output voltage of the error amplifier determines the inductor peak current in order to keep its inverting input equal to the reference voltage (400 mV typ). The compensation network consists of a simple RC series (RCOMP - CCOMP) between the COMP pin and ground. The calculation of RCOMP and CCOMP is fundamental to achieve optimal loop stability and dynamic performance of the boost converter and is strictly related to the operating conditions. 6.1.1 Loop compensation The compensation network can be quickly calculated using equations 12 to 16. Once both RCOMP and CCOMP have been determined, a fine-tuning phase may be required in order to get the optimal dynamic performance from the application. The first parameter to be fixed is the switching frequency. Normally, a high switching frequency allows reducing the size of the inductor and positively affects the dynamic response of the converter (wider bandwidth) but increases the switching losses. For most of applications, the fixed value (660 kHz) represents a good trade-off between power dissipation and dynamic response, allowing to save an external resistor at the same time. In low-profile applications, the inductor value is often kept low to reduce the number of turns; an inductor value in the 4.7 µH-15 µH range is a good starting choice. In order to avoid instability due to interaction between the DC-DC converter's loop and the current generators' loop, the bandwidth of the boost should not exceed the bandwidth of the current generators. A unity-gain frequency (fU) in the order of 30-40 kHz is acceptable. Also, take care not to exceed the CCM-mode right half-plane zero (RHPZ). Equation 12 fU ≤ 0.2 ⋅ FSW Equation 13 2 ⎛ VIN,min ⎞ ⎛ VOUT ⎜ ⎟ ⎜ ⎜ V ⎟ ⎜ 2 M R OUT ⎠ ⎝ IOUT ⎝ fU ≤ 0.2 ⋅ = 0. 2 ⋅ 2π ⋅ L 2π ⋅ L Equation 14 a M= 24/46 VIN,min VOUT ⎞ ⎟ ⎟ ⎠ LED7706 Application information Equation 14b R= VOUT IOUT Where VIN,min is the minimum input voltage and IOUT is the overall output current. Note that, the lower the inductor value (and the higher the switching frequency), the higher the bandwidth can be achieved. The output capacitor is directly involved in the loop of the boost converter and must be large enough to avoid excessive output voltage drop in case of a sudden line transition from the maximum to the minimum input voltages. However a more significant requirement concerns the output voltage ripple. The output capacitor should be chosen in accordance with the following expression: Equation 15 COUT > (IL,peak − IOUT )⋅ TOFF 2 ⋅ ΔVOUT,max where ΔVOUT, max is the maximum acceptable output voltage ripple, IL, peak is the peak inductor current, TOFF is the off-time of the switching cycle (for an extensive explanation see Section 6.4.4 on page 31). Once the output capacitor has been chosen, the RCOMP can be calculated as: Equation 16 R COMP = 2π ⋅ fU ⋅ C GM ⋅ gEA ⋅ M Where GM = 2.7 S and gEA = 375 µS Equation 16 places the loop bandwidth at fU. Then, the CCOMP capacitor is determined to place the frequency of the compensation zero 5 times lower than the loop bandwidth: Equation 17 C COMP = 1 2π ⋅ fZ ⋅ R COMP Where fZ=fU/5. In most of the applications an experimental approach is also very valid to compensate the circuit. A simple technique to optimize different applications is to choose CCOMP = 4.7 nF and to replace RCOMP with a 10 kΩ trimmer adjusting its value to properly damp the output transient response. Insufficient damping will result in excessive ringing at the output and poor phase margin. Figure 17 (a and b) give an example of compensation adjustment for a typical application. 25/46 Application information LED7706 Figure 17. Poor phase margin (a) and properly damped (b) load transient responses a) b) Figure 18. Load transient response measurement set-up 6.8μH VIN= 6V C VBOOST IN 4.7μF MLCC FSW OVSEL LX LDO5 SLOPE VIN AVCC +5V ROW1 BILIM ROW2 RILIM ROW3 LED7706 PGND MODE ROW6 SYNC SGND EN ROW5 DIM ROW4 COMP FAULT SS RL = Up to 10 WLEDs per row VBST 50mA 500Hz AM00607v1 26/46 LED7706 6.2 Application information Thermal considerations In order to prevent the device from exceeding the thermal shutdown threshold (150 °C), it is important to estimate the junction temperature through the following equation: Equation 18 TJ = TAmb + R th,JA ⋅ PD,tot where TA is the ambient temperature, Rth,JA is the equivalent thermal resistance junction to ambient and PD,tot is the power dissipated by the device. The Rth,JA measured on the application demonstration board (described in Section 6.5) is 42 °C/W. The PD,tot has several contributions, listed below. a) Conduction losses due to the RDS(on) of the internal power switch, equal to:\ Equation 19 2 PD,cond = RDSon ⋅ IIN ⋅ D ⋅ DDIM where D is defined as: Equation 20 D = 1− VIN VOUT and DDIM is the duty cycle of the PWM dimming signal. b) Switching losses due to the power MOSFET turn on and off, calculated as: Equation 21 PD,sw = VOUT ⋅ IIN ⋅ fsw ⋅ (tr + t f ) ⋅ DDIM 2 where tr and tf are the power MOSFET rise time and fall time respectively. c) Current generators losses. This contribution is strictly related to the LEDs used in the application. Only the contribution of the leading current generator (“master” current generator) can be predicted, regardless of the LEDs forward voltage: Equation 22 PGEN,Master = IROW ⋅ VIFB ⋅ DDIM where IROW is the current flowing through the row, whereas VIFB is the voltage across the master current generator (typically 400 mV). The voltages across the other current generators depend on the spread of the LEDs forward voltage. The worst case for power dissipation (maximum forward voltage LEDs in the master row, minimum forward voltage LEDs in all other rows) can be estimated as: 27/46 Application information LED7706 Equation 23 PGEN = IROW ⋅ (nROWs − 1) ⋅ (VIFB + ΔVf,LEDs ⋅ nLEDs ) ⋅ DDIM where nROWs is the number of active rows, ΔVf,LEDs is the spread of the LEDs forward voltage and nLEDs is the number of LEDs per row. d) LDO losses, due to the dissipation of the 5 V linear regulator: Equation 24 PD,LDO = (VIN − VLDO ) ⋅ ILDO The LED7706 is housed in a 24 leads 4x4-VFQFPN package with exposed pad that allows good thermal performance. However it is also important to design properly the demonstration board layout in order to assure correct heat dissipation. Figure 19 shows a picture of the LED7706 application demonstration board taken using an infrared camera. The chip temperature, in those application conditions, is kept below 50 °C. Figure 19. Demonstration board thermographic analysis VIN = 12V I ROW = 20mA VOUT = 34V FSW = 660kHz D DIM = 90% T Amb = 30°C 6.3 Component selection 6.3.1 Inductor selection Being the LED7706 mostly dedicated to backlighting, real-estate applications dictate severe constrain in selecting the optimal inductor. The inductor choice must take into account different parameters like conduction losses (DCR), core losses (ferrite or iron-powder), saturation current and magnetic-flux shielding (core shape and technology). The switching frequency of the LED7706 can be set in the 200 kHz-1 MHz range, allowing a wide selecting room for the inductance value. Low switching frequencies takes to high inductance value, resulting in significant DCR and size. On the other hand, high switching frequencies result in significant core losses. The suggested range is 4.7-22 µH, even if the 28/46 LED7706 Application information best trade off between the different loss contributions varies from manufacturer to manufacturer. A 6.8 µH inductor has been experimentally found as the most suitable for applications running at a 660 kHz switching frequency. Table 7. 6.3.2 Recommended inductors Manufacturer Part number Description Size Coilcraft LPS6235-682MLC 6.8 μH, 75 mΩ, 2.7 A 6x6 mm Wurth 7440650068 6.8 μH, 33 mΩ, 3.6 A 10x10 mm Capacitors selection The input and output capacitors should have very low ESR (ceramic capacitors) in order to minimize the ripple voltage. The boost converter of the LED7706 has been designed to support ceramic capacitors. The required capacitance depends on the programmed LED current and the minimum dimming frequency (the boost converter is off when the DIM pin is low and the output capacitor is slowly discharged). Considering the worst case (i.e. 200 Hz dimming frequency and 30 mA/channel), two 2.2 µF MLCCs are suitable for almost all applications. Particular care must be taken when selecting the rated voltage and the dielectric type of the output capacitors: 50 V rated MLCC may show a significant capacitance drop when biased, especially in case of Y5V dielectric. As in most of boost converters, the input capacitor is less critical, although it is necessary to reduce the switching noise on the supply rail. The input capacitor is also important for the internal LDO of the LED7706 and must be kept as close as possible to the chip. The rated voltage of the input capacitor can be chosen according to the supply voltage range; a 10 µF X5R MLCC is recommended. Table 8. 6.3.3 Recommended capacitors Manufacturer Part number Description Package Notes Taiyo Yuden UMK325BJ106KM-T Ceramic, 35V, X5R, 20 % SMD 1210 CIN Murata GRM31CR71H225KA88B Ceramic, 50V, X7R, 20 % SMD 1206 COUT Murata GRM31CR71H475KA88B Ceramic, 50V, X7R, 20 % SMD 1206 COUT Flywheel diode selection The flywheel diode must be a Schottky type to minimize the losses. This component is subject to an average current equal to the output one and must sustain a reverse voltage equal to the maximum output rail voltage. Considering all the channels sinking 30mA each (i.e. 180 mA output current) and the maximum output voltage (36 V), the STP1L40M (If,ave = 1 A, Vr = 40 V) diode is a good choice. Smaller diodes can be used in applications involving lower output voltage and/or lower output current. 29/46 Application information 6.4 LED7706 Design example In order to help the design of an application using the LED7706, in this section a simple step-by-step design example is provided. A possible application could be the LED backlight in a 15” LCD panel using the LED7706. Here below the possible application conditions are listed: 6.4.1 ● VIN = 12 ± 20 % ● 2 strings of 48 white LEDs (20 mA) each (arranged in 6 rows, 8 LEDs per row) ● VF, LEDs = 3.5 V ± 200 mV Switching frequency setting To reduce the number of the external components, the default switching frequency is selected (660 kHz typ.) by connecting the FSW pin to AVCC pin. However, in case a different switching frequency is required, a resistor from FSW pin and ground can be connected, according to the equation (5) in Section 5.1.5. 6.4.2 Row current setting Considering the equation 10 in Section 5.2.1, the RRILIM resistor can be calculated as: Equation 25 RRILIM = KR IROW = 987 V = 49.35kΩ 20 mA The closest standard commercial value is 51 kΩ. The actual value of the row current will be a little lower (19.3 mA). 6.4.3 Inductor choice The boost section, as all DC-DC converters, can work in CCM (continuous conduction mode) or in DCM (discontinuous conduction mode) depending on load current, input and output voltage and other parameters, among which the inductor value. In a boost converter it is usually preferable to work in DCM. Once the load, the input and output voltage, and the switching frequency are fixed, the inductor value defining the boundary between DCM and CCM operation can be calculated as: Equation 26 R 0 ⋅ D ⋅ (1 − D) 2 ⋅ FSW 2 LB = where D is the duty-cycle defined as: 30/46 LED7706 Application information Equation 27 D = 1− ⎧ 0.68 @ VIN,min = 9.6V VIN =⎨ VOUT ⎩0.52 @ VIN,max = 14.4V whereas R0 is: Equation 28 R0 = VOUT = 250Ω IOUT and Equation 29 IOUT = 6 ⋅ IROW = 120mA The output voltage in the above calculations is considered as the maximum value (LED with the maximum forward voltage connected to the leading generator): Equation 30 VOUT,max = 8 ⋅ VF,LEDs,max + 400mV = 30 V Considering the input voltage range, the lower LB will be at the lower input voltage. Hence the condition to assure the DCM operation becomes: Equation 31 L < L B (VIN,min ) = 13.2μH An inductor value of 6.8 µH could be a suitable value, considering also a margin from the boundary condition. It is important to highlight that the inductor choice involves not only the value itself but the saturation current (higher than the current limit, see Section 6.4.4), the rated RMS current (the compliance with the saturation current might be not enough; also the thermal performances must be taken into account), the DCR (which affects the efficiency) and the size (in some application might be a strict requirement). However the DCR can’t be reduced keeping the size small. Hence a trade off between these two requirements must be achieved according to the application. 6.4.4 Output capacitor choice The choice of the output capacitor is mainly affected by the desired output voltage ripple. Since the voltage across the LEDs can be considered almost constant, this ripple is transferred across the current generators, affecting their dynamic response. The output ripple can be estimated as (neglecting the contribution of ESR of COUT, very low in case of MLCC): 31/46 Application information LED7706 Equation 32 ΔVOUT = (IL,peak − IOUT )⋅ TOFF ` 2 ⋅ C OUT where IL, peak is the inductor peak current (see Figure 20) calculated as: Equation 33 IL,peak = VIN ⋅ D ⎧ 1.044 A @ VIN,min = 9.6V =⎨ Fsw ⋅ L ⎩0.914 A @ VIN,max = 14.4V whereas D, working in DCM, is: Equation 34 D= 2 ⋅ Fsw ⋅ L ⋅ M(M − 1) ⎧ 0.488 @ VIN,min = 9.6V =⎨ R0 ⎩0.285 @ VIN,max = 14.4V defining M as: Equation 35 M= VOUT VIN ⎧ 3.125 @ VIN,min = 9.6V ⎨ ⎩2.083 @ VIN,max = 14.4V Figure 20. Inductor current in DCM operation IL IL, peak TON TOFF t TSW = 1/FSW AM00608v1 TOFF can be calculated as: Equation 36 ⎧ 348.5ns @ VIN,min = 9.6V TOFF = TSW ⋅ D 2 = ⎨ ⎩398.5ns @ VIN,max = 14.4V defining D2 as: 32/46 LED7706 Application information Equation 37 D2 = 2 ⋅ FSW ⋅ L ⋅ M ⎧ 0.23 @ VIN,min = 9.6V =⎨ R 0 ⋅ (M − 1) ⎩0.263 @ VIN,max = 14.4V The worst case for the output voltage ripple is when input voltage is lower (VIN,min = 9.6 V). A simple way to select the COUT value is fixing a maximum voltage ripple. In order to affect as less as possible the current generators, it would be better to fix the maximum ripple lower than the typical voltage across the generators. For example considering ΔVOUT lower than 80 mV (i.e. the 20 % of the voltage across the leading generator), the required capacitance is: Equation 38 COUT > (IL,peak − IOUT )⋅ TOFF 2 ⋅ ΔVOUT,max = 2.02μF A margin from the calculated value should be taken into account because of the capacitance drop due to the applied voltage when MLCCs are used. A 4.7 µF MLCC can be a good choice for this application (two 2.2 μF MLCC in parallel can be also a good solution). In case a dimming duty cycle different from 100 % is used, a further contribution to the capacitor discharge (during the off time of the dimming cycle) should be considered. 6.4.5 Input capacitor choice The input capacitor of a boost converter is less critical than the output capacitor, due to the fact that the inductor is in series with the input, and hence, the input current waveform is continuous. A low ESR capacitor is always recommended. A capacitor of 10 µF is tentatively a good choice for most of the applications. 6.4.6 Over-voltage protection divider setting The over-voltage protection (OVP) divider provides a partition of the output voltage to the OVSEL pin. The OVP divider setting not only fixes the OVP threshold, but also the openchannel detection threshold (95 % of the OVP threshold). The proper OVP divider setting can be calculated by the equation (3): Equation 39 R 2 = R1 ⋅ 1.234 V (VOUT,MAX + 2V − 1.234 V) where VOUT, MAX is the maximum output voltage considering the worst case (all LEDs with the maximum VF = VF,max = 3.7 V on the same row): 33/46 Application information LED7706 Equation 40 VOUT,OVP = nLED ⋅ VF,max + 400mV = 30 V R1 can be chosen is in the order of hundreds of kilo-ohms to reduce the leakage current in the resistor divider. For example, setting R1 = 510 kΩ leads to R2 = 21.89 kΩ. The closest standard commercial value is R2 = 22 kΩ. The correct selection of the OVP divider (and, as consequence, the open channel detection threshold) must take into account the shorted LED fault threshold (see Section 5.3.4). When the selected OVP threshold is too high, an undesired triggering of the shorted LED detection circuitry may occur when a channel opens. In fact, as explained in section 4.3.3, once a channel opens, the device reacts increasing the output voltage and the voltage across the working current generators may reach the shorted LED detection threshold before the open channel detection threshold. 6.4.7 Compensation network For the compensation network, the suggestions provided in Section 6.1 are always valid. In this condition, tentatively the following value of R3 and C8 (see Figure 23) are usually a good choice for the loop stability: R3 = 2.4 kΩ C8 = 4.7 nF 6.4.8 Boost current limit The boost current limit is set to protect the internal power switch against excessive current. The slope compensation may reduce the programmed current limit. Hence, to take into account this effect, as a rule of thumb, the current limit can be set as twice as much the maximum inductor peak current (see Section 6.4.4): IBOOST, PEAK > 2.09 A Therefore, using equation (8) and choosing IBOOST, PEAK = 2.5 A, RBILIM will be: Equation 41 RBILIM = 34/46 KB IBOOST,PEAK = 240kΩ LED7706 6.4.9 Application information Power dissipation estimate As explained in Section 6.2, there are several contributions to the total power dissipation. Neglecting the power dissipated by the LDO (surely less significant compared with the other contributions), equation (19), (21), (22) and (23) help to estimate the overall power dissipation. Before starting the power dissipation estimate it is important to highlight that the following calculations are considering the worst case (the actual value of the dissipated power would require measurements). Therefore the power dissipation is estimated according to the following assumptions: 1. Minimum input voltage (9.6 V), which leads to maximum input current (and also D will have the higher value, see Section 6.4.4); 2. Maximum RDS(on) of the internal power MOSFET; 3. LEDs in the row of the leading generator will have the maximum forward voltage, whereas all other LEDs in the other rows will have the minimum forward voltage. 4. 100 % dimming signal duty cycle is considered. The conduction and switching losses on the internal power switch can be calculated as: Equation 42 2 PD,cond = RDSon ⋅ IIN ⋅ D ⋅ DDIM = 34mW Equation 43 PD,sw = VOUT ⋅ IIN ⋅ fsw ⋅ (tr + t f ) ⋅ DDIM = 111mW 2 where tr = tf = 15 ns The power dissipation related to the current generators is given by: Equation 44 PGEN,Master = IROW ⋅ VIFB ⋅ DDIM = 8mW Equation 45 PGEN = IROW ⋅ (nROWs − 1) ⋅ (VIFB + ΔVf,LEDs ⋅ nLEDs ) ⋅ DDIM = 360mW Equation 46 PD,tot ≅ PD,cond + PD,sw + PGEN,Master + PGEN = 513mW The junction temperature can be estimated by equation (18) considering TA = 25 °C: Equation 47 TJ = TAmb + R th,JA ⋅ PD,tot = 46.5°C 35/46 Application information LED7706 In order to estimate also the efficiency, other contributions to the power dissipation must be added to PD, tot (which represents only the power dissipated by the device), that is: Equation 48 PDISS,Diode = VF,Diode ⋅ IIN ⋅ D 2 = 34.5mW where VF, Diode = 0.4 V Equation 49 2 2 PDISS,Ind = DCR ⋅ IInd ,RMS ≅ DCR ⋅ IIN = 11.2mW where DCR = 80 mΩ (typical DCR of the recommended inductors). Therefore the total dissipated power is: Equation 50 PDISS,TOTAL = PD,tot + PDISS,Diode + PDISS,Ind = 559.1mW Considering the input power as the result of input voltage multiplied by the input current, the estimated efficiency is: Equation 51 η= Note: 36/46 PIN − PDISS,TOT PIN = 0.845 It is important to remind that the previous calculations consider the worst case, especially for the power dissipated on the current generators. Statistical analysis (confirmed by bench measurements) shows that the series connection of more LEDs on each channel leads to compensation effects. The hypothesis 3 above mentioned is thus rather unlikely. Therefore PGEN is significantly lower and the overall efficiency is typically around 90 %. LED7706 6.5 Application information Layout consideration 1. A careful PCB layout is important for proper operation. In this section some guidelines are provided in order to achieve a good layout. 2. The device has two different ground pins: signal ground (SGND) and power ground (PGND). The PGND pin handles the switching current related to the boost section; for this reason the PCB traces should be kept as short as possible and with adequate width. 3. The signal ground is the return for the device supply and the current generators and can be connected to the thermal pad. 4. The heat dissipation area (adequate to the application conditions) should be placed backside respect to the device and with the lowest thermal impedance possible (i.e. PCB traces in the backside should be avoided). The dissipation area is thermally and electrically connected to the thermal pad by several vias (nine vias are recommended). 5. The signal and power grounds must be connected together in a single point as close as possible to the PGND pin to reduce ground loops. 6. The R-C components of the compensation network should be placed as close as possible to the COMP pin in order to avoid noise issue and instability of the compensation. 7. Noise sensitive signals (i.e. feedbacks and compensation) should be routed as short as possible to minimize noise collection. The LED7706 pinout makes it easy to separate power components (e.g. inductor, diode) from signal ones. 8. The LX switching node should have and adequate width for high efficiency. 9. The critical power path inductor-LX-PGND must be as short as possible by mounting the inductor, the diode and COUT as close as possible each other. 10. The capacitors of the compensated divider connected to the OVSEL pin should be placed as close as possible to the OVSEL pin. 11. In order to assure good performance in terms of row current accuracy/mismatch, the PCB traces from the rows pins to the LEDs should have similar length and width. 12. The capacitors of the filter connected to LDO5 and VIN pins should be mounted as close as possible to the mentioned pins Figure 21 and Figure 22 shows the demonstration board layout (top view and bottom view respectively). 37/46 Application information LED7706 Figure 21. Demonstration board layout (top view)   Figure 22. Demonstration board layout (bottom view) Figure 23 shows the LED7706 demonstration board application circuit, whereas Table 9 lists the used components and their value. 38/46 LED7706 Application information Figure 23. LED7706 demonstration board schematic AM00609v1 39/46 Application information Table 9. LED7706 LED7706 demonstration board component list Component Description Package Part number MFR Value C1 Ceramic, 35V X5R, 20% SMD 1210 UMK325BJ106KM-T Taiyo Yuden 10 µF SMD 1206 GRM31CR71H225KA88B C2,C3 C4 Ceramic, 50V X7R, 20% 2.2 µF Murata SMD 1206 GRM31CR71H225KA88B N.M. C5 1 µF C6 100 nF C7 3.3 nF C8 C9 4.7 nF Ceramic, 25V X5R, 20% SMD 0603 Standard N.M. C10 220 pF C11 4.7 nF C12 N.M. C13 15 pF R1 510 kΩ R2 16 kΩ R3 R4 Chip resistor 0.1 W, 1% 2.4 kΩ SMD 0603 Standard 4.7 Ω R5 330 kΩ R6 51 kΩ R7 180 kΩ R8 680 kΩ R9, R10 R11 Chip resistor 0.1 W, 1% 100 kΩ SMD 0603 Standard 1.2 kΩ R12 N.M. R13 N.M. L1 6u8, 75 mΩ, 2.7 A 6x6 mm LPS6235-682MLC Coilcraft 6.8 µF D1 Schottky, 40 V, 1 A DO216-AA STPS1L40M ST STPS1L40M D2 Red LED, 3 mA SMD 0603 Standard D3 Signal Schottky SOD-523 BAS69 N.M. U1 Integrated circuit QFN4x4 ST LED7706 J2 PCB pad jumper J8 Header 8 SIL 8 Standard SW1, SW2 Jumper 3 SIL 3 Standard SW3 Push button 6x6 mm 40/46 LED7706 FSM4JSMAT TYCO LED7706 7 Electrical characteristics Electrical characteristics Figure 24. Efficiency versus DIM duty cycle, VIN = 6 V, 6 rows, 10 white LEDs (20 mA) in series Figure 25. Efficiency versus DIM duty cycle, VIN = 12 V, 6 rows, 10 white LEDs (20 mA) in series Figure 26. Efficiency versus DIM duty cycle, VIN = 18 V, 6 rows, 10 white LEDs (20 mA) in series Figure 27. Efficiency versus DIM duty cycle, VIN = 24 V, 6 rows, 10 white LEDs (20 mA) in series Figure 28. Soft-start waveforms (EN, SS, and VOUT monitored) Figure 29. Boost section switching signals (LX, SYNC and inductor current monitored), VIN = 12 V, 10 LEDs 41/46 Electrical characteristics Figure 30. Dimming waveforms (FDIM = 200 Hz) LED7706 Figure 31. Dimming waveforms (FDIM = 1 kHz) Figure 32. Dimming waveforms (FDIM = 5 kHz) Figure 33. Dimming waveforms (FDIM = 20 kHz) Figure 34. Dimming waveforms (TDIM,ON = 10 µs) 42/46 Figure 35. Dimming waveforms with the minimum dimming on time (500 ns) LED7706 8 Package mechanical data Package mechanical data In order to meet environmental requirements, ST offers these devices in different grades of ECOPACK® packages, depending on their level of environmental compliance. ECOPACK® specifications, grade definitions and product status are available at: www.st.com. ECOPACK® is an ST trademark. 43/46 Package mechanical data Table 10. LED7706 VFQFPN-24 4 mm x 4 mm mechanical data mm Dim. Min. Typ. Max. A 0.80 0.90 1.00 A1 0.00 0.02 0.05 A3 0.20 b 0.18 0.25 0.30 D 3.85 4.00 4.15 D2 2.40 2.50 2.60 E 3.85 4.00 4.15 E2 2.40 2.50 2.60 e L 0.50 0.30 ddd Figure 36. Package dimensions 44/46 0.40 0.50 0.08 LED7706 9 Revision history Revision history Table 11. Document revision history Date Revision 08-Feb-2008 1 Initial release 2 Added: Chapter 3 on page 7, Chapter 3 on page 7 and Chapter 3 on page 7 Updated: Chapter 3 on page 7, Chapter 3 on page 7, Chapter 3 on page 7, Table 4, Table 5, Figure 3, Figure 4, Figure 7, Figure 23 and Table 9 09-Apr-2009 Changes 45/46 LED7706 Please Read Carefully: Information in this document is provided solely in connection with ST products. STMicroelectronics NV and its subsidiaries (“ST”) reserve the right to make changes, corrections, modifications or improvements, to this document, and the products and services described herein at any time, without notice. All ST products are sold pursuant to ST’s terms and conditions of sale. Purchasers are solely responsible for the choice, selection and use of the ST products and services described herein, and ST assumes no liability whatsoever relating to the choice, selection or use of the ST products and services described herein. No license, express or implied, by estoppel or otherwise, to any intellectual property rights is granted under this document. If any part of this document refers to any third party products or services it shall not be deemed a license grant by ST for the use of such third party products or services, or any intellectual property contained therein or considered as a warranty covering the use in any manner whatsoever of such third party products or services or any intellectual property contained therein. UNLESS OTHERWISE SET FORTH IN ST’S TERMS AND CONDITIONS OF SALE ST DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY WITH RESPECT TO THE USE AND/OR SALE OF ST PRODUCTS INCLUDING WITHOUT LIMITATION IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE (AND THEIR EQUIVALENTS UNDER THE LAWS OF ANY JURISDICTION), OR INFRINGEMENT OF ANY PATENT, COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT. UNLESS EXPRESSLY APPROVED IN WRITING BY AN AUTHORIZED ST REPRESENTATIVE, ST PRODUCTS ARE NOT RECOMMENDED, AUTHORIZED OR WARRANTED FOR USE IN MILITARY, AIR CRAFT, SPACE, LIFE SAVING, OR LIFE SUSTAINING APPLICATIONS, NOR IN PRODUCTS OR SYSTEMS WHERE FAILURE OR MALFUNCTION MAY RESULT IN PERSONAL INJURY, DEATH, OR SEVERE PROPERTY OR ENVIRONMENTAL DAMAGE. ST PRODUCTS WHICH ARE NOT SPECIFIED AS "AUTOMOTIVE GRADE" MAY ONLY BE USED IN AUTOMOTIVE APPLICATIONS AT USER’S OWN RISK. Resale of ST products with provisions different from the statements and/or technical features set forth in this document shall immediately void any warranty granted by ST for the ST product or service described herein and shall not create or extend in any manner whatsoever, any liability of ST. ST and the ST logo are trademarks or registered trademarks of ST in various countries. Information in this document supersedes and replaces all information previously supplied. The ST logo is a registered trademark of STMicroelectronics. All other names are the property of their respective owners. © 2009 STMicroelectronics - All rights reserved STMicroelectronics group of companies Australia - Belgium - Brazil - Canada - China - Czech Republic - Finland - France - Germany - Hong Kong - India - Israel - Italy - Japan Malaysia - Malta - Morocco - Singapore - Spain - Sweden - Switzerland - United Kingdom - United States of America www.st.com 46/46