NCV8851 Automotive Grade Synchronous Buck Controller The NCV8851 is an adjustable output, synchronous buck controller, which drives dual N−channel MOSFETs, ideal for high power applications. Average current mode control is employed for very fast transient response and tight regulation over wide input voltage and output load ranges. The IC incorporates an internal fixed 6.0 V low−dropout linear regulator (LDO), which supplies charge to the switch mode power supply’s (SMPS) bottom gate driver, limiting the power lost to excess gate drive. The IC is designed for operation over a wide input voltage range (4.5 V to 20 V) and is capable of 10 to 1 voltage conversion at 500 kHz. Additional controller features include undervoltage lockout, internal soft start, low quiescent current sleep mode, programmable frequency, SYNC function, average current limiting, cycle−by−cycle overcurrent protection and thermal shutdown. http://onsemi.com TSSOP−20 SUFFIX DB CASE 948E MARKING DIAGRAM Features • • • • • • • • • • • • • • NCV 8851 ALYWG G Average Current Mode Control 0.8 V ±2% Reference Voltage Input Voltage Range of 4.5 V to 20 V 6.0 V Low−dropout Linear Regulator (LDO) Input UVLO (Undervoltage Lockout) Internal Soft−start 1.0 mA Maximum Quiescent Current in Sleep Mode Adaptive Non−overlap Circuitry 150 ns Minimum High−side Gate Off−time Programmable Fixed Frequency – 170 kHz to 500 kHz External Clock Synchronization up to 600 kHz Average Current Limiting (ACL) Cycle−by−Cycle Overcurrent Protection (OCP) Thermal Shutdown (TSD) NCV8851 = Specific Device Code A = Assembly Location L = Wafer Lot Y = Year W = Work Week G = Pb−Free Package (Note: Microdot may be in either location) ORDERING INFORMATION Applications • Automotive Systems Requiring High Current • Pre−regulated Supply for Low−voltage SMPSs and LDOs © Semiconductor Components Industries, LLC, 2010 April, 2010 − Rev. 2 Package Shipping† NCV8851DBG TSSOP−20 (Pb−Free) 75 Units/Rail NCV8851DBR2G TSSOP−20 (Pb−Free) 2500/Tape & Reel Device †For information on tape and reel specifications, including part orientation and tape sizes, please refer to our Tape and Reel Packaging Specification Brochure, BRD8011/D. 1 Publication Order Number: NCV8851/D NCV8851 VIN 12 EN 11 VIN_IC 3 SYNC 1 ROSC 20 Enable TSD Fault UVLO Logic LDO Enable Fault Soft Start VSS Fixed−Frequency Oscillator Ramp Clock Max Duty VREF Min On Time R Q Reset Dominant PWM Fault CCOMP 15 5 GH 6 VSW 7 GL 8 PGND 2 VIN_CS BST + 19 CSP CSA CFB 16 CSOUT 17 + VEA VFB 13 BST 6VOUT VOCP VCOMP 14 4 Nonoverlap OCP CEA 6VOUT + Q S 9 LDO ILIMIT VREF VACL + VSS 18 CSN ACL 10 AGND VCLAMP Figure 1. Functional Block Diagram VIN VIN EN VIN_IC SYNC ROSC 12 9 11 3 4 1 5 20 6 7 8 CC1 RC1 CFB CC2 RC2 CSOUT + DBST ROSC CCOMP 6VOUT 19 18 13 16 17 14 10 VIN Q1 GH − CBST VSW + L Q2 1W GL RS C VIN_CS − CSP CSN RF1 VFB CV2 VCOMP RV1 CV1 AGND Figure 2. Application Schematic Note: This part is recommended for synchronous use only. http://onsemi.com 2 + VOUT 4300 pF PGND 2 15 + BST RF0 NCV8851 PACKAGE PIN DESCRIPTIONS − 20 Lead TSSOP Package Pin# Pin Symbol Function 1 SYNC External clock synchronization input. 2 VIN_CS Supply input for the internal current sense amplifier. 3 VIN_IC Supply input for internal logic and analog circuitry. 4 BST Supply input for the floating top gate driver. An external diode, DBST, from 6VOUT and a 0.1 mF to 1 mF capacitor, CBST, to VSW forms a boost circuit. 5 GH Gate driver output for the external high−side NMOS FET. 6 VSW Switch−node. This pin connects to the source of the high−side MOSFET and drain of the low−side MOSFET. This pin serves as the switch output to the inductor. 7 GL 8 PGND Power Ground. Ground reference for the high−current path including the NMOS FETs and output capacitor. 9 6VOUT Output of internal fixed 6.0 V LDO. 10 AGND Analog Ground. Ground reference for the internal logic and analog circuitry as well as ROSC and the compensators. 11 EN Enable input. When disabled, the LDO, internal logic and analog circuitry and gate drivers enter sleep mode, drawing under 1 mA. 12 VIN Supply input for the SMPS. 13 VFB SMPS’s voltage feedback. Inverting input to the voltage error amplifier. Connect to VOUT through a resistive divider. 14 VCOMP SMPS’s voltage error amplifier output and non−inverting input to the current error amplifier. 15 CCOMP SMPS’s current error amplifier output and inverting input to the PWM comparator. 16 CFB 17 CSOUT 18 CSN Differential current sense amplifier inverting input. 19 CSP Differential current sense amplifier non−inverting input. 20 ROSC Oscillator’s frequency adjust pin. Resistor to ground sets the oscillator frequency. Gate driver output for the external low−side NMOS FET. SMPS’s current feedback. Inverting input to the current error amplifier. Single−ended output of the differential current sense amplifier. Connect to CFB through a resistor. Non−inverting input to the cycle−by−cycle overcurrent comparator. MAXIMUM RATINGS (Voltages are with respect to AGND unless otherwise indicated.) Rating Dc Supply Voltage (VIN) Peak Transient Voltage (Load Dump, EN = 0 V) Dc Supply Voltage (VIN_CS) Dc Supply Voltage (VIN_IC) Value Unit −0.3 to 20 45 V 46 V 6.5 V −0.7 to 40.7 −2 V 46 wrt PGND 7 wrt VSW V Pin Voltage (GL) −0.3 to 7 wrt PGND V Pin Voltage (EN) −0.3 to 40 V Pin Voltage (CSP, CSN) −0.3 to 10 V Pin Voltage (VFB, VCOMP, CSOUT, CFB, CCOMP, SYNC, ROSC, 6VOUT) −0.3 to 7 V Pin Voltage (VSW) t ≤ 50 ns Pin Voltage (BST, GH) Pin Voltage (PGND) −0.3 to 0.3 V Operating Junction Temperature −40 to 150 °C Storage Temperature Range −65 to 150 °C Peak Reflow Soldering Temperature: Lead−free 60 to 150 seconds at 217°C 265 peak °C Stresses exceeding Maximum Ratings may damage the device. Maximum Ratings are stress ratings only. Functional operation above the Recommended Operating Conditions is not implied. Extended exposure to stresses above the Recommended Operating Conditions may affect device reliability. http://onsemi.com 3 NCV8851 ATTRIBUTES Characteristic Value ESD Capability (All Pins) Human Body Model Machine Model Charge Device Model ≥ 1.5 kV ≥ 200 V ≥ 1.0 kV Package Thermal Resistance Junction–to–Ambient, RqJA (Note 1) Junction–to–Ambient, RqJA (Note 2) 156°C/W 108°C/W 1. 50 mm2, 1.0 oz copper on FR4 board. 2. 500 mm2, 1.0 oz copper on FR4 board. ELECTRICAL CHARACTERISTICS (−40°C < TJ < 150°C, 4.5 V < VIN < 20 V, 4.5 V < BST < 46 V, ROSC = 51.1 kW, unless otherwise specified) Conditions Min Typ Max Unit VIN = 13.2 V, EN = 0 V, Sleep Mode −40°C < TA < 125°C − − 1 mA VIN = 13.2 V, VFB = 1 V EN = 5 V, No Switching − 2.0 3.0 mA VIN = 13.2 V, VFB = 0 V EN = 5 V, Switching − 3.2 5.0 mA VIN = 13.2 V, VFB = 0 V, EN = 5 V Switching, 3.3 nF on GH and GL − 10 20 mA Thermal Shutdown Guaranteed by Design 150 180 210 °C Thermal Shutdown Hysteresis Guaranteed by Design − 10 20 °C Undervoltage Lockout (VIN_IC) VIN_IC increasing 4.1 4.3 4.5 V 50 125 200 mV 0.784 0.8 0.816 V 80 150 220 ns Static Operating − 140 200 ns ROSC = 51.1 kW ROSC = 23.2 kW ROSC = 16.2 kW 153 306 425 170 360 500 187 414 575 kHz kHz kHz 0.9 1.1 1.3 V Characteristic GENERAL Quiescent Current (IVIN + IVIN_CS + IBST) LDO Current Undervoltage Lockout Hysteresis SWITCHING REGULATOR Reference Voltage Minimum GH Off Time Minimum GH Pulse Width OSCILLATOR Switching Frequency Ramp Voltage Amplitude VOLTAGE ERROR AMPLIFIER DC Gain Guaranteed by Design 70 73 − dB Gain−Bandwidth Product Guaranteed by Design 8.0 10 − MHz Charge Currents Source, VCOMP = 0 V 2 4 − mA Sink, VCOMP = 1.75 V 1.3 3 − mA Guaranteed by Design − 0.1 1.0 mA 1.2 − 10.0 V − 1 − V/V FB Bias Current CURRENT SENSE AMPLIFIER Common−Mode Range Amplifier Gain 0 ≤ (CSP−CSN) ≤ 100 mV 1.2 V ≤ CSN ≤ 10.0 V http://onsemi.com 4 NCV8851 ELECTRICAL CHARACTERISTICS (−40°C < TJ < 150°C, 4.5 V < VIN < 20 V, 4.5 V < BST < 46 V, ROSC = 51.1 kW, unless otherwise specified) Characteristic Conditions Min Typ Max Unit CURRENT ERROR AMPLIFIER DC Gain Guaranteed by Design 70 73 − dB Gain−Bandwidth Product Guaranteed by Design 8.0 10 − MHz Source, CCOMP = 1.75 V 2 4 − mA Sink, CCOMP = 1.75 V 1.3 3 − mA Guaranteed by Design − 0.1 1.0 mA 2.7 3.5 − V 80 100 125 mV 115 165 215 mV − 200 − ns 20 − − mV FSW − 600 kHz VSYNC = 0 V VSYNC = 5.0 V − − 0.1 10 0.2 20 mA Logic Low Logic High − 2.0 − − 0.8 − V Output Voltage IOUT = 20 mA 5.8 6.0 6.2 V Dropout Voltage IOUT = 20 mA − − 200 mV 30 75 120 mA VGH = 2 V, VIN_IC = 6 V, Guaranteed by Design VGH = 4 V, VIN_IC = 6 V, Guaranteed by Design − 1.5 − A − 1.5 − A VIN_IC = 6 V VGL = 1.0 V Guaranteed by Design − 1.5 − A − 1.5 − A GH to GL Delay VIN = 13.2 V − 40 70 ns GL to GH Delay VIN = 13.2 V − 40 70 ns FSW = 170 kHz − 14 − ms Input Threshold Logic Low Logic High − 2.0 − − 0.8 − V Input Current EN = 2.0 V − 3.0 10 mA Charge Currents FB Bias Current Clamping Voltage CURRENT LIMIT Average Current Limit Threshold 1.2 V ≤ CSN ≤ 10.0 V Cycle−by−Cycle Current Limit Threshold Voltage Cycle−by−Cycle Current Limit Response Time Guaranteed by Design Cycle−by−Cycle and Average Current Limit Threshold Difference SYNC SYNC Frequency Range SYNC Pin Bias Current SYNC Threshold Voltage 6.0 V LDO Current Limit GATE DRIVERS GH Sink Current GH Source Current GL Sink Current GL Source Current SOFT START Time ENABLE (EN) http://onsemi.com 5 NCV8851 TYPICAL CHARACTERISTICS (TA = +25°C, VIN = 13.2 V, ROSC = 51.1 kW, unless otherwise noted) 100% 90% 4.0 20 3.5 18 3.0 16 Top Gate Bottom Gate 70% 60% 50% 30% 2.0 270 320 370 420 Switching Frequency (kHz) 470 38 Top Gate 220 270 320 370 420 470 0 Switching Frequency (kHz) Figure 3. Soft−start Time vs. Frequency Switching 105% 95% 28 85% Operating Current (%) 33 23 18 No Switching 0 1 2 Load Capacitance (nF) 3 65% 0.6 0.4 0.2 −50 0 50 100 Ambient Temperature (°C) 0 150 −50 0 50 100 Figure 8. Sleep Mode Quiescent Current vs. Temperature 101% 101% 100.5% 98% 100% Switching Frequency (%) Cycle−by−Cycle OCP Threshold (%) Figure 7. Operating Quiescent Current vs. Temperature 99% 99% 98% 97% −50 0 50 100 150 Ambient Temperature (°C) Figure 9. Average Current−Limit Threshold vs. Temperature 150 Ambient Temperature (°C) Figure 6. Driver Rise Time vs. Load Capacitance 100% 4 0.8 75% 4 3 1 45% 8 2 Figure 5. Driver Fall Time vs. Load Capacitance 55% 13 1 Load Capacitance (nF) Figure 4. Driver Quiescent Current vs. Frequency Bottom Gate 12 8 170 Shutdown Current (μA) 220 14 10 1.0 170 Rise Time (ns) 2.5 1.5 40% Average Current Limit Threshold (%) Fall Time (ns) Driver Current (mA) Soft Start Time (%) 80% −50 0 50 100 Ambient Temperature (°C) Figure 10. Cycle−by−Cycle Overcurrent Protection Threshold vs. Temperature http://onsemi.com 6 150 100.0% 99.5% 99.0% −50 0 50 100 Ambient Temperature (°C) Figure 11. VREF vs. Temperature 150 NCV8851 TYPICAL CHARACTERISTICS (TA = +25°C, VIN = 13.2 V, ROSC = 51.1 kW, unless otherwise noted) 102% 170 kHz 360 kHz 70 500 kHz GH to GL 108% GL to GH 65 60 106% Minimum Pulse Width (%) 55 100% 50 Delay (ns) Switching Frequency (%) 101% 99% 98% 45 40 35 30 97% 104% 102% 100% 25 20 96% −50 0 50 100 98% −50 150 0 Figure 12. Oscillator Frequency vs. Temperature 50 100 150 −50 (Ω) 120 100 110 Dropout Voltage (mV) Output Capacitor ESR 6V OUT (V) 99.50% 0.00 STABLE 10 1 UNSTABLE (0.1uF only) 10.00 15.00 20.00 5 10 15 20 Figure 16. LDO Stability Region 100 70 80 60 40 20 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Load Current (A) Figure 18. Efficiency vs. Load Current 5 V, 170 kHz Demo Board http://onsemi.com 7 −50 0 50 100 150 Figure 17. LDO Dropout Voltage vs. Temperature 93% 0.0 80 Ambient Temperature (°C) LDO Load Current (mA) Efficiency (%) Figure 15. LDO Load Regulation 90 50 0 LDO Load Current (mA) 100 60 0.1 5.00 150 130 UNSTABLE 99.75% 100 Figure 14. GH Minimum Pulse Width vs. Temperature 1000 100.00% 50 Ambient Temperature (°C) Figure 13. Non−Overlap Delay vs. Temperature 100.50% 100.25% 0 Ambient Temperature (°C) Ambient Temperature (°C) NCV8851 DETAILED OPERATING DESCRIPTION General inductor current is limited via average current limiting (ACL) and cycle−by−cycle overcurrent protection (OCP). Thermal shutdown (TSD) is also implemented to protect the device from overheating. The NCV8851 is a synchronous buck controller with internal 1.5 A gate drivers designed to drive NMOS FETs. The internal gate drivers simplify design, improve performance and efficiency and minimize board area. The controller uses an 800 mV, 2.0% reference, allowing for a wide range of precise output voltage programmability. The NCV8851 also provides a programmable fixed frequency range of 170 kHz to 500 kHz, allowing more design flexibility in compromising efficiency versus components’ size and cost. This frequency is conveniently set with an external resistor to ground. An external clock signal can also be used to synchronize the NCV8851 to a higher operating frequency during operation. To protect against possible damage of external power−stage components, excessive inrush of current during start−up is prevented by an internal soft−start, and The NCV8851 employs an average current mode control (ACMC) architecture to regulate the output voltage. ACMC uses two loops, as seen in Figure 19. Through the current error amplifier (CEA), the inner current loop monitors the inductor current with the unity gain current sense amplifier (CSA). The current loop responds to input voltage changes, affecting the line transient response. Using the voltage error amplifier (VEA), the outer voltage loop monitors the output voltage, responding to output load changes, affecting the load transient response. Feedback resistors in the voltage loop select the output voltage. Outer Voltage Loop RS L VSW Inner Current Loop VOUT C − + RL CSA Gain=1 − CEA + VEA − PWM and Gate Drivers Average Current Mode Control VREF + Figure 19. ACMC Loops amplifiers, shown as complex impedances in Figure 19. An input resistor from the CSA to the CEA sets the gain of the CEA. The voltage loop also has a pair of feedback resistors from VOUT to set the output voltage and gain of the VEA. Unlike voltage mode control (VMC) of buck regulators, which almost always require the extra components of a Type−III compensation network for adequate transient response, ACMC buck regulators use Type−II compensation. This greatly simplifies the compensator design and optimization process, while offering much faster transient response than a Type−I compensation network. Additionally, the two−loop system separates the effects of output components between the two loops, further simplifying the compensation process. Type−II compensation places a zero and two poles in each of the error loops to offset the effects of the inherent open−loop response. This compensation requires a resistor and two capacitors in the feedback loop for each of the error Enable The enable input (EN) is a TTL−compatible input used to activate the internal LDO. The NCV8851 is disabled when the EN pin is pulled below the enable input logic low threshold voltage, causing a normal shutdown to occur, putting the part into a low quiescent current sleep mode. When the EN pin is pulled above the enable input logic high threshold voltage, the part is enabled, the LDO output is brought up and then the internal soft−start begins. http://onsemi.com 8 NCV8851 to support the internal rails and power the controller. The IC will start up when enabled and VIN_IC surpasses the UVLO threshold and will shutdown when VIN_IC drops below the UVLO threshold minus the UVLO hysteresis. While VIN is less than the set point for VOUT, the output will run at max duty cycle, after soft−start, once VIN_IC surpasses the UVLO threshold. If EN is high and not tied to VIN, the output will begin to rise up while in UVLO, if a minimum output load of 1 kW is not met. VIN REN EN Internal Enable 125 kΩ DZEN2 22 V DZEN1 5.4 V Thermal Shutdown The NCV8851 provides Thermal Shutdown (TSD), which monitors the die temperature and turns off the top and bottom gate drivers if an over temperature condition is detected, for added protection. The internal soft−start capacitor is also discharged. A normal soft−start will occur when the die temperature falls below the TSD threshold minus the TSD hysteresis. Figure 20. Enable Pin Equivalent Structure The EN pin can be tied to VIN in order to enable the part. If EN is above 22 V, DZEN2 will be conducting, as well as DZEN1. The current to DZEN1 is limited by an internal 125 kW resistor. If DZEN2 is conducting, it is recommended at least 250 mA is pulled through this diode. The resistor REN must be used if VIN can go above 20 V as follows. R EN(max) + Duty Cycle and Maximum Pulse Width Limits In steady state dc operation, the duty cycle will stabilize at an operating point defined by the ratio of the input to the output voltage. There is a built in minimum off−time which ensures that the bootstrap supply is charged every cycle, determining the maximum duty cycle at a given frequency. The NCV8851 can achieve at least a 95% duty cycle while operating at frequencies up to 200 kHz (89% at up to 500 kHz). VZ 250 mA Where VZ is the amount of volts where DZEN2 is conducting, but not yet supplied with 250 mA. For example, setting VZ to 1 V means REN must be less than 4 kW for DZEN2 to have at least 250 mA when VIN is at least 23 V; for the range of VIN between 22 V and 23 V, DZEN2 will be conducting, but not with the recommended 250 mA current. Internal Soft−Start The NCV8851 features an internal soft−start function, which reduces inrush current and overshoot of the output voltage. Figures 21 and 22 show a typical soft−start sequence. UVLO Undervoltage Lockout (UVLO) is provided to ensure that unexpected behavior does not occur when VIN_IC is too low VIN UVLO Threshold VSW 6VOUT and VIN_IC VOUT Soft−start Time 90%*VOUT t Soft−start Delay Figure 22. Switch−node in Soft−Start 10%*VOUT t Figure 21. Normal Start−up causing the soft−start time to be inversely related to the frequency set by ROSC. The internal soft−start capacitor is discharged when the part is disabled, enters TSD or enters UVLO, ensuring a proper start−up when the part is re−enabled, leaves TSD or leaves UVLO. Soft−start is achieved by ramping up the internal soft−start voltage (VSS), which is applied to the non−inverting input of the voltage error amplifier, effectively limiting the slew rate of VOUT rising. This ramp is generated by charging an internal soft−start capacitor based on the internal oscillator, http://onsemi.com 9 NCV8851 The 6VOUT pin should be externally connected through a diode to the BST pin, charging the BST capacitor during off−time to generate a voltage for the high−side driver. When the part is enabled and VIN is below the LDO regulated value, the LDO is in dropout and it tracks VIN. The LDO regulates its output once VIN is above the output set point plus the dropout voltage. An external bypass capacitor must be connected from 6VOUT to ground. A short to ground or overcurrent condition on the 6VOUT pin will be mitigated by the LDO current limit and internal thermal shutdown (TSD) circuitry which disables all outputs. A normal soft−start will occur when the die temperature falls below the TSD threshold. This sequence begins once VIN_IC surpasses its UVLO threshold when the part is enabled and the LDO output has risen. After an initial delay to assure a clean start−up, switching begins, the output initially rises quickly and then rises monotonically. The duty cycle is gradually increased until VOUT has reached its set point or until maximum duty cycle is reached. Normal Shutdown Behavior and Sleep Mode Normal shutdown occurs when the IC stops switching because the input supply drops below the UVLO threshold, the part enters TSD or the part is disabled. When disabled, the part enters sleep mode. In sleep mode, the LDO turns off and its output capacitor discharges, causing switching to stop, the internal soft−start capacitor to discharge and GH and GL to go low. The switch node enters a high impedance state and the output inductor and capacitors discharge through the load. The supply current reduces to the sleep mode quiescent current. Drivers The NCV8851 includes 1.5 A gate drivers to switch external N−Channel MOSFETs. This allows the NCV8851 to address high−power, as well as low−power conversion requirements. The gate drivers also include adaptive non−overlap circuitry. The non−overlap circuitry increases efficiency, which minimizes power dissipation, by minimizing the body diode conduction time, while protecting against cross−conduction (shoot−through) of the MOSFETs. A detailed block diagram of the non−overlap and gate drive circuitry used in the chip and related external components is shown in Figure 23. Internal Linear Regulator (LDO) The NCV8851 has an onboard low−dropout linear regulator (LDO) internally connected to drive the low−side gate. The 6VOUT pin should be externally connected to the VIN_IC pin to power the internal rails. Typically, a RC filter is used from 6VOUT to VIN_IC to further decrease noise on the internal rails. BST MainFault GH GL to GH Delay VSW VSW GL Threshold VSW Threshold GH to GL Delay PWM Output 6VOUT GL Main Fault PGND Figure 23. Gate Driver Block Diagram Since the BST capacitor only recharges when the low−side MOSFET is on, pulling VSW down to ground, the NCV8851 has a minimum off−time. This also means that the BST capacitor cannot be arbitrarily large, since 6VOUT needs to be able to charge it up during this minimum off−time so the high−side gate driver doesn’t run out of headroom. 6VOUT needs to supply charge both to the BST capacitor and also the low−side driver, so the LDO capacitor must be sufficiently larger than the BST capacitor. A 1 mF LDO capacitor is recommended. A capacitor is placed from VSW to BST and a diode is placed from 6VOUT to BST to create a bootstrap supply on the BST pin for the high−side floating gate driver. This ensures that the voltage on BST is about 6VOUT higher than VSW, less a diode drop, yielding a gate drive voltage high enough to enhance the high−side MOSFET. The BST capacitor supplies the charge used by the gate driver to charge up the input capacitance of the high−side MOSFET, and is typically chosen to be at least a decade larger than this capacitance. A 0.1 mF BST capacitor is recommended. http://onsemi.com 10 NCV8851 above the ACL threshold. This causes the PWM pulse to be terminated very quickly and disables the part from switching back on until the current through the inductor has dropped below the OCP threshold. Once the inductor current is below the OCP threshold, the part will begin switching again and the current will be limited by ACL, until the inductor current drops below the ACL threshold. An advantage of this current limiting scheme is that the NCV8851 will limit large transient currents yet resume normal operation on the following cycle. Additionally, the current will not run away, nor will the part latch off in case of a short, which is typical of other current limiting schemes employing high−side current sensing. Careful selection and layout of external components is required to realize the full benefit of the onboard drivers. The capacitors between VIN and GND and between BST and VSW must be placed as close as possible to the IC. The current paths for the GH and GL connections must be optimized, minimizing parasitic resistance and inductance. Current Limiting and Overcurrent Protection The NCV8851 contains average current limiting (ACL) and cycle−by−cycle overcurrent protection (OCP) to protect the power switches, inductor, current sense resistor and other external components. The current through the inductor is continuously sensed using the CSP and CSN pins. A sense resistor is placed between these pins to translate the output current to a proportional voltage. This voltage is compared to a fixed internal voltage threshold. When the differential voltage exceeds the ACL threshold, the PWM pulse is terminated for this cycle, limiting the current through the inductor. In steady−state operation, decreasing the load resistance while in ACL will cause the duty cycle and VOUT to decrease proportionally without skipping pulses or jitter. There is also a fast OCP path which is tripped when the differential voltage exceeds the OCP threshold, which is SYNC Feature An external clock signal can synchronize the NCV8851 to a higher frequency. The rising edge of the SYNC pulse turns on the power switch to start a new switching cycle, as shown in Figure 24. There is a 0.5 ms delay between the rising edge of the SYNC pulse and rising edge of the VSW pin voltage. The SYNC threshold is TTL logic compatible, and duty cycle of the SYNC pulses can vary from 10% to 90%. The SYNC frequency must be higher than the internal oscillator frequency set by ROSC. Figure 24. Synchronization from 170 kHz to an external 600 kHz signal Snubber 0.5 W of power dissipation for switching frequencies below 280 kHz and 1 W of power dissipation for switching frequencies above 280 kHz. The snubber should be placed close to the switch node pin. A snubber consisting of a 4300 pF ceramic capacitor and a 1 W reistor from switch node to ground is required to decrease noise susceptibility. The resistor should be rated for http://onsemi.com 11 NCV8851 APPLICATIONS INFORMATION Design Methodology leading to decreased efficiency, especially noticeable at light loads. Typically, the switching frequency is selected to avoid interfering with signals of known frequencies. Often, in this case, the frequency can be programmed to a lower value with ROSC and then a higher−frequency signal can be applied to the SYNC pin to increase the frequency dynamically to avoid given frequencies. A spread spectrum signal could also be used for the SYNC input, as long as the lowest frequency in the range is above the programmed frequency set by ROSC. Additionally, the highest SYNC frequency must not exceed maximum switching frequency limits. There are two limits on the maximum allowable switching frequency: minimum off−time and minimum on−time. These set two different maximum switching frequencies, as follows: Choosing external components for the NCV8851 encompasses the following design process: 1. Define operational parameters 2. Select switching frequency 3. Select current sensor 4. Select output inductor 5. Select output capacitors 6. Select input capacitors 7. Select compensator components (1) Operational Parameter Definition Before proceeding with the rest of the design, certain operational parameters must be defined. These are application−dependent and include the following: VIN: input voltage, range from minimum to maximum with a typical value [V] VOUT: output voltage [V] IOUT: output current, range from minimum to maximum with initial start−up value [A] ICL: desired typical current−limit [A] A number of basic calculations must be performed up−front to use in the design process, as follows: V OUT V IN(typ) D MAX + 1 * D MAX T MinOff F SW(max)2 + D MIN T MinOn Where: FSW(max)1: maximum switching frequency due to minimum off−time [Hz] TMinOff: minimum off−time [s] FSW(max)2: maximum switching frequency due to minimum on−time [Hz] TMinOn: minimum on−time [s] Alternatively, the minimum and maximum operational input voltage can be calculated as follows: V OUT D MIN + V IN(max) D+ F SW(max)1 + V OUT V IN(min) Where: DMIN: minimum duty cycle (ideal) [%] VIN(max): maximum input voltage [V] D: typical duty cycle (ideal) [%] VIN(typ): typical input voltage [V] DMAX: maximum duty cycle (ideal) [%] VIN(min): minimum input voltage [V] It should be noted that these are the ideal duty cycles; the actual duty cycles will be marginally higher than these calculated values. The actual duty cycles are dependent on load due to voltage drops in the MOSFETs, inductor and current sensor. V IN(min) + V OUT 1 * T MinOff @ F SW V IN(max) + V OUT T MinOn @ F SW Where: FSW: switching frequency [Hz] The switching frequency is programmed by selecting the resistor connected between the ROSC pin and ground. The grounded side of this resistor should be directly connected to the AGND pin. Avoid running any noisy signals under the resistor, since injected noise could cause frequency jitter. The graph in Figure 25 shows the required resistance to program the frequency. From 150 to 450 kHz, the following formula is accurate to within 3%: (2) Switching Frequency Selection Selecting the switching frequency is a trade−off between component size and power losses. Operation at higher switching frequencies allows the use of smaller inductor and capacitor values to achieve the same inductor current ripple and output voltage ripple. However, increasing the frequency increases the switching losses of the MOSFETs, R OSC + 8687000 F SW Where: ROSC: frequency program resistor [W] Some specific values for switching frequency with standard 1% resistors can be seen in Table 1. http://onsemi.com 12 NCV8851 600 smaller physical size. Since the inductor is often one of the largest components in the power supply, a minimum inductor value is particularly important in space− constrained applications. From an electrical perspective, an inductor is chosen for a set amount of current ripple and to assure adequate transient response. Larger inductor values limit the switcher’s ability to slew current through the output inductor in response to output load transients, impacting the dynamic response. While the inductor is slewing current during this time, output capacitors must supply the load current. Therefore, decreasing the inductance allows for less output capacitance to hold the output voltage up during a load step. Load transient simulation is a powerful tool in anticipating this response. For switchers with both cycle−by−cycle overcurrent protection (OCP) and average current limiting (ACL), the OCP and ACL references are compared to the sensed current via sense resistance, RS. A minimum inductance is required to prevent the OCP from tripping during the onset of ACL during typical operation as follows: FSW (kHz) 500 400 300 200 100 0 10 20 30 40 50 60 70 80 90 ROSC (kW) Figure 25. Frequency vs. ROSC Table 1. Frequency vs. ROSC FSW (kHz) ROSC (kW) 170 51.1 250 34.8 300 28.7 360 23.2 500 16.2 L MIN + Where: LMIN: minimum inductance to assure OCP and ACL do not both trip [H] DVCL: difference between OCP and ACL threshold voltages [V] For switchers that use the current signal of the inductor for control purposes, the voltage ripple over the sense resistance must be sufficient in magnitude to counteract the contribution due to inherent comparator offsets and other errors, as follows: The soft−start time can be estimated as follows: T SS [ V OUT(1 * D) RS @ 2 @ F SW DV CL F0 @ T SS0 F SW Where: TSS: soft−start time [s] F0: specified frequency [Hz] TSS0: soft−start time at specified frequency [s] L MAX + (3) Current Sensor Selection Current sensing for average current mode control relies on the inductor current signal. This is translated into a voltage via a current sensor, which is then measured differentially by the current sense amplifier, generating a single−ended output to use as a control signal. The easiest means of implementing this transresistance is through the use of a sense resistor in series with the output inductor and capacitors. A sense resistor should be selected as follows: V OUT @ (1 * D MAX) RS @ F SW Ë L @ V CL Where: LMAX: maximum inductance to assure adequate voltage ripple over the sense resistance [H] κL: inductor current ripple to current limit ratio [%] VCL: threshold voltage for the current limit [V] The inductor current ripple to current limit ratio is expressed as follows: ËL + V R S + CL I CL iL I CL As a rule of thumb, ensuring that κL is at least 5% to 10% has been empirically sufficient. Smaller values of inductance increase the regulator’s maximum achievable slew rate and decrease the necessary capacitance, at the expense of higher ripple current, which causes higher output voltage ripple. The peak−to−peak ripple current is given by the following equation: Where: RS: sense resistor [W] VCL: current limit threshold voltage [V] ICL: desired current limit [A] Alternative methods, such as lossless inductor current sensing, are feasible but beyond the scope of this document. (4) Output Inductor Selection iL + Both mechanical and electrical considerations influence the selection of an output inductor. From a mechanical perspective, smaller inductor values generally correspond to V OUT @ (1 * D) L @ F SW Where: iL: peak−to−peak output current ripple [App] http://onsemi.com 13 NCV8851 (5) Output Capacitor Selection The ripple current is at a maximum when the duty cycle is at a minimum value and vice versa, as follows: The output capacitor is a basic component for the fast response of the power supply. During a load step, for the first few microseconds, it supplies the current to the load. The controller immediately recognizes the load step and increases the duty cycle, but the current slope is limited by the inductor’s slew rate. During a load release, the output voltage will overshoot. The capacitance will dampen this undesirable response, decreasing the amount of voltage overshoot. In the case of stepping into a short, the inductor current approaches zero with the worst case initial current at the current limit and the initial voltage at the output voltage set point, calculating the voltage overshoot as follows: @ (1 * D MIN) V i L(max) + OUT L @ F SW i L(min) + V OUT @ (1 * D MAX) L @ F SW Where: iL(max): maximum inductor current ripple [App] iL(min): minimum inductor current ripple [App] From this equation it is clear that the ripple current increases as L decreases, emphasizing the trade−off between dynamic response and ripple current. The peak and valley values of the triangular current waveform are as follows: iL 2 i I L(vly) + I OUT * L 2 I L(pk) + I OUT ) DV OS + ǸL @CI CL 2 ) V OUT 2 * V OUT Accordingly, a minimum amount of capacitance can be chosen for a maximum allowed output voltage overshoot: Where: IL(pk): peak (maximum) value of ripple current [A] IL(vly): valley (minimum) value of ripple current [A] Saturation current is specified by inductor manufacturers as the current at which the inductance value has dropped a certain percentage from the nominal value, typically 10%. For stable operation, the output inductor must be chosen so that the inductance is close to the nominal value even at the peak output current, IL(pk). It is recommended to choose an inductor with saturation current sufficiently higher than the peak output current, such that the inductance is very close to the nominal value at the peak output current. This introduces a safety factor and allows for more optimized compensation. Inductor efficiency is another consideration when selecting an output inductor. Inductor losses include dc and ac winding losses and core losses. Core losses include eddy current losses, which are very low due to high core resistance, and magnetic hysteresis losses, which increase with peak−to−peak ripple current. Core losses also increase as switching frequency increases. Ac winding losses are based on the ac resistance of the winding and the RMS ripple current through the inductor, which is much lower than the dc current. The ac winding losses are due to skin and proximity effects and are typically much less than the dc losses, but increase with frequency. Dc winding losses account for a large percentage of output inductor losses and are the dominant factor at switching frequencies at or below 500 kHz. The dc winding losses in the inductor can be calculated with the following equation: C MIN + L @ I CL 2 (V OUT ) DV OS(max)) 2 * V OUT 2 Where: CMIN: minimum amount of capacitance to minimize voltage overshoot to DVOS(max) [F] DVOS(max): maximum allowed voltage overshoot during a short [V] A maximum amount of capacitance can be found based on the inrush current and current limit. To calculate the input startup current, the following equation can be used: I INRUSH + C OUT @ V OUT ) I OUT(i) t SS Where: IINRUSH: input current during startup IOUT(i): initial output current If the inrush current is higher than the steady−state input current with the maximum load, then the input fuse should be rated accordingly, if one is used. During soft−start, the inductor current must provide current to the load, as well as current to charge the output capacitor. The maximum current which the inductor is allowed to conduct is the current limit. Setting the inrush current to the current limit, this puts a limit on the maximum capacitor size, as follows: C MAX + (I CL * I OUT(i)) @ t SS V OUT Where: CMAX: maximum output capacitance [F] Capacitors should also be chosen to provide acceptable output voltage ripple with a dc load, in addition to limiting voltage overshoot during a dynamic response. Key specifications are equivalent series resistance (ESR) and equivalent series inductance (ESL). The output capacitors must have very low ESL for best transient response. The PCB traces will add to the ESL, but by putting the output capacitors close to the load, this effect can be minimized and ESL neglected in determining output voltage ripple. P L(dc) + I OUT 2 @ R dc Where: PL(dc) : dc winding losses in the output inductor Rdc: dc resistance of the output inductor (DCR) As can be seen from the above equation, to minimize inductor losses, an inductor with very low DCR should be chosen. http://onsemi.com 14 NCV8851 capacitors must be rated to handle a ripple current of one−half the maximum output current at the switching frequency. ESR is the majority cause of losses in the input capacitors. Losses in the input capacitors can be calculated with the following equation: The capacitance itself causes a voltage ripple due to the current ripple. This is as follows: VQ + iL @ D C @ F SW Where: vQ: output voltage ripple due to output capacitance [Vpp] Also, the ripple current through the inductor causes a voltage ripple over the output capacitor due to its ESR as follows: P CIN + I IN(RMS) 2 @ R ESR(CIN) Where: PCIN = power loss in the input capacitors RESR(CIN) = effective series resistance of the input capacitance Due to large current transients through the input capacitors, electrolytic, polymer or ceramics should be used. If a tantalum must be used, it must be surge protected, to prevent against capacitor failure. Due to the large ripple current, it is common to put small ceramic capacitors in parallel with the bulk input capacitors, which will handle a significant portion of the ripple current. A value of 0.01 mF to 0.1 mF placed near the MOSFETs is recommended. V ESR + i L @ R ESR Where: vESR: output voltage ripple due to the effects of ESR [Vpp] RESR: total ESR of output capacitors [W] Typically, the ripple due to ESR dominates, having the largest effect on output voltage ripple. The total output voltage ripple in steady−state operation can be calculated as follows: V OUT + V Q ) V ESR + Ë C @ V OUT (7) Compensator Design Where: vOUT: total output voltage ripple [Vpp] κC: percent output voltage ripple [%] Typically, the voltage ripple percentage is a performance parameter used to decide on the desired output capacitor. The maximum total effective ESR of the output capacitors is calculated as follows: R ESR(max) + The purpose of the compensators is to stabilize the dynamic response of the converter. By optimizing the compensators, stable regulation with fast input line and output load transient response is achieved. Compensator design is related to the placement of zeros and poles in the closed loop, in order to assure stability with optimized transient response. The general approach is to use some rule of thumb values and then tune them through simulation to optimize load step response, while assuring stability over line and load variations. Type−II compensators are used with the two error amplifiers in average current mode control. The CEA closes the inner current−loop and the VEA closes the outer voltage−loop. As a rule of thumb, a zero is placed in each loop with the intent to compensate the effects of the double pole from the output inductor and capacitor. Additionally, a pole is placed at origin, due to the negative feedback, and a pole is also placed in each loop with the intent to compensate the effects of the double right−half−plane zero from the current sampling function. The crossover frequency is then set so that gain limitations of the error amplifier are not exceeded. The compensator must assure there is adequate phase margin in the total closed−loop response, which can be analyzed on a small−signal basis. Further reduction in loop gain, via decreasing the crossover frequency, may be required to avoid large−signal clamping limitations; this effect can be seen in simulation and taken care of in the compensator tuning process. V OUT * V Q i L(max) Where: RESR(max): maximum allowable total ESR of output capacitors It should be noted that these values of ESR are at the switching frequency and ESR decreases as frequency increases. The steady−state power lost due to the ESR of the output capacitor can be calculated as follows: P C(ESR) + 1 i L 2 @ R ESR 3 (6) Input Capacitor Selection The input capacitors have to sustain the ripple current produced during the on time of the high−side MOSFET and must have a low ESR to minimize the losses. The RMS value of this ripple is: I IN(RMS) + I OUT ǸD @ (1 * D) Where: IIN(RMS) = input RMS current The large majority of the ripple spectrum will be at the switching frequency. The above equation reaches its maximum value with D = 0.5, IIN(RMS) = IOUT/2. The input http://onsemi.com 15 NCV8851 Equations for placement of pole, zero and crossover frequency follow: Current−loop Compensator w iz + Voltage−loop Compensator 1 ǸL @ C w vz + F SW @ p 4 w ip + w vp + 2 ǸL @ C F SW @ p 4 w v + 2 @ w vp w i + 2 @ w ip feedback capacitors (CC1, CV1) in series with feedback resistor are chosen, on the order of less than 3 nF. The values are calculated as follows: The implementation of the above compensators is through a resistance on the negative input (RC2, RV2), resistor (RC1, RV1) and capacitor (CC1, CV1) in series in feedback and another capacitor (CC2, CV2) in feedback of an opamp. The Current−loop Compensator Voltage−loop Compensator R C1 + 1 w iz @ C C1 R V1 + 1 w vz @ C V1 C CE + 1 w ip @ R C1 C VE + 1 w vp @ R V1 C C2 + C C1 C C R C2 + C1 CE C V2 + *1 C 1 w i @ (C C1 ) C C2) R V2 + The resistance on the negative input of the VEA (RV2) also sets the output voltage. This resistance comprises a resistor divider with a resistor from the output voltage to the negative input of the VEA (RF1) and a resistor from the negative input of the VEA to ground (RF0). The resistor values are calculated as follows: R F1 + C V1 C V1 VE *1 1 w v @ (C V1 ) C V2) P TG + Q TG @ F SW @ V BST Where: QTG: total high−side MOSFET gate charge at VBST VBST: BST pin voltage The low−side synchronous rectifier MOSFET gate driver losses are: P BG + Q BG @ F SW @ V CC V OUT @ R V2 V REF Where: QBG: total low−side MOSFET gate charge at VIN The junction temperature of the controller can then be calculated as follows: R @ R F1 R F0 + V2 R F1 * R V2 T J + T A ) P IC @ R qJA Thermal Considerations Where: TJ = junction temperature of the IC TA = ambient temperature The power dissipation of the NCV8851 varies with the MOSFETs used, VIN and the boost voltage (VBST). The average MOSFET gate current typically dominates the control IC power dissipation. The IC power dissipation can be estimated as follows: RqJA = junction−to−ambient thermal resistance of the IC package The package thermal resistance (RqJA) can be obtained from the specifications section of this data sheet and a calculation can be made to determine the IC junction temperature. It should be noted that the physical layout of the board, the proximity of other heat sources such as MOSFETs and inductors and the amount of metal connected to the IC impact the temperature of the device. Use these calculations as a guide, but measurements should be taken in the actual application. P IC + V IN @ I Q ) P HS ) P L Where: PIC: control IC power dissipation IQ: IC measured supply current (quiescent current) PTG: high−side MOSFET gate driver losses PBG: low−side MOSFET gate driver losses The high−side switching MOSFET gate driver losses are: http://onsemi.com 16 NCV8851 PACKAGE DIMENSIONS TSSOP−20 CASE 948E−02 ISSUE C 20X 0.15 (0.006) T U 2X L K REF 0.10 (0.004) S L/2 20 M T U S V K K1 ÍÍÍÍ ÍÍÍÍ ÍÍÍÍ S J J1 11 B SECTION N−N −U− PIN 1 IDENT 0.25 (0.010) N 1 10 M 0.15 (0.006) T U S A −V− NOTES: 1. DIMENSIONING AND TOLERANCING PER ANSI Y14.5M, 1982. 2. CONTROLLING DIMENSION: MILLIMETER. 3. DIMENSION A DOES NOT INCLUDE MOLD FLASH, PROTRUSIONS OR GATE BURRS. MOLD FLASH OR GATE BURRS SHALL NOT EXCEED 0.15 (0.006) PER SIDE. 4. DIMENSION B DOES NOT INCLUDE INTERLEAD FLASH OR PROTRUSION. INTERLEAD FLASH OR PROTRUSION SHALL NOT EXCEED 0.25 (0.010) PER SIDE. 5. DIMENSION K DOES NOT INCLUDE DAMBAR PROTRUSION. ALLOWABLE DAMBAR PROTRUSION SHALL BE 0.08 (0.003) TOTAL IN EXCESS OF THE K DIMENSION AT MAXIMUM MATERIAL CONDITION. 6. TERMINAL NUMBERS ARE SHOWN FOR REFERENCE ONLY. 7. DIMENSION A AND B ARE TO BE DETERMINED AT DATUM PLANE −W−. N F DETAIL E −W− C G D H DETAIL E 0.100 (0.004) −T− SEATING DIM A B C D F G H J J1 K K1 L M MILLIMETERS MIN MAX 6.40 6.60 4.30 4.50 --1.20 0.05 0.15 0.50 0.75 0.65 BSC 0.27 0.37 0.09 0.20 0.09 0.16 0.19 0.30 0.19 0.25 6.40 BSC 0_ 8_ PLANE SOLDERING FOOTPRINT 7.06 1 0.65 PITCH 16X 0.36 16X 1.26 DIMENSIONS: MILLIMETERS http://onsemi.com 17 INCHES MIN MAX 0.252 0.260 0.169 0.177 --0.047 0.002 0.006 0.020 0.030 0.026 BSC 0.011 0.015 0.004 0.008 0.004 0.006 0.007 0.012 0.007 0.010 0.252 BSC 0_ 8_ NCV8851 ON Semiconductor and are registered trademarks of Semiconductor Components Industries, LLC (SCILLC). SCILLC reserves the right to make changes without further notice to any products herein. SCILLC makes no warranty, representation or guarantee regarding the suitability of its products for any particular purpose, nor does SCILLC assume any liability arising out of the application or use of any product or circuit, and specifically disclaims any and all liability, including without limitation special, consequential or incidental damages. “Typical” parameters which may be provided in SCILLC data sheets and/or specifications can and do vary in different applications and actual performance may vary over time. All operating parameters, including “Typicals” must be validated for each customer application by customer’s technical experts. SCILLC does not convey any license under its patent rights nor the rights of others. SCILLC products are not designed, intended, or authorized for use as components in systems intended for surgical implant into the body, or other applications intended to support or sustain life, or for any other application in which the failure of the SCILLC product could create a situation where personal injury or death may occur. Should Buyer purchase or use SCILLC products for any such unintended or unauthorized application, Buyer shall indemnify and hold SCILLC and its officers, employees, subsidiaries, affiliates, and distributors harmless against all claims, costs, damages, and expenses, and reasonable attorney fees arising out of, directly or indirectly, any claim of personal injury or death associated with such unintended or unauthorized use, even if such claim alleges that SCILLC was negligent regarding the design or manufacture of the part. SCILLC is an Equal Opportunity/Affirmative Action Employer. This literature is subject to all applicable copyright laws and is not for resale in any manner. PUBLICATION ORDERING INFORMATION LITERATURE FULFILLMENT: Literature Distribution Center for ON Semiconductor P.O. Box 5163, Denver, Colorado 80217 USA Phone: 303−675−2175 or 800−344−3860 Toll Free USA/Canada Fax: 303−675−2176 or 800−344−3867 Toll Free USA/Canada Email: [email protected] N. American Technical Support: 800−282−9855 Toll Free USA/Canada Europe, Middle East and Africa Technical Support: Phone: 421 33 790 2910 Japan Customer Focus Center Phone: 81−3−5773−3850 http://onsemi.com 18 ON Semiconductor Website: www.onsemi.com Order Literature: http://www.onsemi.com/orderlit For additional information, please contact your local Sales Representative NCV8851/D