NCP1239 Low−Standby High Performance PWM Controller Housed in SO−16 the NCP1239 represents a major leap toward ultra−compact Switch Mode Power Supplies specifically tailored for medium to high power off−line applications, e.g. notebook adapters. The NCP1239 offers everything needed to build a rugged and efficient power supply, including a dedicated event management to drive a Power Factor Correction front−end circuitry. The circuit disables the front−end PFC stage while still in fault or standby conditions by interrupting the PFC controller powering for improved no−load consumption figures. As soon as normal operating mode recovers, the NCP1239 feeds back the PFC that wakes−up. When power demand is low, the IC automatically enters the so−called skip cycle mode and provides excellent efficiency at light loads. Because this occurs at a user adjustable low peak current, no acoustic noise takes place. Features • • • • • • • • • • • • • Current−Mode Operation with Internal Ramp Compensation Internal High−Voltage Current Source for loss−less Start−up Adjustable Skip−Cycle Capability Selectable Soft−Start Period Internal Frequency Dithering for Improved EMI Signature Go−to−Standby Signal for PFC Front−Stage Large Vcc Operation from 12.2 V to 25 V 500 mV Over−Current Limit ±5% 500 mA/−800 mA Peak Current Capability 5 V / 10 mA Pinned−out Reference Voltage Adjustable Switching Frequency up to 250 kHz. Over−Load Protection Independent of the Auxiliary Vcc Adjustable Over−Power Compensation • High Power AC/DC Adapters for Notebooks etc. • Offline Battery Chargers • Set−Top Boxes Power Supplies, TV, Monitors etc. February, 2004 − Rev. 0 MARKING DIAGRAM 16 SO−16 FD SUFFIX CASE 751B 16 NCP1239FD AWLYWW 1 1 NCP1239 = Device Code A = Assembly Location WL = Wafer Lot Y = Year WW = Work Week PIN CONNECTIONS 1 16 GTS REF5V Fault Detect Rt Brown−out SS / Timer Skip Adjust FB HV NC NC Vcc Drv Gnd CS Over Power Limit ORDERING INFORMATION Device NCP1239FDR2 Typical Applications Semiconductor Components Industries, LLC, 2004 http://onsemi.com Package Shipping† SO−16 2500/Tape & Reel †For information on tape and reel specifications, including part orientation and tape sizes, please refer to our Tape and Reel Packaging Specifications Brochure, BRD8011/D. 1 Publication Order Number NCP1239/D NCP1239 Vbulk Rbo1 R10 to PFC_VCC OVP REF5V + NTC 16 2 15 3 14 4 13 5 12 6 11 7 10 8 9 Vout REF5V (5V/10mA) Thermistor + CBulk 1 BO Rbo2 Gnd VCC Rramp Rcomp NCP1239 Rt Cbo + Css Gnd Figure 1. Typical Application Example MAXIMUM RATINGS Rating Symbol Power Supply Voltage Vcc 25 V V 500 V RθJA 145 °C/W TJMAX 150 °C −60 to +150 °C 2 kV 200 V Maximum Voltage on pin 16 (HV) Maximum Junction Temperature Storage Temperature Range ESD Capability, HBM model (All pins except Vcc and HV) ESD Capability, Machine Model http://onsemi.com 2 Unit −0.3, +10 Pins 1 to 10 (except Vref pin) maximum voltage Thermal Resistance − Junction−to−Air, SOIC version Value NCP1239 ELECTRICAL CHARACTERISTICS (For typical values Tj = 25°C, for min/max values Tj = 0°C to +125°C, Max TJ = 150°C, VCC = 20 V unless otherwise noted.) Rating Symbol Pin Min Typ Max Unit Supply Section VCCON Turn−on threshold level, Vcc going up 13 15.5 16.4 17.5 V VCCOFF Minimum operating voltage after turn−on 13 10.5 11.2 12.2 V HYST1 Difference (VccON−VccOFF) 13 4.5 5.2 − V VCClatch Vcc decreasing level at which the latch−off phase ends 13 6.5 6.9 7.2 V VCCreset Vcc level at which the internal logic gets reset 13 − 4.0 − V ICC1 Internal IC consumption, no output load on pin 12 (@ IRt = 20 µA) 13 − 2.1 3.0 mA ICC2 Internal IC consumption, 1 nF output load on pin 12, FSW = 65 kHz 13 − 3.1 3.8 mA ICC2 Internal IC consumption, 1 nF output load on pin 12, FSW = 100 kHz 13 − 3.9 5.0 mA ICC2 Internal IC consumption, 1 nF output load on pin 12, FSW = 130 kHz 13 − 4.6 5.9 mA ICC3 Internal IC consumption, latch−off phase 13 − 400 700 µA High−Voltage Current Source (sunk by pin16), Vcc = 10 V 16 2.2 5.6 9.6 mA Start−up Charge Current flowing out of the Vcc pin, Vcc=10 V 13 1.7 5.2 7.2 mA High−Voltage Current Source, Vcc = 0 16 − 6.8 − mA 4.7 4.6 5.0 4.9 5.2 5.1 Internal Start−up Current Source IC1_hv IC1_Vcc IC2 5 V Reference Voltage (REF5V) REF5V Iref Reference Voltage @ no load on pin2 @ Ipin2 = 5 mA 2 V Current Capability 2 5.0 10 − mA Drive Output Vcl Output Voltage Positive Clamp 12 11.5 13.6 16 V Trise Output Voltage rise−time @ CL = 1 nF, 10−90% of output signal 12 − 40 − ns Tfall Output Voltage fall−time @ CL = 1 nF, 10−90% of output signal 12 − 20 − ns Vsource High State Voltage Drop @ Ipin12 = 3 mA and Vcc = 12 V 12 − 2.5 3.3 V Isource Source Current Capability (@ Vpin12 = 0 V) 12 − 500 − mA ROL Sink Resistance @ Vpin12=1 V 12 − 3.8 7.5 Ω Isink Sink Current Capability (@ Vpin12 = 10 V) 12 − 800 − mA Recommended Switching Frequency Range 12 25 − 250 kHz Vosc Pin 4 Voltage @ Rt = 100 kΩ 4 − 1.6 − V Kosc Product (Switching Freq. times the Rt pin 4 resistance) (Note 1) @ 65 kHz and 130 kHz 6050 6500 6950 Oscillator fsw kHz*kΩ ∆fsw Internal modulation swing, in percentage of Fsw − ±3.5 − % Dmax Maximum duty−cycle 74 80 85 % 60 120 80 160 100 185 0.48 0.47 0.50 0.50 0.52 0.52 − 140 220 Current Control Iocp Vopl TDEL_OCP Internal Current source of the Over Power Limit Pin @ 1 V on pin 5 and Vpin9 = 0.5 V @ 2 V on pin 5 and Vpin9 = 0.5 V 9 Over Power Limitation Threshold @ TJ = 25°C @ TJ = 0°C to 125°C 9 Propagation delay from Vpin9 > Vopl to gate turned off (pin12 loaded by 1 nF) 9 http://onsemi.com 3 µA V ns NCP1239 ELECTRICAL CHARACTERISTICS (For typical values Tj = 25°C, for min/max values Tj = 0°C to +125°C, Max TJ = 150°C, VCC = 20 V unless otherwise noted.) Symbol Rating Pin Min Typ Max Unit Maximum internal current set−point 10 0.84 0.90 0.95 V Propagation delay from Vpin10 > Ilimit to gate turned off (pin12 loaded by 1 nF) 10 − 140 220 ns Current Control ILimit TDEL CS TLEB−65kHz Leading Edge Blanking Duration (pins 9 and 10) @ 65 kHz 9, 10 − 420 − ns TLEB−130kHz Leading Edge Blanking Duration (pins 9 and 10) @ 130 kHz 9, 10 − 230 − ns Soft Start or Jittering charge current @ Vpin6 = 2.4 V 6 60 95 110 µA Idisch Jittering Discharge Current @ Vpin6 = 2.4 V 6 80 115 135 µA Vjitter Jittering Saw−Tooth Lower Threshold 6 1.67 1.80 1.89 V VjitterH Jittering Saw−Tooth Upper Threshold 6 2.85 3.00 3.20 V VtimerL Timer Peak Threshold 6 4.0 4.3 4.6 V ItimerC Timer Charge Current @ Vpin6 = 3.5 V and pin 8 open 6 4.1 5.5 6.6 µA ItimerD Timer Discharge Current @ Vpin6 = 3.5 V and pin8 open 6 − 400 − µA Internal Pull−up Resistor 8 − 20 − kΩ Source Current @ Vpin8 = 0.5 V 8 − 200 − µA Pin 8 to current set−point division ratio − − 3.0 − − Soft Start and Timer Ich Feedback Section Rup Ifb Iratio Internal Ramp Compensation Rramp Internal Resistor 10 − 32 − kΩ Vramp Internal Saw−Tooth Amplitude 10 − 3.2 − V Skipping Mode and Stand−By Management Rgts Pin 1 output impedance in stand−by state (pin 8 grounded, Vpin6 > 4.5 V) @ Vcc = 12.5 V 1 4.0 8.0 18 kΩ Igts Sink Current Source in Normal Mode @ Vpin8 = 2 V, pin7 open @ Vcc−Vpin1=0.7 V 1 0.6 1.0 − mA Default Feed−Back Level for Skip Cycle Operation and Stand−by Detection 7 380 430 480 mV FB_stby−out Default Feed−Back Level to Leave Standby 7 650 740 810 mV Vstby−out/Vskip Ratio leave stand−by set−point to skip cycle set−point 1.5 1.7 1.9 − − 110 − kΩ − 3.0 − − FB−skip Rpin7 Internal pin 7 Impedance 7 Pin 7 to skipping set−point ratio Brown−Out Detection BOthH Brown−Out Detection Upper Threshold 5 0.45 0.50 0.55 V BOthL Brown−Out Detection Low Threshold 5 0.20 0.24 0.28 V BOhyst Brown−Out Hysteresis 5 0.20 0.26 0.30 V Protections TSD Vfault °C Thermal Shutdown: Thermal Shutdown Threshold Hysteresis 140 30 Fault Detection Threshold 3 2.2 2.4 2.6 V 1. The nominal switching frequency fsw equals: fsw = KOSC / Rt. The implemented jittering makes the switching frequency continuously vary around this nominal value (+/−3.5% variation). http://onsemi.com 4 NCP1239 PIN FUNCTION DESCRIPTION Pin No. Pin Name Function 1 GTS Shuts the PFC down in standby The stand−by detection block changes pin1 state in accordance to the mode (stand−by or normal mode). Pin1 is designed to drive an external pnp transistor that connects or disconnects the NCP1239’s Vcc to the PFC’s. Pin Description 2 REF5V A 5V reference voltage This pin helps to internally bias the controller but can also be used to power surrounding logic gates for any purposes. The typical output current is 10 mA. This voltage source is disabled during the circuit start−up and latched−off phases. A 100 nF filtering capacitor must be placed between pin2 and ground. 3 Fault Detect Enables to permanently shutdown the part If the pin3 voltage exceeds 2.4 V, the circuit is permanently shut down. This pin can be used to monitor the voltage accross a thermistor in order to protect the application from excessive heating and/or to detect an Over−Voltage condition. 4 Rt Timing resistor 5 Brown−Out Brown−Out 6 SS / Timer Performs soft−start and fault timeout During Power on and fault conditions, the capacitor connected to this pin ensures a soft−start period. When a fault is detected, this pin is internally brought high by a current source. If 4.3 V are reached, the fault is confirmed and the circuit enters an auto−recovery burst mode, otherwise the pin goes back to a lower value and oscillates to perform frequency jittering. 7 Skip Adjust Adjust skip level By adjusting the skip cycle level, it is possible to fight against noisy transformers and modify the stand−by detection thresholds. Keep pin7 open to operate with the default levels (skip threshold set−point: 140 mV, normal mode recovery set−point: 250 mV). 8 FB Feedback signal An opto−coupler collector pulls this pin low to regulate 9 Over Power Limit Enables a precise peak current clamp and then an accurate Over Power Detection This pin delivers a current proportional to Vpin5, an image of the high voltage rail. Inserting a resistor between pin 9 and the current sense resistor, an offset proportional to the input voltage is built. Such offset compensates the circuit and power switch propagation delays for an accurate power limitation in the whole input voltage range. 10 CS The current sense input This pin receives the primary current information via a sense element. By inserting a resistor in series with this pin, it becomes possible to introduce ramp compensation. 11 Ground The IC ground 12 Drv Drives the MOSFET 13 Vcc Supplies the controller 14 NC − Creepage distance 15 NC − Creepage distance 16 HV The high−voltage startup Pin4 resistor allows a precise frequency programming from 20 kHz up to 250 kHz. This pin receives a portion of the Bulk capacitor to authorize operation above a certain level of mains only. It also serves to elaborate an offset voltage on pin 9 used for Over Power Compensation. − By offering up to +500 mA/−800 mA peak, this pin lets you drive large Qg MOSFET’s. It is clamped to 16 V maximum not to exceed the maximum gate−source voltage of most power MOSFET’s. This pin accepts up to 25 V from an auxiliary winding This pin connects to the bulk capacitor to generate the start−up current http://onsemi.com 5 NCP1239 FB<Vpin1 => Skip high Skip + Skip 7 adjust − FB 100k + Stby_detect S UVLOs Latch Reset Q 15r 450mV 16 HV R + Q 25r Internal Thermal Shutdown − FB>1.6*Vpin1 =>Stby_detect RESET 15 UVLO Fault detect TSD 3 (Vcc<VccOFF) − OVL 2.5V + S regOUT Fault 14 Q Vcc R Q Vdd PFC_Vcc Regul Vcc < 4V 10k 1 pfcON stdwn pfcOFF 1mA 13 Vcc Start−Up_Phase Stby Vstop S Q Vdd Vdd R Q OVL Divider by 2 Vcc<7V Stby_detect SS / timer Soft−Start and timer management 6 Error_Flag Output Buffer OUTon Soft Start Ipk limit REF5V 2 12 Drv Jittering Modulation 14V clamp + 5V Ramp Compensation pfcON 3.2V 32k 11 CLK − 5 PWM Latch BO_in BO BO_out + + 0.5V / 0.25V Gnd S Q R Q 10 CS LEB Vdd Vstop Vdd BO_in Oscillator Rt 4 75A/V x Vpin5 CLK − + 2.5V 9 Over Power Limit Skip Jittering Modulation LEB Vdd + − ”Jittered” Reference − + + 0.5V 20k FB 8 /3 0.9V to Skip Soft Start Ipk limit Error flag Figure 2. Internal Circuit Architecture http://onsemi.com 6 NCP1239 TYPICAL PERFORMANCE CHARACTERISTICS 8.5 8.0 7.5 6.0 IC2 (mA) IC1_HV, (mA) 7.0 5.0 6.5 5.5 4.0 3.0 −25 0 25 50 75 100 4.5 −25 125 0 Figure 3. High Voltage Current Source vs. Temperature @ VCC = 10 V Pin16 Leakage Current (A) 50 75 100 125 TEMPERATURE (°C) TEMPERATURE (°C) Figure 4. High Voltage Current Source vs. Temperature @ VCC = 0 V 50 16.7 45 16.6 40 16.5 VCC_on (V) 35 30 25 16.4 16.3 16.2 20 16.1 15 10 −25 25 0 25 50 75 100 16.0 −25 125 0 25 50 75 100 125 TEMPERATURE (°C) TEMPERATURE (°C) Figure 6. VCC Start−up Threshold vs. Temperature Figure 5. High Voltage Pin Leakage Current vs. Temperature 6.90 11.40 11.30 6.88 VCC_off (V) VCCLATCH (V) 11.20 11.10 11.00 6.86 6.84 10.90 10.80 −25 0 25 50 75 100 125 6.82 −25 0 25 50 75 100 TEMPERATURE (°C) TEMPERATURE (°C) Figure 7. VCC Turn−Off Threshold vs. Temperature Figure 8. VCC Latched−Off vs. Temperature http://onsemi.com 7 125 NCP1239 TYPICAL PERFORMANCE CHARACTERISTICS 2.80 5.0 2.60 4.5 2.40 4.0 ICC2 (mA) ICC1 (mA) 130 kHz 2.20 3.5 65 kHz 2.00 3.0 1.80 2.5 1.60 −25 0 25 50 75 100 2.0 −25 125 100 kHz 0 TEMPERATURE (°C) 25 50 75 100 125 TEMPERATURE (°C) Figure 9. No Load Circuit Consumption vs. Temperature Figure 10. Circuit Consumption (1nF on driver pin 12) vs. Temperature 0.60 5.05 0.55 5.00 0 mA 4.95 0.45 REF5V (V) ICC3 (mA) 0.50 0.40 0.35 4.90 5 mA 4.85 4.80 0.30 10 mA 4.75 0.25 0.20 −25 0 25 50 75 100 4.70 −25 125 0 TEMPERATURE (°C) 25 50 75 100 125 TEMPERATURE (°C) Figure 11. Latched−Off Mode Consumption vs. Temperature Figure 12. REF5V Voltage Source vs. Temperature 2.75 7.0 2.70 6.0 5.0 2.60 SINK () Vdrop (V) 2.65 2.55 2.50 4.0 3.0 2.45 2.0 2.40 2.35 −25 0 25 50 75 100 1.0 −25 125 0 25 50 75 100 125 TEMPERATURE (°C) TEMPERATURE (°C) Figure 13. Driver High State Voltage Drop vs. Temperature Figure 14. Driver Sink Resistance vs. Temperature http://onsemi.com 8 NCP1239 TYPICAL PERFORMANCE CHARACTERISTICS 83 82 14.5 81 Dmax (%) Vcl CLAMP VOLTAGE (V) 15.0 14.0 13.5 80 79 13.0 78 12.5 −25 0 25 50 75 100 77 −25 125 0 TEMPERATURE (°C) 65 kHz 125 160 BO=2 V 6540 140 130 kHz 6520 Iocp (A) Kosc (kHz*k) 100 180 6560 6500 6480 120 100 80 6460 BO=1 V 200 kHz 6440 0 25 60 50 75 100 40 −25 125 0 Figure 17. Oscillator Kosc Parameter vs. Temperature (Kosc=fsw*Rpin4) 0.52 0.92 0.51 0.91 ILimit (V) 0.93 0.50 0.89 0.48 0.88 50 75 75 100 125 0.90 0.49 25 50 Figure 18. Pin9 Current vs. Temperature (@ Vpin9 = 0.5 V) 0.53 0 25 TEMPERATURE (°C) TEMPERATURE (°C) Vopl (V) 75 Figure 16. Maximum Duty Cycle vs. Temperature 6580 0.47 −25 50 TEMPERATURE (°C) Figure 15. Driver Voltage Clamp vs. Temperature 6420 −25 25 100 0.87 −25 125 0 25 50 75 100 TEMPERATURE (°C) TEMPERATURE (°C) Figure 19. Over−Power Limitation Threshold vs. Temperature Figure 20. Maximum Current Set−Point vs. Temperature http://onsemi.com 9 125 NCP1239 TYPICAL PERFORMANCE CHARACTERISTICS 450 770 445 760 FBstby−out (mV) FBskip (mV) 440 435 430 425 750 740 730 420 720 415 410 −25 0 25 50 75 100 710 −25 125 0 25 50 75 100 125 TEMPERATURE (°C) TEMPERATURE (°C) Figure 21. Default Feed−back Threshold for Stand−by Detection vs. Temperature Figure 22. Default Feed−back Level for Normal Operation Recovery 0.497 0.250 0.496 0.248 BO_th_L (V) 0.494 0.493 0.246 0.244 0.492 0.242 0.491 0.490 −25 0 25 50 75 100 125 0.240 −25 0 TEMPERATURE (°C) 25 75 100 Figure 24. Brown−Out Low Threshold vs. Temperature 2.48 2.46 2.44 2.42 2.40 2.38 2.36 −25 50 TEMPERATURE (°C) Figure 23. Brown−Out Upper Threshold vs. Temperature Vfault (V) BO_th_H (V) 0.495 0 25 50 75 100 125 TEMPERATURE (°C) Figure 25. Fault Detect Threshold vs. Temperature http://onsemi.com 10 125 NCP1239 SS / timer pin Fault Management 4.3V New Start−up attempt Fault confirmed 100ms Fault not confirmed 100ms 10ms Jittering 3.0V 1.8V fmax fmin OVL signal (Over−Load) 1V Error flag 1V Error flag 1V Error flag Reset at UVLO PFC off PFC off PFC on DRV Vcc 16.4V 11.2V Fault occurs here 6.9V Fb is ok In the example with Cpin6=390 nF, the pin 6 capacitor approximately forces a 7.5 ms soft−start, a 10 ms jittering period and a 100 ms timer. Figure 26. Fault Management http://onsemi.com 11 NCP1239 Stand−by Detection Vpin6 (SS/timer) Standby is not confirmed Standby is confirmed, 4.3V Jittering 10ms 3.0V 1.8V 100ms delay 100ms PFC is down PFC running Bunches of pulses Drv FB No delay FB−stby−out (1.7*Vpin7) FB−skip (Vpin7) Stby_detect latch is armed Stby_detect latch is reset Skip activity Fb is ok Standby is entered Standby is left In the example with Cpin6=390 nF, the pin 6 capacitor approximately forces a 7.5 ms soft−start, a 10 ms jittering period and a 100 ms timer. Figure 27. Stand−by Detection http://onsemi.com 12 NCP1239 APPLICATION INFORMATION a stand−by situation. Pin1 state changes in accordance to the detected mode (stand−by or normal mode). Simply connect a pnp transistor between the NCP1239 Vcc and the PFC controller one and drive it using pin1, to enable the PFC stage in normal mode and disable it in stand−by. Soft−start: the capacitor connected to pin6 provides a soft−start sequence that precludes the main power switch from being stressed upon start−up. The same voltage is also used to perform frequency jittering and timing for the fault condition detection. Major Fault Detection: the circuit detects when pin3 voltage exceeds 2.4 V. When this occurs, the NCP1239 considers that a major fault is present and as a consequence, the circuit gets permanently latched−off. In this mode, the circuit needs the Vcc to go down below 4 V to reset, for instance when the user un−plugs the SMPS. This capability is mainly intended to detect an over−voltage condition or/and an over−heating of the application that would be sensed by a thermistor. Brown−out detection: by monitoring the level on pin5 during normal operation, the controller protects the SMPS against low mains conditions. When the pin5 voltage falls below 250 mV, the controllers stops pulsing until this level goes back to 500 mV to prevent any instability. Short−circuit protection: short−circuit and especially over−load protections are difficult to implement when a strong leakage inductance affects the transformer (the auxiliary winding level does not properly collapse…). Here, every time the feed−back pin is at its maximum (higher than 5 V practically), an error flag is asserted and a time period starts, thanks to pin6 capacitor. If pin6 reaches 4.6 V while the error flag is still present, the controller stops the pulses and goes into a latch−off phase, operating in a low−frequency burst−mode. As soon as the fault disappears, the SMPS resumes its operation. The latch−off phase can also be initiated, more classically, when Vcc drops below UVLO (11.2 V typical). Adjustable frequency and Internal dithering for improved EMI signature: pin4 offers a means to precisely adjust the switching frequency through a simple resistor to ground. Frequency operation is allowed up to 250 kHz. By modulating the internal switching frequency with the pin6 saw−tooth (100 Hz with 390 nF), natural energy spread appears and softens the controller’s EMI signature. 5 V reference voltage: a 5 V regulator is provided to help biasing any external circuitry in the vicinity of the controller. This reference voltage can typically supply up to 10 mA. The NCP1239 includes all necessary features to help building a rugged and safe switch−mode power supply. The following details the major benefits brought by implementing the NCP1239 controller: Current−mode operation with internal ramp compensation: implementing peak current mode control, the NCP1239 offers an internal ramp compensation signal that can easily be summed up to the sensed current. Subharmonic oscillations can thus be fought via the inclusion of a simple resistor, 500 mV Current Sense threshold for Over Power Limit: the NCP1239 operating in current mode, the circuit pin10 monitors the current to modulate its level according to the power demand. Due to the ramp compensation, one must generally note that the pin10 voltage is not the exact image of the inductor current. A precise current limitation being essential, the NCP1239 features a separate current sense pin (pin9) for an accurate over−current detection. The low threshold of this protection (500 mV) avoids excessive losses in the current sense resistor and improves the efficiency. In addition, pin9 sources a current that proportional to the high−voltage rail, compensates the current−sense and turn off delays at high line. A resistor inserted between pin9 and the sensing resistor offsets the pin9 current−sense information to build a precise overload protection, independent of the mains input. Large Vcc operation: the NCP1239 offers an extended Vcc range up to 25 V, bringing greater flexibility in Flyback or Forward applications. Internal high−voltage startup switch: reaching low levels of stand−by power represents a difficult exercise when the controller requires an external, lossy, resistor connected to the bulk capacitor. Thanks to an internal logic, the controller disables the high−voltage current source after start−up which no longer hampers the consumption in no−load situations. Skip−cycle capability: a continuous flow of pulses is not compatible with no−load standby power requirements. Slicing the switching pattern in bunch of pulses drastically reduces overall losses but can, in certain cases, bring acoustic noise in the transformer. Thanks to a skip operation taking place at low peak currents only, no mechanical noise appears in the transformer. Furthermore, the skip threshold is made programmable to allow the best trade−off between noise and efficiency. Stand−by detect / Shutdown of the PFC front−stage: The NCP1239 incorporates an internal logic that is able to detect http://onsemi.com 13 NCP1239 Start−up Sequence When the power supply is first connected to the mains outlet, the internal current source (typically 5.2 mA) is biased and charges up the Vcc capacitor. When the voltage on this Vcc capacitor reaches the VccON level (typically 16.4 V), the current source turns off and no longer wastes any power. At this time, the energy stored by the Vcc capacitor serves to supply the controller and the auxiliary supply is supposed to take over before Vcc collapses below VccOFF. Figure 28 shows the internal arrangement of this structure: 16 16.4 V / 11.2 V + − As soon as Vcc reaches 16.4 V, driving pulses are delivered on pin12 and the auxiliary winding grows up the Vcc pin. Because the output voltage is below the target (the SMPS is starting up), the feed−back pin is at its maximum voltage. A resistor divider outputs the third of the feed−back voltage that forms the current set−point. This set−point is clamped and the limitation level slowly increases until it reaches 0.9V during the soft start time. In nominal operation, the set−point clamp keeps equal to 0.9 V (refer to Figure 29). As soon as the feed−back voltage is high enough to activate the 0.9 V set−point clamp (during the start−up period but also anytime an overload occurs), an internal error flag is asserted, testifying that the system is pushed to the maximum power. At that moment, a 100 ms time period (typically, with Cpin6=390 nF that also corresponds to 7.5 ms soft −start) starts while a logic block observes this error flag. If the error flag keeps asserted all along the 100ms period, then the controller assumes that the power supply really undergoes a fault condition and immediately stops all pulses to enter a safe burst operation. The 100 ms timer enables to distinguish a start−up phase (shorter than 100 ms) from an over−load condition. If the error flag is released before the 100 ms period has elapsed, the controller concludes that no error is present and resets the timer to use it for other purposes (e.g. frequency dithering). HV 5.2 mA / 0 13 CVcc Aux 10 The current source brings Vcc above 16.4 V and then turns off Figure 28. to Stand−by Management (Skipping, GTS) Vdd CLK PWM Latch 20k S 8 Q Q Vin Feed−back /3 R 0.9 V Current Sense Comparator Soft−start oscillator − Ramp Compensation Rramp + 10 LEB Current Sense Rsense pin5 (Brown−Out) Over Power Comparator + Over−Currents Compensation LEB Rcomp 9 Over Power Limit − + 500 mV Pin10 monitors the power switch current and compares it to the current set−point (one third of the feed−back voltage). The current set−point is limited by the soft−start during the power−on sequence and permanently clamped to 0.9 V. A second pin (pin9) monitors the current to clamp the power. Figure 29. http://onsemi.com 14 NCP1239 Figure 30 depicts the Vcc evolution during a proper startup sequence, showing the state of the error flag: Vcc VccON VccOFF Latch−off phase level Logic reset level FB Full power User Powers up! Feedback loops reacts... regulation Skip level Ip max Error Flag Timer 7.5ms SS No error has been confirmed An error flag gets asserted after the soft−start period and as soon as the current set−point is maximum Figure 30. http://onsemi.com 15 NCP1239 PFC Start−up Sequence To ensure an adequate start−up sequence of both PWM section and the PFC stage, some logic and timing need to be included as shown on the internal diagram. The key point here is the fact that the PFC always starts after the PWM section. As a result, the SMPS must be designed to cope with transient universal mains operation. Why this? Because of the light−to−heavy load transition where a case exists when the PFC is off, the PWM in standby and the load is suddenly applied. In this scenario, the PWM section must sustain the entire transient period that lasts until the PFC re−starts since it has been de−activated for standby. The stand−by detection block generates an internal signal “pfcON” that controls pin1 in accordance to the operation mode: − “pfcON” is high in normal mode and a current source draws 1mA from pin1, − “pfcON” is low in stand−by to disable the 1 mA current source. A 10 kΩ resistor pulls up pin1 to Vcc. This configuration makes it ideal to drive a pnp transistor that connects or disconnects the NCP1239 Vcc to the PFC controller one (refer to Figure 32). The “pfcON” signal is activated following Figure 31 diagram. Let’s split this drawing in different time periods to clearly depict signal assertions: Power on: during this time, Vcc rises up, the Vcc capacitor being charged by the 5.2 mA current source. When Vcc exceeds VccON (16.4 V typ.), driving pulses are delivered to the MOSFET in an attempt to crank the power supply. Vcc collapses (because the Vcc capacitor alone delivers the energy) until sufficient auxiliary voltage is built up in order to take over the start−up sequence and thus self−supply the controller. As long as the output voltage has not reached its wished value, the controller pushes for the maximum peak current. During the soft−start (7.5 ms with 390 nF on pin6), the maximum permissible current linearly increases till the maximum peak set−point is reached, the internal 0.9 V zener diode actively clamping the current amplitude to (0.9 V/Rsense). During this time, the NCP1239 asserts an error flag. A maximum current condition being observed, the circuit determines if this state results from either a normal response (start−up or a transient period) or a fault condition. To make the difference, each time the error flag is asserted, a 100 ms timer starts to count down. If the error flag keeps asserted for the 100 ms period, there is a fault and the PWM controller enters a safe, auto−recovery, burst mode to limit the dissipated heat (see below for more details). During the Power−on sequence, “pfcON” keeps low to pull−up pin1 to Vcc until the error flag is down. When the error flag is down, the power supply has entered regulation, its auxiliary voltage is stable, then pin1 can turn low (1 mA sink current) to safely allow PFC operation. Entering Standby: when skip cycle starts to activate, a 100 ms countdown takes place and the logic observes the skip activity. If the skip activity is still there at the end of the 100 ms, then stand−by is confirmed and the NCP1239 pulls up pin1 to Vcc to shut down the PFC. Leaving standby: in this case, as soon as the skip cycle activity disappears, the circuit immediately re−activates the 1 mA sinking current source of pin1, to enable the PFC: there is no reaction delay in this situation. Short−circuit condition: a short circuit is detected on the primary side by measuring the time the error flag is asserted. As explained, if this flag is asserted longer than 100 ms, then the PWM stops oscillating and enters a safe burst mode. In this case, pin1 is pulled up to Vcc and the PFC is shut down. During the burst, it is not activated (PFC is off) until the fault goes away and the power supply resumes operation. The PFC being shut off in short−circuit conditions, it naturally reduces the main MOSFET stress. Latch−off mode: if the controller is permanently latched−off due to a major fault (pin3 detection of an OVP or an excessive external temperature), the PFC is kept off (pin1 being tied to Vcc). http://onsemi.com 16 NCP1239 Vcc PWM regulation Short−circuit Short−circuit Stby stby is left 16.4V Nom Pout 11.2V 6.9V Timer 100ms 100ms 100ms 100ms One Vcc cycle is skipped to lower the burst mode duty cycle to typically 5% in fault conditions. 0.9V flag 7.5ms SS PFC Vcc If the fault had disappeared the SMPS would recover normal operation Stand−by is confirmed Connection of both controllers and PFC stage require a proper startup sequence Figure 31. http://onsemi.com 17 NCP1239 away as it is fully supplied by the PWM auxiliary winding and even high quiescent current devices do not hamper the standby power since they are completely disconnected in standby. The PFC controller connection is really straightforward as testified by Figure 32: simply connect to pin1, the base of a pnp transistor that connects the PFC’s Vcc to the NCP1239 one (perhaps add a small decoupling capacitor like a 0.1 F on the PFC) and this is all! The PFC startup network goes PFC stage Rectified ac line Q1 PFC_VCC 1 8 2 7 3 6 4 1 16 2 15 3 14 4 13 5 12 6 11 7 10 8 9 VCC 5 + PFC Controller + NCP1239 The NCP1239 turns off the pnp Q1 during the stand−by so that the PFC controller is no longer supplied in this mode. Figure 32. Short−Circuit or Overload Condition The NCP1239 differs from other controllers in the sense that a fault condition is detected independently of the auxiliary voltage level. In auxiliary supply−based power supplies, it is necessary that the (isolated) secondary output conditions properly reflects on the (non−isolated) auxiliary winding in order to instruct the controller on what is happening on the other side of the transformer. For the following reasons, it sometimes becomes extremely difficult to build an efficient short−circuit protection circuitry and even more difficult to implement over power detection (e.g. the output load is 25% above the nominal value but Vout is still present): The primary leakage inductance is high: this is probably the main reason why building efficient short−circuit detection is difficult. When the power switch opens, the leakage inductance superimposes a large over−voltage spike on the drain voltage. This spike is seen on the secondary side but also on the auxiliary winding. Unfortunately, since the Vcc capacitor and the auxiliary diode form a peak rectifier, the auxiliary Vcc often depends on this peak value rather than the true plateau which corresponds to the output level: http://onsemi.com 18 NCP1239 Leakage effect: Vpeak = 24.2V 25.0 ”clean” plateau V = 13.4V 15.0 0V 5.00 − 5.00 − 15.0 236U 240U 244U 248U 252U The leakage effect seen on the auxiliary side pulls−up the final level peak−rectified by the diode Figure 33. level before the 100 ms period has elapsed, nothing happens and the controller continues working normally. When a fault is detected, we have seen that the controller stops delivering pulses. At this time, Vcc starts to drop because the power supply is locked off. When the Vcc drops below VccOFF (11.2 V typical), it enters a so−called latch−off phase where the internal consumption is reduced down to about 400 µA. The Vcc capacitor continues to deplete, but at a lower rate. When Vcc finally reaches the latch−off level (around 6.9 V), the start−up current source turns on and pulls Vcc above VccON, exactly as a start−up sequence would do. When Vcc exceeds VccON (16.4 V), pulses are delivered and can last 100 ms maximum if there is enough voltage or can be prematurely interrupted if Vcc falls below VccOFF. Figure 34 shows the difference between these two cases. As already explained, in short−circuit bursts, the PFC section is not validated. The short−circuit protection features a so−called auto−recovery circuitry. That is to say, during the 100 ms period, the power supply attempts to start−up. If the fault has gone, then the controller resumes from the fault and the power supply operates again. If the fault is still present, the pulses are stopped at the end of the 100 ms section (Tpulse) for a given time period Tfault. At the end of Tfault, a new 100 ms attempt is made and so on. To avoid any thermal runaway, a burst duty−cycle defined by Tpulse/(Tfault+ Tpulse) below 10% is desirable ((Tfault+Tpulse) is the burst period). If the 100 ms is made by an internal timer in conjunction with the pin6 capacitor, the Tfault duration builds on the Vcc capacitor which is charged / discharged two times. Figure 35 portrays this behavior: On Figure 33’s example, one can clearly observe the difference between the peak and the real plateau DC level. The delta is around 10 V, which obviously degrades the auxiliary image of the secondary side. When a short−circuit occurs, the leakage can be so strong that the whole plateau has dropped to a few volts, but the leakage contribution becomes so energetic (Ip = Ip max.) that even a few µs duration is enough to prevent Vcc auxiliary from collapsing and thus stopping the pulses. Needless to say that over−power detection is simply impossible. Low standby power requirement decreases Vcc at no−load: this is particularly true if you try to reach less than 100 mW at high line. Thanks to skip cycle, the continuous flow of pulses turns into bunches of pulses (sometimes 1−2 pulses only) that can be spaced by 50ms or more in certain cases. The energy content in each bunch of pulses does not suffer any attenuation. For instance, to lower Figure 33’s peak, you could think of inserting a resistor with the auxiliary diode to form a low pass filter with the Vcc capacitor. Unfortunately, it would drastically reduce the Vcc capacitor refueling current and Vcc could not be maintained. To compensate that effect, a solution could be to increase the turn ratio, but then the peak rectification problem comes back again. As one can see, a short−circuit protection free of the Vcc level would be the best solution. This is exactly what the NCP1239 brings you with an internal 100 ms timer. As soon as the internal 0.9 V error flag is asserted high, a 100 ms timer gets started. If the error flag keeps asserted during the 100 ms period, then the controller detects a true fault condition and stops pulsing the output. If this is a simple transient overload, e.g. the error flag goes back to a normal http://onsemi.com 19 NCP1239 Cpin6=390 nF Vcc VccON VccOFF Drv 100ms < 100ms Bunch length given by timer Bunch length given by VccOFF When Vcc drops faster than the timer, it prematurely interrupts the pulses flow. The 100 ms delay could be shortened or lengthened by changing the pin6 capacitor. Figure 34. Cpin6=390 nF Vcc VccON t3 t1 VccOFF t’1 t2 Latch−off phase level t’2 Logic reset level Drv 100ms 100ms The burst period is ensured by the Vcc capacitor charge / discharge cycle The 100 ms delay could be shortened or lengthened by changing the pin6 capacitor. Figure 35. http://onsemi.com 20 NCP1239 If by design we have selected a 47 F Vcc capacitor, it becomes easy to evaluate the burst period and its duty−cycle. This can be done by properly identifying all time events on figure 35 and applying the classical formula: t = C * V / i. To simplify, let’s consider t1 starts while Vcc=VccOFF. Then: • t1: I = ICC3 = 400 A, ∆V= 11.2 – 6.9 = 3 V t1 = 505ms • t2: I = 5.2 mA, ∆V= 16.4 – 6.9 = 9.5 V t2 = 86 ms • t3: I = 400 A, ∆V= 16.4 – 11.2 = 5.2 V t3 = 611 ms • t’1=t1=505 ms • t’2=t2=86 ms The total period duration is thus the sum of all these events which leads to Tfault = 1793 ms. If Tpulse = 100 ms, then our burst duty−cycle equals 100/(1793+100) ≈ 5%, which is excellent. In fact, the calculation assumption, t1 starts while Vcc=VccOFF, gives the worse case since the duty cycle is calculated in the case where Tpulse exactly equals the active phase duration (switching period when Vcc decreases from VccON to VccOFF). In fact, Tpulse is generally: − − shorter than the switching phase period. In this case, t1 is longer since the latched off phase starts earlier (at a Vcc higher than VccOFF). As a consequence, the final duty cycle is lower than previously estimated, longer than the switching phase period. In this case, the circuit detects an over−load condition simply because Vcc drops below VccOFF (11.2 V) before the fault timer has elapsed. Tpulse is lower than 100 ms and as a result the duty cycle is also lower. (Major) Fault Detection and Latched Off Mode The NCP1239 features a fast comparator that permanently monitors the “Fault Detect” pin level. If for any reason this level exceeds 2.4 V (typical), the part immediately stops oscillating and stays latched off until the user cycles down the power supply. This enables the SMPS designer to externally shut down the part in particular when a major default occurs, e.g. an Over Voltage Protection (OVP). Figure 36 shows what happens when the part is latched: Vcc VccON VccOFF Latch−off phase level Logic reset level The user has unplugged, reset! Drv pin3 Stop! 2.4 V When Vpin3 exceeds 2.4 V, NCP1239 permanently latches−off the output pulsesuntil its Vcc goes below 4 V. The figure can illustrate a case where a thermistor supplied by REF5V is connected to pin3 to detect excessive temperatures of the application (refer to application schematic). Figure 36. http://onsemi.com 21 NCP1239 application must not exceed. Choosing R equal to 5k, the pin3 voltage at 130°C that equates: Pin3 can serve to build an Over Voltage Protection by placing a zener between the voltage to measure (e.g., Vcc) and pin3 (refer to application schematic). If a 15 V zener is applied, the pin3 comparator will switch when (Vcc−15 V) exceeds the 2.4 V internal reference, that is, when Vcc is higher than 17.5 V. This pin can also monitor the temperature using an external thermistor (refer to application schematic). Thermistors can be of NTC type (Negative Temperature Coefficient – the resistance decreases versus the temperature) or of PTC type (Positive Temperature Coefficient – the resistance increases versus the temperature). Let’s assume that a NTC thermistor is used (as in the application schematic). Placing it between the 5 V reference voltage (REF5V) and pin3, and a classical resistance between pin3 and ground, the pin3 voltage equals: Vpin3 = Vpin3(130C ) = 5k * 5V = 2.5V 5k + 5k triggers the fault comparator. This example illustrates that one must just select the bottom resistor so that it exhibits the same resistance as the thermistor at the temperature to be detected. If the thermistor is a PTC, it must be placed between pin3 and ground. One must place a resistor between the 5 V reference voltage and pin3. Similarly, the resistor must be selected so that its resistance equals the thermistor one at the temperature to be detected. Brown−Out and Over Power Limitation SMPS are designed for a given input range. When the input voltage is too low (brown−out), the SMPS tends to compensate by sinking an increased current from the line. As a result the power components may suffer from an excessive heating and ultimately the SMPS may be destroyed. To avoid such a risk, the NCP1239 incorporates a brown−out detection that monitors the portion of the input voltage that is applied to pin5. R * 5V R + Rthermistor , where R and Rthermistor are respectively the resistor and the thermistor resistance. Rthermistor decreasing versus the temperature, the pin3 voltage (Vpin3) increases when the temperature grows up. For instance, the thermistor resistance can be in the range of 500 kΩ at 25°C and as low as 5 kΩ at 130°C that as an example, one can take as the temperature limit the HV CMP Rupper CMP 5 Driver + − Driver is off as long as CMP is low Rlower + 500 mV if CMP is low 240 mV if CMP is high Vpin5 240 mV 500 mV An hysteresis comparator monitors the SMPS input voltage Figure 37. comparator toggles, the internal reference voltage changes from 500 mV to 240 mV. This effect is not latched: that is to say, when the bulk capacitor is below the target, the controller does not deliver pulses. As soon as the input voltage grows−up and reaches the level imposed by the resistive divider, pulses are passed to the internal driver and activate the MOSFET. Figure 38 offers a way to connect the elements around pin5 to create a Brown−Out detection: Also called “Bulk OK” signal (BOK), the Brown−Out (BO) protection prevents the power supply from being adversely destroyed in case the mains drops to a very low value. When it detects such a situation, the NCP1239 no longer pulses but waits until the bulk voltage goes back to its normal level. A certain amount of hysteresis needs to be provided since the bulk capacitor is affected by some ripple, especially at low input levels. For that reason, when the BO http://onsemi.com 22 NCP1239 to converter PFC Preconverter Rupper ac line Input Filtering Capacitor Cbulk + 5 Rlower Cfil Example where the voltage of the bulk capacitor is used for the brown−out Protection Figure 38. amount of power, actually the power of your converter (35W in our example). The equation associated to Bload instructs the simulator not to draw current until the Brown−Out converter gives the order, just like what the real converter will do. As a result, Vbulk is free of ripple until the node CMP goes high, giving the green light to switch pulses. The input line is modulated by the “timing” node which ramps up and down to simulate a slow startup / turn−off sequence. Then, by adjusting the Cfil value, it becomes possible to select the right turn−off AC voltage. Figure 40 portrays the typical signal you can expect from the simulator. We measured a turn−on voltage of 85VAC whereas the turn−off voltage is 72VAC. Further increasing Cfil lowers this level (for instance, a 1 uF capacitor gives VBO=65VAC in the example). As we have seen, the load variations will modify this turn−off level. To remove the dependency between VBO and the load, it is possible to directly sense the rectified input line present at the PFC stage input, as figure 41 offers. In that case, there still exists the input line ripple, but this ripple is independent of the load. By adjusting Cfil capacitance and the divider section, you can build a brown−out detection independent of the load. The calculation procedure for Rupper and Rlower is easy. The first level transition is always clean: the SMPS is not working during the start−up sequence and there exists no ripple superimposed on Cbulk. Supposed we want to start the operation at Vbulk = Vtrip = 120 VDC (i.e., VinAC = 85V). 1. Fix a bridge current Ib compatible with your standby requirements, for instance an Ib of 50µA. 2. Then evaluate Rlower by: Rlower = 0.5 / Ib = 10kΩ 3. Calculate Rupper by: (Vtrip – 0.5V) / Ib = (120 – 0.5) / 50µA = 2.39MΩ The second threshold, the level at which the power supply stops (VBO), depends on the capacitor Cfil but also on the selected bulk capacitor. Furthermore, when the load varies, the ripple also does and increases as Vin drops. If Cfil allows a too high ripple, chances exist to prematurely stop the converter. By increasing Cfil, you have the ability to select the amount of hysteresis you want to apply. The less ripple appears on pin5, the larger the gap between Vtrip and VBO (the maximum being VBO = Vtrip/2). The best way to assess the right value of Cfil, is to use a simple simulation sketch as the one depicted by figure 39. A behavioral source loads the rectified DC line and adjusts itself to draw a given http://onsemi.com 23 NCP1239 bulk + VBulk Vline ∆ 2 B1 Voltage V(line)*V(timing) Cbulk 47uF IC = 40 IN Bload Current V(CMP) >3 ? 35/V(bulk) : 0 − 3 PSpice: EBload Value = { IF ( V(CMP)>3, 35/V(bulk), 0) } bulk timing line Rupper 2.4Meg V1 cmp + V2 BrownOut 5 Cfil 220n Rlower 10k − Bbrown Voltage V(CMP) > 3 ? 250m : 500m PSpice: EBbrown Value = { IF ( V(CMP)>3, 250m, 0) } V2 timing 0 PWL 0 0.2 3s 1 7s 1 10s 0.2 V1 line 0 SIN 0 150 50 A simple simulation configuration helps to tailor the right value for Cfil Figure 39. 200 16.0 100 12.0 0 8.0 −100 4.0 −200 0 Turn−off voltage occurs at: VinRMS = 72.3 volts Vbrown−out 8.156 8.175 8.195 8.215 Typical signals obtained from the simulator Figure 40. http://onsemi.com 24 8.235 NCP1239 Rectified ac line sensing to converter PFC Preconverter ac line Rupper Input filtering Capacitor + 5 Cbulk Rlower Cfil A second option to directly sense the mains Figure 41. In addition, it is not recommended to provide the output with more power than normally necessary. To the light of these statements, it becomes interesting to accurately limit the amount of power drawn from the AC line in fault conditions. The easiest way to do so consists of clamping the peak current since in a discontinuous mode flyback converter, the input power (Pin) can be calculated as follows: Pin = 1/2 * Lp * Ippk2 * fsw, where Lp is the primary inductor, Ippk is the inductor peak current and fsw is the switching frequency. Practically, a sense resistor converts the primary current into a voltage that is compared to a voltage reference. When the voltage representative of the current exceeds the voltage reference, the controller turns off the power switch. The theoretical maximum peak current is then: Imax = Vocp / Rsense, where Vocp is the reference voltage (or over current protection threshold) and Rsense is the sense resistor. Unfortunately, the controller cannot turn off the power switch immediately when it detects that the current exceeds its maximum permissible level. Internal propagation delays differ the drive turn low. In addition, the power switch needs some time to turn off. Finally, the real current stop can be 250ns or more delayed. During this time, the current continues ramping up so that an over−current is obtained. This second option that directly senses the input voltage (see figure 41), enables a more direct under−mains detection. Even in a brown−out conditions, the PFC pre−converter may be able to maintain a sufficient bulk voltage, possibly at the price of some excessive stress. Measuring the rectified ac line instead of the bulk voltage, the NCP1239 more surely protects the PFC stage in brown−out conditions. Using: − Rlower = 10 kΩ, − Rupper = 2. 39 MΩ, − Cfil = 1 F, One obtains the following voltage thresholds: − Vtrip = 85 Vrms, − VBO = 65 Vrms. Over−Power Limit Overload conditions may push the converter to draw an excessive power (which generally increases versus the input voltage). One must avoid such a behavior: a) not to have to dimension the converter for a power higher than the nominal one, b) to meet SMPS specifications that often request the power not to exceed a given level. http://onsemi.com 25 NCP1239 Actual Peak Current Low Input Voltage δIHL δILL Vopl/Rsense Wished Maximum Peak Current High Input Voltage δt δt The propagation delay (t) produces over−currents (ILL at low line, IHL at high line in the figure) that are proportional to the input voltage. As a consequence, the actual maximum current and then the power limit gets higher when the ac line increases. Figure 42. Imax = Vocp Vin * δt + Rsense Lp Then, Ipth = , where Vin is the converter input voltage and δt is the total delay in turning off the power switch. The NCP1239 enables the compensation of the second term in the Imax equation for a precise limitation of the peak current. A current source (Ipin9) proportional to the pin5 voltage flows out of pin9. Since pin5 receives a voltage proportional to the input voltage for brown−out detection, Ipin9 is proportional to the input voltage too. An external resistor Rcomp can be connected between pin9 and the positive terminal of Rsense, so that pin9 monitors the following voltage: Vpin9 = [Rsense * (Ip + Ipin9)] + (Rcomp * Ipin9) Ipin9 being small compared to the inductor current, the pin9 voltage simplifies as follows: Vpin9 = (Rsense * Ip) + (Rcomp * Ipin9) Ipin9 is proportional to the pin5 voltage (80 µA/V*Vpin5 – see parameters specification table) and Vpin5 is a portion of the input voltage (Vpin5=kBO*Vin). Finally, Ipin9 = 80 A/V * kBO * Vin The voltage Vpin9 is compared to the internal reference Vocp. When Vpin9 reaches Vocp, the corresponding threshold current (Ipth) is deducted from: Vopl = (Rsense * Ipth) + (Rcomp * 80 A/V * kBO * Vin) Vopl − Rcomp * 80 A/V * k BO * Vin Rsense Taking into account the over−current resulting from the propagation delays, the maximum current is finally: Rcomp * 80A/V * k BO * Vin Vin * δt Vocp − + Imax = Rsense Rsense Lp Rcomp * 80 A/V * k BO δt = Rsense Lp Choosing Rcomp so that , the current limit is made constant in the whole input voltage range (Imax = Vocp / Rsense). As an example, let’s assume that: − the minimum input voltage for operation is 100 V => kBO=0.5/100=0.005, − Rsense is 0.25 Ω, − Lp=500 µH, − The total propagation delays are δt=350 ns, Then, the Rcomp resistor should be: Rcomp = http://onsemi.com 26 δt * Rsense 350n * 0.25 = ≈ 438Ω 80 * k BO * Lp 80 * 0.005 * 500 m NCP1239 Vdd CLK Rbo1 Vin PWM Latch Vpin5 80A/V*Vpin5 S 5 Q Rbo2 Brown−Out Q Cbo to Brown−Out Comparator R Vcomp = k*Rcomp*Vpin5 Rcomp + LEB 9 Over Power Limit − + Vcomp Rsense 0.5V to Current Sense Comparator Rramp 10 Current Sense An (averaged) portion of the input voltage is applied to the brown−out pin. A current source proportional to this voltage, flows through an external resistor Rcomp to form an offset proportional to the (average) input voltage. Rcomp should be selected so that the offset compensates the over−current sensed by the current sensing resistor Rsense. Figure 43. Soft−Start The NCP1239 features an internal soft−start activated during the power on sequence (PON). As soon as Vcc reaches 16.4 V, the current set−point is gradually increased from nearly zero up to the maximum clamping level (e.g. 0.9V/Rsense). This situation lasts a programmable time that is adjusted by the pin6 capacitor (7.5 ms typically with Cpin6=390 nF). Further to that time period, the current set−point is blocked to 0.9V/Rsense until the supply enters regulation. The soft−start is also activated at each start of the active phase of fault burst operation. Every re−start attempt is followed by a soft−start activation. 16.4V 6.9V Soft−start is activated during a start−up sequence or an OVL condition Figure 44. http://onsemi.com 27 NCP1239 Generally speaking, the soft−start will be activated when Vcc ramps up either from zero (fresh power−on sequence) or 6.9 V, the latch−off threshold after an over−load detection (OVL) for instance. Figure 44 portrays the soft−start behavior which a 390 nF soft−start capacitor on pin6. The time scales are purposely shifted to offer a better zoom portion. In the NCP1239, the ramp features a swing of 3.2 V. Suppose we select a 65 kHz version. Over a 65 kHz frequency, it corresponds to a 130 mV/ms ramp. In our FLYBACK design, let’s assume that our primary inductance Lp is 350 mH, and the SMPS delivers 12 V with a Np:Ns ratio of 1:0.1. The OFF time slope of the primary current is: Vout + Vf ⋅ Ns Np Lp that is, 371 mA/ms or 37 mV/ms, once Internal Ramp Compensation Ramp compensation is a known mean to cure sub−harmonic oscillations. These oscillations take place at half the switching frequency and occur only during Continuous Conduction Mode (CCM) with a duty−cycle greater than 50%. To lower the current loop gain, one usually injects between 50 and 100% of the inductor down−slope. Figure 45 depicts how internally the ramp is generated: projected over a 0.1 Rsense for instance. If we select 75% of the down−slope as the required amount of ramp compensation, then we shall inject 27 mV/ms. Our internal compensation being of 208 mV/ms, the divider ratio (divratio) between Rramp and the 32 k is 0.178. A few lines of algebra to determine Rramp: 19k ⋅ divratio Rramp = 1 − divratio = 6.92k. 3.2V The ramp is disabled during stand−by (i.e., when pfcON is low). This inhibition avoids that the ramp compensation modifies the set−point above which the NCP1239 enables PFC. 0V 32k Rramp Frequency Jittering Frequency jittering is a method used to soften the EMI signature by spreading the energy in the vicinity of the main switching component. NCP1239 offers a +3.5% deviation of the nominal switching frequency. The sweep saw−tooth is internally generated and modulates the clock up and down with a period depending on the pin6 capacitor (10 ms typically with 390 nF). Again, if one selects a 65 kHz version, the frequency will equal 65 kHz in the middle of the ripple and will increase as Vpin6 rises or decrease as Vpin6 ramps down. Figure 46 portrays the behavior we have adopted: LEB CS + − Rsense from setpoint Inserting a resistor in series with the current sense information brings ramp compensation Figure 45. Internal ramp 67.6kHz 65kHz Internal sawtooth 62.4kHz 10ms The Vpin6 ramp is used to introduce frequency jittering on the oscillator saw−tooth Figure 46. http://onsemi.com 28 NCP1239 Skipping Cycle Mode The NCP1239 automatically skips switching cycles when the output power demand drops below a given level. This is accomplished by monitoring the FB pin. In normal operation, pin8 imposes a current set−point accordingly to the load value. If the load demand decreases, the internal loop asks for less peak current. When this set−point reaches a fixed determined level, the IC prevents the current from decreasing further down and starts to blank the output pulses: the IC enters the so−called skip cycle mode, also named controlled burst operation. The default skip cycle current is internally frozen to 30% of the maximum peak current which is 0.5V/Rsense The power transfer now depends upon the width of the pulse bunches (figure 47). Suppose we have the following component values: Lp, primary inductance = 350 mH Fsw , switching frequency = 65 kHz Ip skip = 600 mA (or 150 mV / Rsense) The theoretical power transfer is therefore: 1/2 * Lp * Ip2 * Fsw = 4 W If this IC enters skip cycle mode with a bunch length of 10ms over a recurrent period of 100ms, then the total power transfer is: 4 W * 10 ms / 100 ms = 400 mW To better understand how this skip cycle mode takes place, a look at the operation mode versus the FB level immediately gives the necessary insight: FB pin Voltage 5V, FB pin open 2.7V upper dynamic range Normal current mode operation 0.45V Skip cycle operation Ip MIN = 150mV / Rsense Figure 47. When FB is below the skip cycle threshold (0.45 V by default), the circuit skips the switching cycle. When the IC enters the skip cycle mode, the peak current cannot go below (0.45 V/3)/Rsense or 150 mV/Rsense. Figure 48 shows different values of pulse widths when the SMPS starts to skip cycles at different power levels: Power P1 Power P2 Power P3 Output pulses at various power levels (X = 5s/div) P1 < P2 < P3 Figure 48. http://onsemi.com 29 NCP1239 300.0M Max peak current 200.0M 25% of max Ip 100.0M 0 315.4U 882.7U 1.450M 2.017M 2.585M The skip cycle takes place at low peak currents which guaranties noise free operation Figure 49. PFC Inhibition in Stand−by The circuit detects a light load condition by permanently monitoring the skip cycle comparator activity: in normal load condition this comparator keeps quiet. As soon as the load strongly decreases, this comparator starts to toggle at a low frequency rate: we are entering skip cycle and the opto−coupler operates in a digital manner, ON/OFF. Figure 49 shows the way skip cycle is detected. In skip mode, the feed−back voltage oscillates around Vpin7 (If no voltage is applied to the pin7, a 430 mV voltage source supplies a default value through a high impedance resistor). In these conditions, the skip comparator (“COMP1”) that turns on and off (to adjust the skip mode bunches of pulses), sets the stand−by detection latch. A second comparator (“COMP2”) compares the feed−back voltage (FB or Vpin8) to 1.7*Vpin7. As long as the load keeps light, FB does not exceed 1.7*Vpin7 (i.e., 0.74 V typical if no voltage is forced to pin7). A timer counts down and if COMP2 keeps high for 100 ms (typically with 390 nF on pin6), the NCP1239 considers that the system runs in the stand−by mode. Pin1 turns high, a 10kΩ resistor tying the pin to Vcc. If as shown in figure 32, pin1 directly drives a pnp transistor that is connected between Vcc and the PFC Vcc, this switch turns off in stand−by. As a result, this transistor stops feeding the PFC Vcc and ultimately shuts the PFC down. As soon as FB exceeds 1.7*Vpin7, the circuit leaves the stand−by mode without any delay by forcing a 1mA sinking current source on pin1, that re−activates the pnp transistor and then the PFC stage. One can note that there is a 1/3 ratio between the actual current set−point and the feed−back value FB. Therefore the default thresholds for stand−by detection and normal mode recovery (0.43 V, 0.74 V) actually corresponds to the 140 mV and 250 mV set−points. 70% A delay is inserted to avoid false tripping of the GTS signal Figure 50. http://onsemi.com 30 NCP1239 One clearly sees that the GTS signal does not react to the fugitive low FB pin condition during startup Figure 51. FB < Vpin7 => Skip high REF5V Skip + R1 Skip Adjust 100k − FB COMP1 7 + 0.43V R2 Stby_detect S Q 15r Q 25r COMP2 + R − FB > 1.7*Vpin7 => Stby_detect RESET GTS 100 ms timer (*) (SS and timer block) pin1 (*) the 100 ms delay is programmed by the pin6 capacitor Internal Go−To−Standby signal elaboration Figure 52. Suppose our Flyback controller is built with a transformer primary inductance of 250 µH. To pass 120 W, we assume that a peak current of 4.2 A was needed. Thanks to these numbers, we can easily now when the GTS signal will be asserted: Lp, primary inductance = 250 H = 85% Fsw , switching frequency = 65 kHz Ip = 2⋅Pout η⋅Lp⋅Fsw Ip skip = 30% of Ip max = 1.26 A The theoretical region at which the SMPS will enter standby is: 1/2 * Lp * Ip * Fsw * 11 W. This number can vary depending on the line level since the propagation delay becomes a sensitive parameter, and on the efficiency that is difficult to precisely predict in light load conditions. The peak current at which the SMPS will leave standby is 48% of the peak current which means that a power of 28 W is necessary to re−trigger the PFC. = 4.2 A http://onsemi.com 31 NCP1239 INFORMATIVE WAVEFORMS The following plots were obtained using a 150 W application (output 19 V/7 A). The NCP1239 enables the PFC Vcc as soon as the FB pin voltage has gone below a threshold (about 2.7 V), that is when the internal error flag stops being asserted. Figure 53. Start−up Sequence http://onsemi.com 32 NCP1239 The feed−back voltage goes high and asserts the internal error flag. The pin6 timer counts for about 100 ms (Cpin6=390ns) before shutting down the SMPS. One “Vcc cycle over two is skipped” to limit the duty cycle in overload. Figure 54. Overload Conditions http://onsemi.com 33 NCP1239 When the load current falls to a low level (CH4), the FB pin voltage diminishes to take into account the decay of the power demand. As a consequence, the FB pin voltage goes below the “Vskip” threshold and the soft start timer counts about 100ms (if Cpin6=330nF). When the 100ms time has elapsed, the PFC Vcc stops being fed. Figure 55. Transition Normal to Stand−by http://onsemi.com 34 NCP1239 When the load current increases from 1A to 5A, the FB pin increases too so that the supplied power matches the new demand. The normal mode is recovered without delay. Figure 56. Transition stand−by to normal http://onsemi.com 35 NCP1239 PACKAGE DIMENSIONS SO−16 FD SUFFIX CASE 751B−05 ISSUE J −A− 16 NOTES: 1. DIMENSIONING AND TOLERANCING PER ANSI Y14.5M, 1982. 2. CONTROLLING DIMENSION: MILLIMETER. 3. DIMENSIONS A AND B DO NOT INCLUDE MOLD PROTRUSION. 4. MAXIMUM MOLD PROTRUSION 0.15 (0.006) PER SIDE. 5. DIMENSION D DOES NOT INCLUDE DAMBAR PROTRUSION. ALLOWABLE DAMBAR PROTRUSION SHALL BE 0.127 (0.005) TOTAL IN EXCESS OF THE D DIMENSION AT MAXIMUM MATERIAL CONDITION. 9 −B− 1 P 8 PL 0.25 (0.010) 8 M B S G R K F X 45 C −T− SEATING PLANE J M D 16 PL 0.25 (0.010) M T B S A S DIM A B C D F G J K M P R MILLIMETERS MIN MAX 9.80 10.00 3.80 4.00 1.35 1.75 0.35 0.49 0.40 1.25 1.27 BSC 0.19 0.25 0.10 0.25 0 7 5.80 6.20 0.25 0.50 INCHES MIN MAX 0.386 0.393 0.150 0.157 0.054 0.068 0.014 0.019 0.016 0.049 0.050 BSC 0.008 0.009 0.004 0.009 0 7 0.229 0.244 0.010 0.019 The product described herein (NCP1239), may be covered by one or more of the following U.S. patents: 6,362,067, 6,385,060, 6,429,709. There may be other patents pending. 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 ON Semiconductor Website: http://onsemi.com Order Literature: http://www.onsemi.com/litorder Japan: ON Semiconductor, Japan Customer Focus Center 2−9−1 Kamimeguro, Meguro−ku, Tokyo, Japan 153−0051 Phone: 81−3−5773−3850 http://onsemi.com 36 For additional information, please contact your local Sales Representative. NCP1239/D